platform/kernel/linux-rpi.git
22 months agoBluetooth: btusb: Ignore zero length of USB packets on ALT 6 for specific chip
Hilda Wu [Wed, 5 Oct 2022 08:43:31 +0000 (16:43 +0800)]
Bluetooth: btusb: Ignore zero length of USB packets on ALT 6 for specific chip

For USB ALT 6 settings some Realtek chips need to transmit mSBC data
continuously without the zero length of USB packets.
In this commit, create BTUSB_ALT6_CONTINUOUS_TX to manage the behavior.
Therefore, create REALTEK_ALT6_CONTINUOUS_TX_CHIP to manage the specific
chip model for the behavior.

Signed-off-by: Max Chou <max.chou@realtek.com>
Signed-off-by: Hilda Wu <hildawu@realtek.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
22 months agoBluetooth: btrtl: Add btrealtek data struct
Hilda Wu [Wed, 5 Oct 2022 08:43:30 +0000 (16:43 +0800)]
Bluetooth: btrtl: Add btrealtek data struct

This patch adds a data structure for btrealtek object, and the
definition of vendor behavior flags. It also adds macros to set/test/get
the flags.

Signed-off-by: Hilda Wu <hildawu@realtek.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
22 months agoBluetooth: virtio_bt: fix device removal
Michael S. Tsirkin [Mon, 10 Oct 2022 17:14:37 +0000 (13:14 -0400)]
Bluetooth: virtio_bt: fix device removal

Device removal is clearly out of virtio spec: it attempts to remove
unused buffers from a VQ before invoking device reset. To fix, make
open/close NOPs and do all cleanup/setup in probe/remove.

NB: This is a hacky way to handle this - virtbt_{open,close} as NOP is
not really what a driver is supposed to be doing. These are transport
enable/disable callbacks from the BT core towards the driver. It maps to
a device being enabled/disabled by something like bluetoothd for
example. So if disabled, users expect that no resources/queues are in
use.  It does work with all other transports like USB, SDIO, UART etc.
There should be no buffer used if the device is powered off. We also
don’t have any USB URBs in-flight if the transport is not active.

The way to implement a proper fix would be using vq reset if supported,
or even using a full device reset.

The cost of the hack is a single skb wasted on an unused bt device.

NB2: with this fix in place driver still suffers from a race condition
if an interrupt triggers while device is being reset.  To fix, in the
virtbt_close() callback we should deactivate all interrupts.  To be
fixed.

squashed fixup: bluetooth: virtio_bt: fix an error code in probe()

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20220811080943.198245-1-mst@redhat.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Tested-by: Igor Skalkin <Igor.Skalkin@opensynergy.com>
22 months agoBluetooth: hci_sync: cancel cmd_timer if hci_open failed
Archie Pusaka [Wed, 5 Oct 2022 07:09:47 +0000 (15:09 +0800)]
Bluetooth: hci_sync: cancel cmd_timer if hci_open failed

If a command is already sent, we take care of freeing it, but we
also need to cancel the timeout as well.

Signed-off-by: Archie Pusaka <apusaka@chromium.org>
Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@google.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
22 months agoBluetooth: hci_sync: Fix not able to set force_static_address
Luiz Augusto von Dentz [Sat, 8 Oct 2022 01:08:43 +0000 (18:08 -0700)]
Bluetooth: hci_sync: Fix not able to set force_static_address

force_static_address shall be writable while hdev is initing but is not
considered powered yet since the static address is written only when
powered.

Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Tested-by: Brian Gix <brian.gix@intel.com>
22 months agoBluetooth: hci_sync: Fix not setting static address
Luiz Augusto von Dentz [Fri, 7 Oct 2022 20:25:38 +0000 (13:25 -0700)]
Bluetooth: hci_sync: Fix not setting static address

This attempts to program the address stored in hdev->static_addr after
the init sequence has been complete:

@ MGMT Command: Set Static A.. (0x002b) plen 6
        Address: C0:55:44:33:22:11 (Static)
@ MGMT Event: Command Complete (0x0001) plen 7
      Set Static Address (0x002b) plen 4
        Status: Success (0x00)
        Current settings: 0x00008200
          Low Energy
          Static Address
@ MGMT Event: New Settings (0x0006) plen 4
        Current settings: 0x00008200
          Low Energy
          Static Address
< HCI Command: LE Set Random.. (0x08|0x0005) plen 6
        Address: C0:55:44:33:22:11 (Static)
> HCI Event: Command Complete (0x0e) plen 4
      LE Set Random Address (0x08|0x0005) ncmd 1
        Status: Success (0x00)
@ MGMT Event: Command Complete (0x0001) plen 7
      Set Powered (0x0005) plen 4
        Status: Success (0x00)
        Current settings: 0x00008201
          Powered
          Low Energy
          Static Address
@ MGMT Event: New Settings (0x0006) plen 4
        Current settings: 0x00008201
          Powered
          Low Energy
          Static Address

Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Tested-by: Brian Gix <brian.gix@intel.com>
22 months agoBluetooth: Work around SCO over USB HCI design defect
Nicolas Cavallari [Wed, 5 Oct 2022 15:06:21 +0000 (17:06 +0200)]
Bluetooth: Work around SCO over USB HCI design defect

The USB interface between the host and the bluetooth adapter used for
SCO packets uses an USB isochronous endpoint with a fragmentation scheme
that does not tolerate errors.  Except USB isochronous transfers do
not provide a reliable stream with guaranteed delivery. (There is no
retry on error, see USB spec v2.0 5.6 and 8.5.5.)

To fragment a packet, the bluetooth HCI simply splits it in parts and
transfer them as-is.  The receiver is expected to reconstruct the packet
by assuming the first fragment contains the header and parsing its size
field.  There is no error detection either.

If a fragment is lost, the end result is that the kernel is no longer
synchronized and will pass malformed data to the upper layers, since it
has no way to tell if the first fragment is an actual first fragment or
a continuation fragment.  Resynchronization can only happen by luck and
requires an unbounded amount of time.

The typical symptom for a HSP/HFP bluetooth headset is that the
microphone stops working and dmesg contains piles of rate-limited
"Bluetooth: hci0: SCO packet for unknown connection handle XXXX"
errors for an indeterminate amount of time, until the kernel accidentally
resynchronize.

A workaround is to ask the upper layer to prevalidate the first fragment
header.  This is not possible with user channels so this workaround is
disabled in this case.

This problem is the most severe when using an ath3k adapter on an i.MX 6
board, where packet loss occur regularly, possibly because it is an USB1
device connected on an USB2 hub and this is a special case requiring
split transactions.

Signed-off-by: Nicolas Cavallari <nicolas.cavallari@green-communications.fr>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
22 months agoBluetooth: btusb: Introduce generic USB reset
Archie Pusaka [Thu, 6 Oct 2022 09:09:31 +0000 (17:09 +0800)]
Bluetooth: btusb: Introduce generic USB reset

On cmd_timeout with no reset_gpio, reset the USB port as a last
resort.

This patch changes the behavior of btusb_intel_cmd_timeout and
btusb_rtl_cmd_timeout.

Signed-off-by: Archie Pusaka <apusaka@chromium.org>
Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@google.com>
Reviewed-by: Ying Hsu <yinghsu@chromium.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
22 months agobtusb: Avoid reset of ISOC endpoint alt settings to zero
Chethan Tumkur Narayan [Mon, 3 Oct 2022 06:48:36 +0000 (12:18 +0530)]
btusb: Avoid reset of ISOC endpoint alt settings to zero

In case of suspend/resume and HCI_RESET (BT On and Off),
ISOC endpoint set to alt setting 0 when no SCO connection exists.
This patch shall avoid resetting of ISOC endpoint to alt setting to 0.

Signed-off-by: Chethan Tumkur Narayan <chethan.tumkur.narayan@intel.com>
Signed-off-by: Kiran K <kiran.k@intel.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
22 months agodt-bindings: net: Convert Socionext NetSec Ethernet to DT schema
Rob Herring [Fri, 9 Dec 2022 17:15:52 +0000 (11:15 -0600)]
dt-bindings: net: Convert Socionext NetSec Ethernet to DT schema

Convert the Socionext NetSec Ethernet binding to DT schema format.

Signed-off-by: Rob Herring <robh@kernel.org>
Acked-by: Jassi Brar <jaswinder.singh@linaro.org>
Link: https://lore.kernel.org/r/20221209171553.3350583-1-robh@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMAINTAINERS: Update email address for Marvell Prestera Ethernet Switch driver
Taras Chornyi [Fri, 9 Dec 2022 15:45:21 +0000 (17:45 +0200)]
MAINTAINERS: Update email address for Marvell Prestera Ethernet Switch driver

Taras's Marvell email account will be shut down soon so change it to Plvision.

Signed-off-by: Taras Chornyi <taras.chornyi@plvision.eu>
Signed-off-by: Vadym Kochan <vadym.kochan@plvision.eu>
Link: https://lore.kernel.org/r/20221209154521.1246881-1-vadym.kochan@plvision.eu
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: hns3: use strscpy() to instead of strncpy()
Xu Panda [Fri, 9 Dec 2022 07:38:59 +0000 (15:38 +0800)]
net: hns3: use strscpy() to instead of strncpy()

The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.

Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/202212091538591375035@zte.com.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agosctp: sysctl: make extra pointers netns aware
Firo Yang [Fri, 9 Dec 2022 05:48:54 +0000 (13:48 +0800)]
sctp: sysctl: make extra pointers netns aware

Recently, a customer reported that from their container whose
net namespace is different to the host's init_net, they can't set
the container's net.sctp.rto_max to any value smaller than
init_net.sctp.rto_min.

For instance,
Host:
sudo sysctl net.sctp.rto_min
net.sctp.rto_min = 1000

Container:
echo 100 > /mnt/proc-net/sctp/rto_min
echo 400 > /mnt/proc-net/sctp/rto_max
echo: write error: Invalid argument

This is caused by the check made from this'commit 4f3fdf3bc59c
("sctp: add check rto_min and rto_max in sysctl")'
When validating the input value, it's always referring the boundary
value set for the init_net namespace.

Having container's rto_max smaller than host's init_net.sctp.rto_min
does make sense. Consider that the rto between two containers on the
same host is very likely smaller than it for two hosts.

So to fix this problem, as suggested by Marcelo, this patch makes the
extra pointers of rto_min, rto_max, pf_retrans, and ps_retrans point
to the corresponding variables from the newly created net namespace while
the new net namespace is being registered in sctp_sysctl_net_register.

Fixes: 4f3fdf3bc59c ("sctp: add check rto_min and rto_max in sysctl")
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Firo Yang <firo.yang@suse.com>
Link: https://lore.kernel.org/r/20221209054854.23889-1-firo.yang@suse.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: ethernet: ti: am65-cpsw: Fix PM runtime leakage in am65_cpsw_nuss_ndo_slave_open()
Roger Quadros [Thu, 8 Dec 2022 10:55:34 +0000 (12:55 +0200)]
net: ethernet: ti: am65-cpsw: Fix PM runtime leakage in am65_cpsw_nuss_ndo_slave_open()

Ensure pm_runtime_put() is issued in error path.

Reported-by: Jakub Kicinski <kuba@kernel.org>
Fixes: 93a76530316a ("net: ethernet: ti: introduce am65x/j721e gigabit eth subsystem driver")
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Link: https://lore.kernel.org/r/20221208105534.63709-1-rogerq@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge tag 'wireless-next-2022-12-12' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Mon, 12 Dec 2022 20:15:23 +0000 (12:15 -0800)]
Merge tag 'wireless-next-2022-12-12' of git://git./linux/kernel/git/wireless/wireless-next

Kalle Valo says:

====================
wireless-next patches for v6.2

Fourth set of patches for v6.2. Few final patches, a big change is
that rtw88 now has USB support.

Major changes:

rtw88
 * support USB devices rtw8821cu, rtw8822bu, rtw8822cu and rtw8723du

* tag 'wireless-next-2022-12-12' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (43 commits)
  wifi: rtl8xxxu: fixing IQK failures for rtl8192eu
  wifi: rtlwifi: btcoexist: fix conditions branches that are never executed
  wifi: rtlwifi: rtl8192se: remove redundant rtl_get_bbreg() call
  wifi: rtw88: Add rtw8723du chipset support
  wifi: rtw88: Add rtw8822cu chipset support
  wifi: rtw88: Add rtw8822bu chipset support
  wifi: rtw88: Add rtw8821cu chipset support
  wifi: rtw88: Add common USB chip support
  wifi: rtw88: iterate over vif/sta list non-atomically
  wifi: rtw88: Drop coex mutex
  wifi: rtw88: Drop h2c.lock
  wifi: rtw88: Drop rf_lock
  wifi: rtw88: Call rtw_fw_beacon_filter_config() with rtwdev->mutex held
  wifi: rtw88: print firmware type in info message
  wifi: rtw89: add join info upon create interface
  wifi: rtw89: fix unsuccessful interface_add flow
  wifi: rtw89: stop mac port function when stop_ap()
  wifi: rtw89: add mac TSF sync function
  wifi: rtw89: request full firmware only once if it's early requested
  wifi: rtw89: don't request partial firmware if SECURITY_LOADPIN_ENFORCE
  ...
====================

Link: https://lore.kernel.org/r/20221212093026.5C5AEC433D2@smtp.kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf...
Jakub Kicinski [Mon, 12 Dec 2022 19:27:41 +0000 (11:27 -0800)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf-next

Alexei Starovoitov says:

====================
pull-request: bpf-next 2022-12-11

We've added 74 non-merge commits during the last 11 day(s) which contain
a total of 88 files changed, 3362 insertions(+), 789 deletions(-).

The main changes are:

1) Decouple prune and jump points handling in the verifier, from Andrii.

2) Do not rely on ALLOW_ERROR_INJECTION for fmod_ret, from Benjamin.
   Merged from hid tree.

3) Do not zero-extend kfunc return values. Necessary fix for 32-bit archs,
   from Björn.

4) Don't use rcu_users to refcount in task kfuncs, from David.

5) Three reg_state->id fixes in the verifier, from Eduard.

6) Optimize bpf_mem_alloc by reusing elements from free_by_rcu, from Hou.

7) Refactor dynptr handling in the verifier, from Kumar.

8) Remove the "/sys" mount and umount dance in {open,close}_netns
  in bpf selftests, from Martin.

9) Enable sleepable support for cgrp local storage, from Yonghong.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (74 commits)
  selftests/bpf: test case for relaxed prunning of active_lock.id
  selftests/bpf: Add pruning test case for bpf_spin_lock
  bpf: use check_ids() for active_lock comparison
  selftests/bpf: verify states_equal() maintains idmap across all frames
  bpf: states_equal() must build idmap for all function frames
  selftests/bpf: test cases for regsafe() bug skipping check_id()
  bpf: regsafe() must not skip check_ids()
  docs/bpf: Add documentation for BPF_MAP_TYPE_SK_STORAGE
  selftests/bpf: Add test for dynptr reinit in user_ringbuf callback
  bpf: Use memmove for bpf_dynptr_{read,write}
  bpf: Move PTR_TO_STACK alignment check to process_dynptr_func
  bpf: Rework check_func_arg_reg_off
  bpf: Rework process_dynptr_func
  bpf: Propagate errors from process_* checks in check_func_arg
  bpf: Refactor ARG_PTR_TO_DYNPTR checks into process_dynptr_func
  bpf: Skip rcu_barrier() if rcu_trace_implies_rcu_gp() is true
  bpf: Reuse freed element in free_by_rcu during allocation
  selftests/bpf: Bring test_offload.py back to life
  bpf: Fix comment error in fixup_kfunc_call function
  bpf: Do not zero-extend kfunc return values
  ...
====================

Link: https://lore.kernel.org/r/20221212024701.73809-1-alexei.starovoitov@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge tag 'linux-can-next-for-6.2-20221212' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Mon, 12 Dec 2022 12:11:37 +0000 (12:11 +0000)]
Merge tag 'linux-can-next-for-6.2-20221212' of git://git./linux/kernel/git/mkl/linux-can-next

Marc Kleine-Budde says:

====================
linux-can-next-for-6.2-20221212

this is a pull request of 39 patches for net-next/master.

The first 2 patches are by me fix a warning and coding style in the
kvaser_usb driver.

Vivek Yadav's patch sorts the includes of the m_can driver.

Biju Das contributes 5 patches for the rcar_canfd driver improve the
support for different IP core variants.

Jean Delvare's patch for the ctucanfd drops the dependency on
COMPILE_TEST.

Vincent Mailhol's patch sorts the includes of the etas_es58x driver.

Haibo Chen's contributes 2 patches that add i.MX93 support to the
flexcan driver.

Lad Prabhakar's patch updates the dt-bindings documentation of the
rcar_canfd driver.

Minghao Chi's patch converts the c_can platform driver to
devm_platform_get_and_ioremap_resource().

In the next 7 patches Vincent Mailhol adds devlink support to the
etas_es58x driver to report firmware, bootloader and hardware version.

Xu Panda's patch converts a strncpy() -> strscpy() in the ucan driver.

Ye Bin's patch removes a useless parameter from the AF_CAN protocol.

The next 2 patches by Vincent Mailhol and remove unneeded or unused
pointers to struct usb_interface in device's priv struct in the ucan
and gs_usb driver.

Vivek Yadav's patch cleans up the usage of the RAM initialization in
the m_can driver.

A patch by me add support for SO_MARK to the AF_CAN protocol.

Geert Uytterhoeven's patch fixes the number of CAN channels in the
rcan_canfd bindings documentation.

In the last 11 patches Markus Schneider-Pargmann optimizes the
register access in the t_can driver and cleans up the tcan glue
driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoMerge patch series "can: m_can: Optimizations for tcan and peripheral chips"
Marc Kleine-Budde [Mon, 12 Dec 2022 11:01:26 +0000 (12:01 +0100)]
Merge patch series "can: m_can: Optimizations for tcan and peripheral chips"

Markus Schneider-Pargmann <msp@baylibre.com> says:

as requested I split the series into two parts. This is the first
parts with simple improvements to reduce the number of SPI transfers.
The second part will be the rest with coalescing support and more
complex optimizations.

Changes since v1: https://lore.kernel.org/all/20221116205308.2996556-1-msp@baylibre.com
- Fixed register ranges
- Added fixes: tag for two patches

Link: https://lore.kernel.org/all/20221206115728.1056014-1-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: tcan4x5x: Specify separate read/write ranges
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:28 +0000 (12:57 +0100)]
can: tcan4x5x: Specify separate read/write ranges

Specify exactly which registers are read/writeable in the chip. This
is supposed to help detect any violations in the future.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-12-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: tcan4x5x: Fix register range of first two blocks
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:27 +0000 (12:57 +0100)]
can: tcan4x5x: Fix register range of first two blocks

According to the datasheet 0x10 is the last register in the first block,
not register 0x2c.

The datasheet lists the last register of the second block as 0x830, not
0x83c.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-11-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: tcan4x5x: Fix use of register error status mask
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:26 +0000 (12:57 +0100)]
can: tcan4x5x: Fix use of register error status mask

TCAN4X5X_ERROR_STATUS is not a status register that needs clearing
during interrupt handling. Instead this is a masking register that masks
error interrupts. Writing TCAN4X5X_CLEAR_ALL_INT to this register
effectively masks everything.

Rename the register and mask all error interrupts only once by writing
to the register in tcan4x5x_init.

Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-10-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: tcan4x5x: Remove invalid write in clear_interrupts
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:25 +0000 (12:57 +0100)]
can: tcan4x5x: Remove invalid write in clear_interrupts

Register 0x824 TCAN4X5X_MCAN_INT_REG is a read-only register. Any writes
to this register do not have any effect.

Remove this write. The m_can driver aldready clears the interrupts in
m_can_isr() by writing to M_CAN_IR which is translated to register
0x1050 which is a writable version of this register.

Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-9-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Batch acknowledge rx fifo
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:24 +0000 (12:57 +0100)]
can: m_can: Batch acknowledge rx fifo

Instead of acknowledging every item of the fifo, only acknowledge the
last item read. This behavior is documented in the datasheet. The new
getindex will be the acknowledged item + 1.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-8-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Batch acknowledge transmit events
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:23 +0000 (12:57 +0100)]
can: m_can: Batch acknowledge transmit events

Transmit events from the txe fifo can be batch acknowledged by
acknowledging the last read txe fifo item. This will save txe_count
writes which is important for peripheral chips.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-7-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Count read getindex in the driver
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:22 +0000 (12:57 +0100)]
can: m_can: Count read getindex in the driver

The getindex gets increased by one every time. We can calculate the
correct getindex in the driver and avoid the additional reads of rxfs.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-6-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Count TXE FIFO getidx in the driver
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:21 +0000 (12:57 +0100)]
can: m_can: Count TXE FIFO getidx in the driver

The getindex simply increases by one for every iteration. There is no
need to get the current getidx every time from a register. Instead we
can just count and wrap if necessary.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-5-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Read register PSR only on error
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:20 +0000 (12:57 +0100)]
can: m_can: Read register PSR only on error

Only read register PSR if there is an error indicated in irqstatus.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-4-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Avoid reading irqstatus twice
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:19 +0000 (12:57 +0100)]
can: m_can: Avoid reading irqstatus twice

For peripheral devices the m_can_rx_handler is called directly after
setting cdev->irqstatus. This means we don't have to read the irqstatus
again in m_can_rx_handler. Avoid this by adding a parameter that is
false for direct calls.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-3-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Eliminate double read of TXFQS in tx_handler
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:18 +0000 (12:57 +0100)]
can: m_can: Eliminate double read of TXFQS in tx_handler

The TXFQS register is read first to check if the fifo is full and then
immediately again to get the putidx. This is unnecessary and adds
significant overhead if read requests are done over a slow bus, for
example SPI with tcan4x5x.

Add a variable to store the value of the register. Split the
m_can_tx_fifo_full function into two to avoid the hidden m_can_read call
if not needed.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-2-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agodt-bindings: can: renesas,rcar-canfd: Fix number of channels for R-Car V3U
Geert Uytterhoeven [Wed, 7 Dec 2022 16:50:21 +0000 (17:50 +0100)]
dt-bindings: can: renesas,rcar-canfd: Fix number of channels for R-Car V3U

According to the bindings, only two channels are supported.
However, R-Car V3U supports eight, leading to "make dtbs" failures:

        arch/arm64/boot/dts/renesas/r8a779a0-falcon.dtb: can@e6660000: Unevaluated properties are not allowed ('channel2', 'channel3', 'channel4', 'channel5', 'channel6', 'channel7' were unexpected)

Update the number of channels to 8 on R-Car V3U.
While at it, prevent adding more properties to the channel nodes, as
they must contain no other properties than a status property.

Fixes: d6254d52d70de530 ("dt-bindings: can: renesas,rcar-canfd: Document r8a779a0 support")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/all/7d41d72cd7db2e90bae069ce57dbb672f17500ae.1670431681.git.geert+renesas@glider.be
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: raw: add support for SO_MARK
Marc Kleine-Budde [Fri, 9 Dec 2022 09:10:08 +0000 (10:10 +0100)]
can: raw: add support for SO_MARK

Add support for SO_MARK to the CAN_RAW protocol. This makes it
possible to add traffic control filters based on the fwmark.

Link: https://lore.kernel.org/all/20221210113653.170346-1-mkl@pengutronix.de
Acked-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: Call the RAM init directly from m_can_chip_config
Vivek Yadav [Wed, 7 Dec 2022 10:06:31 +0000 (15:36 +0530)]
can: m_can: Call the RAM init directly from m_can_chip_config

When we try to access the mcan message ram addresses during the probe,
hclk is gated by any other drivers or disabled, because of that probe
gets failed.

Move the mram init functionality to mcan chip config called by
m_can_start from mcan open function, by that time clocks are
enabled.

Suggested-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Vivek Yadav <vivek.2311@samsung.com>
Link: https://lore.kernel.org/all/20221207100632.96200-2-vivek.2311@samsung.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agoMerge patch series "can: usb: remove pointers to struct usb_interface in device's...
Marc Kleine-Budde [Mon, 12 Dec 2022 10:41:58 +0000 (11:41 +0100)]
Merge patch series "can: usb: remove pointers to struct usb_interface in device's priv structures"

Vincent Mailhol <mailhol.vincent@wanadoo.fr> says:

The gs_can and ucan drivers keep a pointer to struct usb_interface in
their private structure. This is not needed. For gs_can the only use
is to retrieve struct usb_device, which is already available in
gs_usb::udev. For ucan, the field is set but never used.

Remove the struct usb_interface fields and clean up.

Link: https://lore.kernel.org/all/20221208081142.16936-1-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: gs_usb: remove gs_can::iface
Vincent Mailhol [Thu, 8 Dec 2022 08:11:42 +0000 (17:11 +0900)]
can: gs_usb: remove gs_can::iface

The iface field of struct gs_can is only used to retrieve the
usb_device which is already available in gs_can::udev.

Replace each occurrence of interface_to_usbdev(dev->iface) with
dev->udev. This done, remove gs_can::iface.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221208081142.16936-3-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: ucan: remove unused ucan_priv::intf
Vincent Mailhol [Thu, 8 Dec 2022 08:11:41 +0000 (17:11 +0900)]
can: ucan: remove unused ucan_priv::intf

Field intf of struct ucan_priv is set but never used. Remove it.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221208081142.16936-2-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agonet: af_can: remove useless parameter 'err' in 'can_rx_register()'
Ye Bin [Thu, 8 Dec 2022 09:09:40 +0000 (17:09 +0800)]
net: af_can: remove useless parameter 'err' in 'can_rx_register()'

Since commit bdfb5765e45b remove NULL-ptr checks from users of
can_dev_rcv_lists_find(). 'err' parameter is useless, so remove it.

Signed-off-by: Ye Bin <yebin10@huawei.com>
Link: https://lore.kernel.org/all/20221208090940.3695670-1-yebin@huaweicloud.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: ucan: use strscpy() to instead of strncpy()
Xu Panda [Wed, 7 Dec 2022 01:09:09 +0000 (09:09 +0800)]
can: ucan: use strscpy() to instead of strncpy()

The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.

Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Link: https://lore.kernel.org/all/202212070909095189693@zte.com.cn
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agoMerge patch series "can: etas_es58x: report firmware, bootloader and hardware version"
Marc Kleine-Budde [Mon, 12 Dec 2022 10:40:16 +0000 (11:40 +0100)]
Merge patch series "can: etas_es58x: report firmware, bootloader and hardware version"

Vincent Mailhol <mailhol.vincent@wanadoo.fr> says:

The goal of this series is to report the firmware version, the
bootloader version and the hardware revision of ETAS ES58x devices.

These are already reported in the kernel log but this isn't best
practice. Remove the kernel log and instead export all these through
devlink. The devlink core automatically exports the firmware and the
bootloader version to ethtool, so no need to implement the
ethtool_ops::get_drvinfo() callback anymore.

Patch one and two implement the core support for devlink (at device
level) and devlink port (at the network interface level).

Patch three export usb_cache_string() and patch four add a new info
attribute to devlink.h. Both are prerequisites for patch five.

Patch five is the actual goal: it parses the product information from
a custom usb string returned by the device and expose them through
devlink.

Patch six removes the product information from the kernel log.

Finally, patch seven add a devlink documentation page with list all
the information attributes reported by the driver.

* Sample outputs following this series *

| $ devlink dev info
| usb/1-9:1.1:
|   serial_number 0108954
|   versions:
|       fixed:
|         board.rev B012/000
|       running:
|         fw 04.00.01
|         fw.bootloader 02.00.00

| $ devlink port show can0
| usb/1-9:1.1/0: type eth netdev can0 flavour physical port 0 splittable false

| $ ethtool -i can0
| driver: etas_es58x
| version: 6.1.0-rc7+
| firmware-version: 04.00.01 02.00.00
| expansion-rom-version:
| bus-info: 1-9:1.1
| supports-statistics: no
| supports-test: no
| supports-eeprom-access: no
| supports-register-dump: no
| supports-priv-flags: no

* Changelog *

v4 -> v5: https://lore.kernel.org/all/20221130174658.29282-1-mailhol.vincent@wanadoo.fr

  * [PATH 2/7] add devlink port support. This extends devlink to the
    network interface.

  * thanks to devlink port, 'ethtool -i' is now able to retrieve the
    firmware version from devlink. No need to implement the
    ethtool_ops::get_drvinfo() callback anymore: remove one patch from
    the series.

  * [PATCH 4/7] A new patch to add a new info attribute for the
    bootloader version in devlink.h. This patch was initially sent as
    a standalone patch here:
      https://lore.kernel.org/netdev/20221129031406.3849872-1-mailhol.vincent@wanadoo.fr
    Merging it to this series so that it is both added and used at the
    same time.

  * [PATCH 5/7] use the newly info attribute defined in patch 4/7 to
    report the bootloader version instead of the custom string "bl".

  * [PATCH 5/7] because the series does not implement
    ethtool_ops::get_drvinfo() anymore, the two helper functions
    es58x_sw_version_is_set() and es58x_hw_revision_is_set() are only
    used in devlink.c. Move them from es58x_core.h to es58x_devlink.c.

  * [PATCH 5/7] small rework of the helper function
    es58x_hw_revision_is_set(): it is OK to only check the letter (if
    the letter is '\0', it will not be possible to print the next
    numbers).

  * [PATCH 5/7 and 6/7] add reviewed-by Andrew Lunn tag.

  * [PATCH 7/7] Now, 'ethtool -i' reports both the firmware version
    and the bootloader version (this is how the core export the
    information from devlink to ethtool). Update the documentation to
    reflect this fact.

  * Reoder the patches.

v3 -> v4: https://lore.kernel.org/all/20221126162211.93322-1-mailhol.vincent@wanadoo.fr

  * major rework to use devlink instead of sysfs following Andrew's
    comment.

  * split the series in 6 patches.

  * [PATCH 1/6] add Acked-by: Greg Kroah-Hartman

v2 -> v3: https://lore.kernel.org/all/20221113040108.68249-1-mailhol.vincent@wanadoo.fr

  * patch 2/3: do not spam the kernel log anymore with the product
    number. Instead parse the product information string, extract the
    firmware version, the bootloadar version and the hardware revision
    and export them through sysfs.

  * patch 2/3: rework the parsing in order not to need additional
    fields in struct es58x_parameters.

  * patch 3/3: only populate ethtool_drvinfo::fw_version because since
    commit edaf5df22cb8 ("ethtool: ethtool_get_drvinfo: populate
    drvinfo fields even if callback exits"), there is no need to
    populate ethtool_drvinfo::driver and ethtool_drvinfo::bus_info in
    the driver.

v1 -> v2: https://lore.kernel.org/all/20221104171604.24052-1-mailhol.vincent@wanadoo.fr

  * was a single patch. It is now a series of three patches.
  * add a first new patch to export  usb_cache_string().
  * add a second new patch to apply usb_cache_string() to existing code.
  * add missing check on product info string to prevent a buffer overflow.
  * add comma on the last entry of struct es58x_parameters.

v1: https://lore.kernel.org/all/20221104073659.414147-1-mailhol.vincent@wanadoo.fr

Link: https://lore.kernel.org/all/20221130174658.29282-1-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agoDocumentation: devlink: add devlink documentation for the etas_es58x driver
Vincent Mailhol [Wed, 30 Nov 2022 17:46:58 +0000 (02:46 +0900)]
Documentation: devlink: add devlink documentation for the etas_es58x driver

List all the version information reported by the etas_es58x driver
through devlink. Also, update MAINTAINERS with the newly created file.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-8-mailhol.vincent@wanadoo.fr
[mkl: fixed version information table: "bl" -> "fw.bootloader"
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: etas_es58x: remove es58x_get_product_info()
Vincent Mailhol [Wed, 30 Nov 2022 17:46:57 +0000 (02:46 +0900)]
can: etas_es58x: remove es58x_get_product_info()

Now that the product information are available under devlink, no more
need to print them in the kernel log. Remove es58x_get_product_info().

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20221130174658.29282-7-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: etas_es58x: export product information through devlink_ops::info_get()
Vincent Mailhol [Wed, 30 Nov 2022 17:46:56 +0000 (02:46 +0900)]
can: etas_es58x: export product information through devlink_ops::info_get()

ES58x devices report below product information through a custom usb
string:

  * the firmware version
  * the bootloader version
  * the hardware revision

Parse this string, store the results in struct es58x_dev, export:

  * the firmware version through devlink's "fw" name
  * the bootloader version through devlink's "fw.bootloader" name
  * the hardware revisionthrough devlink's "board.rev" name

Those devlink entries are not critical to use the device, if parsing
fails, print an informative log message and continue to probe the
device.

In addition to that, use usb_device::serial to report the device
serial number.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20221130174658.29282-6-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agonet: devlink: add DEVLINK_INFO_VERSION_GENERIC_FW_BOOTLOADER
Vincent Mailhol [Wed, 30 Nov 2022 17:46:55 +0000 (02:46 +0900)]
net: devlink: add DEVLINK_INFO_VERSION_GENERIC_FW_BOOTLOADER

As discussed in [1], abbreviating the bootloader to "bl" might not be
well understood. Instead, a bootloader technically being a firmware,
name it "fw.bootloader".

Add a new macro to devlink.h to formalize this new info attribute name
and update the documentation.

[1] https://lore.kernel.org/netdev/20221128142723.2f826d20@kernel.org/

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-5-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: c_can: use devm_platform_get_and_ioremap_resource()
Minghao Chi [Fri, 11 Nov 2022 06:43:00 +0000 (14:43 +0800)]
can: c_can: use devm_platform_get_and_ioremap_resource()

Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Link: https://lore.kernel.org/all/202211111443005202576@zte.com.cn
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agoUSB: core: export usb_cache_string()
Vincent Mailhol [Wed, 30 Nov 2022 17:46:54 +0000 (02:46 +0900)]
USB: core: export usb_cache_string()

usb_cache_string() can also be useful for the drivers so export it.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/all/20221130174658.29282-4-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agodt-bindings: can: renesas,rcar-canfd: Document RZ/Five SoC
Lad Prabhakar [Tue, 15 Nov 2022 12:38:11 +0000 (12:38 +0000)]
dt-bindings: can: renesas,rcar-canfd: Document RZ/Five SoC

The CANFD block on the RZ/Five SoC is identical to one found on the
RZ/G2UL SoC. "renesas,r9a07g043-canfd" compatible string will be used
on the RZ/Five SoC so to make this clear, update the comment to include
RZ/Five SoC.

No driver changes are required as generic compatible string
"renesas,rzg2l-canfd" will be used as a fallback on RZ/Five SoC.

Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/20221115123811.1182922-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: etas_es58x: add devlink port support
Vincent Mailhol [Wed, 30 Nov 2022 17:46:53 +0000 (02:46 +0900)]
can: etas_es58x: add devlink port support

Add support for devlink port which extends the devlink support to the
network interface level. For now, the etas_es58x driver will only rely
on the default features that devlink port has to offer and not
implement additional feature ones.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-3-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agodt-bindings: can: fsl,flexcan: add imx93 compatible
Haibo Chen [Tue, 22 Nov 2022 11:32:31 +0000 (19:32 +0800)]
dt-bindings: can: fsl,flexcan: add imx93 compatible

Add a new compatible string for imx93.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/all/1669116752-4260-2-git-send-email-haibo.chen@nxp.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: etas_es58x: add devlink support
Vincent Mailhol [Wed, 30 Nov 2022 17:46:52 +0000 (02:46 +0900)]
can: etas_es58x: add devlink support

Add basic support for devlink at the device level. The callbacks of
struct devlink_ops will be implemented next.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-2-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: flexcan: add auto stop mode for IMX93 to support wakeup
Haibo Chen [Tue, 22 Nov 2022 11:32:30 +0000 (19:32 +0800)]
can: flexcan: add auto stop mode for IMX93 to support wakeup

IMX93 do not contain a GPR to config the stop mode, it will set
the flexcan into stop mode automatically once the ARM core go
into low power mode (WFI instruct) and gate off the flexcan
related clock automatically. But to let these logic work as
expect, before ARM core go into low power mode, need to make
sure the flexcan related clock keep on.

To support stop mode and wakeup feature on imx93, this patch
add a new fsl_imx93_devtype_data to separate from imx8mp.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Link: https://lore.kernel.org/all/1669116752-4260-1-git-send-email-haibo.chen@nxp.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: etas_es58x: sort the includes by alphabetic order
Vincent Mailhol [Sat, 26 Nov 2022 16:05:25 +0000 (01:05 +0900)]
can: etas_es58x: sort the includes by alphabetic order

Follow the best practices, reorder the includes.

While doing so, bump up copyright year of each modified files.

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221126160525.87036-1-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: ctucanfd: Drop obsolete dependency on COMPILE_TEST
Jean Delvare [Thu, 24 Nov 2022 13:16:04 +0000 (14:16 +0100)]
can: ctucanfd: Drop obsolete dependency on COMPILE_TEST

Since commit 0166dc11be91 ("of: make CONFIG_OF user selectable"), it
is possible to test-build any driver which depends on OF on any
architecture by explicitly selecting OF. Therefore depending on
COMPILE_TEST as an alternative is no longer needed.

It is actually better to always build such drivers with OF enabled,
so that the test builds are closer to how each driver will actually be
built on its intended target. Building them without OF may not test
much as the compiler will optimize out potentially large parts of the
code. In the worst case, this could even pop false positive warnings.
Dropping COMPILE_TEST here improves the quality of our testing and
avoids wasting time on non-existent issues.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Cc: Ondrej Ille <ondrej.ille@gmail.com>
Acked-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Link: https://lore.kernel.org/all/20221124141604.4265225f@endymion.delvare
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agoMerge patch series "R-Car CAN FD driver enhancements"
Marc Kleine-Budde [Mon, 12 Dec 2022 10:37:56 +0000 (11:37 +0100)]
Merge patch series "R-Car CAN FD driver enhancements"

Biju Das <biju.das.jz@bp.renesas.com> says:

The CAN FD IP found on RZ/G2L SoC has some HW features different to
that of R-Car. For example, it has multiple resets, dedicated channel
tx and error interrupts, separate global rx and error interrupts
compared to shared irq for R-Car. it does not s ECC error flag
registers and clk post divider present on R-Car.

Similarly, R-Car V3U has 8 channels whereas other SoCs has only 2
channels. Currently all the HW differences are handled by comparing
with chip_id enum.

This patch series aims to replace chip_id with struct
rcar_canfd_hw_info to handle the HW feature differences and driver
data present on both IPs.

The changes are trivial and tested on RZ/G2L SMARC EVK.

This patch series depend upon [1].

[1] https://lore.kernel.org/all/20221025155657.1426948-1-biju.das.jz@bp.renesas.com

changes since v2: https://lore.kernel.org/all/20221026131732.1843105-1-biju.das.jz@bp.renesas.com
 * Replaced data type of max_channels from unsigned int->u8 to save memory.
 * Replaced data type of postdiv from unsigned int->u8 to save memory.

changes since v1: https://lore.kernel.org/all/20221022104357.1276740-1-biju.das.jz@bp.renesas.com
 * Updated commit description for R-Car V3U SoC detection using
   driver data.
 * Replaced data type of max_channels from u32->unsigned int.
 * Replaced multi_global_irqs->shared_global_irqs to make it
   positive checks.
 * Replaced clk_postdiv->postdiv driver data variable.
 * Simplified the calcualtion for fcan_freq.
 * Replaced info->has_gerfl to gpriv->info->has_gerfl and wrapped
   the ECC error flag checks inside a single if statement.
 * Added Rb tag from Geert patch#1,#2,#3 and #5

Link: https://lore.kernel.org/all/20221027082158.95895-1-biju.das.jz@bp.renesas.com
[mkl: only take patches 1...5 to avoid merge conflict]
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agoMerge branch 'ovs-tc-dedup'
David S. Miller [Mon, 12 Dec 2022 10:14:03 +0000 (10:14 +0000)]
Merge branch 'ovs-tc-dedup'

Xin Long says:

====================
net: eliminate the duplicate code in the ct nat functions of ovs and tc

The changes in the patchset:

  "net: add helper support in tc act_ct for ovs offloading"

had moved some common ct code used by both OVS and TC into netfilter.

There are still some big functions pretty similar defined and used in
each of OVS and TC. It is not good to maintain such big function in 2
places. This patchset is to extract the functions for NAT processing
from OVS and TC to netfilter.

To make this change clear and safe, this patchset gets the common code
out of OVS and TC step by step: The patch 1-4 make some minor changes
in OVS and TC to make the NAT code of them completely the same, then
the patch 5 moves the common code to the netfilter and exports one
function called by each of OVS and TC.

v1->v2:
  - Create nf_nat_ovs.c to include the nat functions, as Pablo suggested.
v2->v3:
  - fix a typo in subject of patch 2/5, as Marcelo noticed.
  - fix in openvswitch to keep OVS ct nat and TC ct nat consistent in
    patch 3/5 instead of in tc, as Marcelo noticed.
  - use BIT(var) macro instead of (1 << var) in patch 5/5, as Marcelo
    suggested.
  - use ifdef in netfilter/Makefile to build nf_nat_ovs only when OVS
    or TC ct action is enabled in patch 5/5, as Marcelo suggested.
v3->v4:
  - add NF_NAT_OVS in netfilter/Kconfig and add select NF_NAT_OVS in
    OVS and TC Kconfig instead of using ifdef in netfilter/Makefile,
    as Pablo suggested.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: move the nat function to nf_nat_ovs for ovs and tc
Xin Long [Thu, 8 Dec 2022 16:56:12 +0000 (11:56 -0500)]
net: move the nat function to nf_nat_ovs for ovs and tc

There are two nat functions are nearly the same in both OVS and
TC code, (ovs_)ct_nat_execute() and ovs_ct_nat/tcf_ct_act_nat().

This patch creates nf_nat_ovs.c under netfilter and moves them
there then exports nf_ct_nat() so that it can be shared by both
OVS and TC, and keeps the nat (type) check and nat flag update
in OVS and TC's own place, as these parts are different between
OVS and TC.

Note that in OVS nat function it was using skb->protocol to get
the proto as it already skips vlans in key_extract(), while it
doesn't in TC, and TC has to call skb_protocol() to get proto.
So in nf_ct_nat_execute(), we keep using skb_protocol() which
works for both OVS and TC contrack.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Aaron Conole <aconole@redhat.com>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: sched: update the nat flag for icmp error packets in ct_nat_execute
Xin Long [Thu, 8 Dec 2022 16:56:11 +0000 (11:56 -0500)]
net: sched: update the nat flag for icmp error packets in ct_nat_execute

In ovs_ct_nat_execute(), the packet flow key nat flags are updated
when it processes ICMP(v6) error packets translation successfully.

In ct_nat_execute() when processing ICMP(v6) error packets translation
successfully, it should have done the same in ct_nat_execute() to set
post_ct_s/dnat flag, which will be used to update flow key nat flags
in OVS module later.

Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoopenvswitch: return NF_DROP when fails to add nat ext in ovs_ct_nat
Xin Long [Thu, 8 Dec 2022 16:56:10 +0000 (11:56 -0500)]
openvswitch: return NF_DROP when fails to add nat ext in ovs_ct_nat

When it fails to allocate nat ext, the packet should be dropped, like
the memory allocation failures in other places in ovs_ct_nat().

This patch changes to return NF_DROP when fails to add nat ext before
doing NAT in ovs_ct_nat(), also it would keep consistent with tc
action ct' processing in tcf_ct_act_nat().

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoopenvswitch: return NF_ACCEPT when OVS_CT_NAT is not set in info nat
Xin Long [Thu, 8 Dec 2022 16:56:09 +0000 (11:56 -0500)]
openvswitch: return NF_ACCEPT when OVS_CT_NAT is not set in info nat

Either OVS_CT_SRC_NAT or OVS_CT_DST_NAT is set, OVS_CT_NAT must be
set in info->nat. Thus, if OVS_CT_NAT is not set in info->nat, it
will definitely not do NAT but returns NF_ACCEPT in ovs_ct_nat().

This patch changes nothing funcational but only makes this return
earlier in ovs_ct_nat() to keep consistent with TC's processing
in tcf_ct_act_nat().

Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoopenvswitch: delete the unncessary skb_pull_rcsum call in ovs_ct_nat_execute
Xin Long [Thu, 8 Dec 2022 16:56:08 +0000 (11:56 -0500)]
openvswitch: delete the unncessary skb_pull_rcsum call in ovs_ct_nat_execute

The calls to ovs_ct_nat_execute() are as below:

  ovs_ct_execute()
    ovs_ct_lookup()
      __ovs_ct_lookup()
        ovs_ct_nat()
          ovs_ct_nat_execute()
    ovs_ct_commit()
      __ovs_ct_lookup()
        ovs_ct_nat()
          ovs_ct_nat_execute()

and since skb_pull_rcsum() and skb_push_rcsum() are already
called in ovs_ct_execute(), there's no need to do it again
in ovs_ct_nat_execute().

Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agomyri10ge: use strscpy() to instead of strncpy()
Xu Panda [Fri, 9 Dec 2022 07:37:29 +0000 (15:37 +0800)]
myri10ge: use strscpy() to instead of strncpy()

The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.

Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoliquidio: use strscpy() to instead of strncpy()
Xu Panda [Fri, 9 Dec 2022 07:34:49 +0000 (15:34 +0800)]
liquidio: use strscpy() to instead of strncpy()

The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.

Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agohns: use strscpy() to instead of strncpy()
Xu Panda [Fri, 9 Dec 2022 07:33:25 +0000 (15:33 +0800)]
hns: use strscpy() to instead of strncpy()

The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.

Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: stmmac: Add check for taprio basetime configuration
Michael Sit Wei Hong [Thu, 8 Dec 2022 09:03:15 +0000 (17:03 +0800)]
net: stmmac: Add check for taprio basetime configuration

Adds a boundary check to prevent negative basetime input from user
while configuring taprio.

Signed-off-by: Michael Sit Wei Hong <michael.wei.hong.sit@intel.com>
Signed-off-by: Lai Peter Jun Ann <jun.ann.lai@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoMerge branch 'tun-vnet-uso'
David S. Miller [Mon, 12 Dec 2022 09:29:57 +0000 (09:29 +0000)]
Merge branch 'tun-vnet-uso'

Andrew Melnychenko says:

====================
TUN/VirtioNet USO features support.

Added new offloads for TUN devices TUN_F_USO4 and TUN_F_USO6.
Technically they enable NETIF_F_GSO_UDP_L4
(and only if USO4 & USO6 are set simultaneously).
It allows the transmission of large UDP packets.

UDP Segmentation Offload (USO/GSO_UDP_L4) - ability to split UDP packets
into several segments. It's similar to UFO, except it doesn't use IP
fragmentation. The drivers may push big packets and the NIC will split
them(or assemble them in case of receive), but in the case of VirtioNet
we just pass big UDP to the host. So we are freeing the driver from doing
the unnecessary job of splitting. The same thing for several guests
on one host, we can pass big packets between guests.

Different features USO4 and USO6 are required for qemu where Windows
guests can enable disable USO receives for IPv4 and IPv6 separately.
On the other side, Linux can't really differentiate USO4 and USO6, for now.
For now, to enable USO for TUN it requires enabling USO4 and USO6 together.
In the future, there would be a mechanism to control UDP_L4 GSO separately.

New types for virtio-net already in virtio-net specification:
https://github.com/oasis-tcs/virtio-spec/issues/120

Test it WIP Qemu https://github.com/daynix/qemu/tree/USOv3

Changes since v4 & RFC:
 * Fixed typo and refactored.
 * Tun USO offload refactored.
 * Add support for guest-to-guest segmentation offload (thx Jason).
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agodrivers/net/virtio_net.c: Added USO support.
Andrew Melnychenko [Wed, 7 Dec 2022 11:35:58 +0000 (13:35 +0200)]
drivers/net/virtio_net.c: Added USO support.

Now, it possible to enable GSO_UDP_L4("tx-udp-segmentation") for VirtioNet.

Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agolinux/virtio_net.h: Support USO offload in vnet header.
Andrew Melnychenko [Wed, 7 Dec 2022 11:35:57 +0000 (13:35 +0200)]
linux/virtio_net.h: Support USO offload in vnet header.

Now, it's possible to convert USO vnet packets from/to skb.
Added support for GSO_UDP_L4 offload.

Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agouapi/linux/virtio_net.h: Added USO types.
Andrew Melnychenko [Wed, 7 Dec 2022 11:35:56 +0000 (13:35 +0200)]
uapi/linux/virtio_net.h: Added USO types.

Added new GSO type for USO: VIRTIO_NET_HDR_GSO_UDP_L4.
Feature VIRTIO_NET_F_HOST_USO allows to enable NETIF_F_GSO_UDP_L4.
Separated VIRTIO_NET_F_GUEST_USO4 & VIRTIO_NET_F_GUEST_USO6 features
required for Windows guests.

Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agodriver/net/tun: Added features for USO.
Andrew Melnychenko [Wed, 7 Dec 2022 11:35:55 +0000 (13:35 +0200)]
driver/net/tun: Added features for USO.

Added support for USO4 and USO6.
For now, to "enable" USO, it's required to set both USO4 and USO6 simultaneously.
USO enables NETIF_F_GSO_UDP_L4.

Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agouapi/linux/if_tun.h: Added new offload types for USO4/6.
Andrew Melnychenko [Wed, 7 Dec 2022 11:35:54 +0000 (13:35 +0200)]
uapi/linux/if_tun.h: Added new offload types for USO4/6.

Added 2 additional offlloads for USO(IPv4 & IPv6).
Separate offloads are required for Windows VM guests,
g.e. Windows may set USO rx only for IPv4.

Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoudp: allow header check for dodgy GSO_UDP_L4 packets.
Andrew Melnychenko [Wed, 7 Dec 2022 11:35:53 +0000 (13:35 +0200)]
udp: allow header check for dodgy GSO_UDP_L4 packets.

Allow UDP_L4 for robust packets.

Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoMerge branch 'stricter register ID checking in regsafe()'
Alexei Starovoitov [Sat, 10 Dec 2022 21:20:53 +0000 (13:20 -0800)]
Merge branch 'stricter register ID checking in regsafe()'

Eduard Zingerman says:

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

This patch-set consists of a series of bug fixes for register ID
tracking in verifier.c:states_equal()/regsafe() functions:
 - for registers of type PTR_TO_MAP_{KEY,VALUE}, PTR_TO_PACKET[_META]
   the regsafe() should call check_ids() even if registers are
   byte-to-byte equal;
 - states_equal() must maintain idmap that covers all function frames
   in the state because functions like mark_ptr_or_null_regs() operate
   on all registers in the state;
 - regsafe() must compare spin lock ids for PTR_TO_MAP_VALUE registers.

The last point covers issue reported by Kumar Kartikeya Dwivedi in [1],
I borrowed the test commit from there.
Note, that there is also an issue with register id tracking for
scalars described here [2], it would be addressed separately.

[1] https://lore.kernel.org/bpf/20221111202719.982118-1-memxor@gmail.com/
[2] https://lore.kernel.org/bpf/20221128163442.280187-2-eddyz87@gmail.com/

Eduard Zingerman (6):
  bpf: regsafe() must not skip check_ids()
  selftests/bpf: test cases for regsafe() bug skipping check_id()
  bpf: states_equal() must build idmap for all function frames
  selftests/bpf: verify states_equal() maintains idmap across all frames
  bpf: use check_ids() for active_lock comparison
  selftests/bpf: test case for relaxed prunning of active_lock.id
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agoselftests/bpf: test case for relaxed prunning of active_lock.id
Eduard Zingerman [Fri, 9 Dec 2022 13:57:33 +0000 (15:57 +0200)]
selftests/bpf: test case for relaxed prunning of active_lock.id

Check that verifier.c:states_equal() uses check_ids() to match
consistent active_lock/map_value configurations. This allows to prune
states with active spin locks even if numerical values of
active_lock ids do not match across compared states.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-8-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agoselftests/bpf: Add pruning test case for bpf_spin_lock
Kumar Kartikeya Dwivedi [Fri, 9 Dec 2022 13:57:32 +0000 (15:57 +0200)]
selftests/bpf: Add pruning test case for bpf_spin_lock

Test that when reg->id is not same for the same register of type
PTR_TO_MAP_VALUE between current and old explored state, we currently
return false from regsafe and continue exploring.

Without the fix in prior commit, the test case fails.

Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-7-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agobpf: use check_ids() for active_lock comparison
Eduard Zingerman [Fri, 9 Dec 2022 13:57:31 +0000 (15:57 +0200)]
bpf: use check_ids() for active_lock comparison

An update for verifier.c:states_equal()/regsafe() to use check_ids()
for active spin lock comparisons. This fixes the issue reported by
Kumar Kartikeya Dwivedi in [1] using technique suggested by Edward Cree.

W/o this commit the verifier might be tricked to accept the following
program working with a map containing spin locks:

  0: r9 = map_lookup_elem(...)  ; Returns PTR_TO_MAP_VALUE_OR_NULL id=1.
  1: r8 = map_lookup_elem(...)  ; Returns PTR_TO_MAP_VALUE_OR_NULL id=2.
  2: if r9 == 0 goto exit       ; r9 -> PTR_TO_MAP_VALUE.
  3: if r8 == 0 goto exit       ; r8 -> PTR_TO_MAP_VALUE.
  4: r7 = ktime_get_ns()        ; Unbound SCALAR_VALUE.
  5: r6 = ktime_get_ns()        ; Unbound SCALAR_VALUE.
  6: bpf_spin_lock(r8)          ; active_lock.id == 2.
  7: if r6 > r7 goto +1         ; No new information about the state
                                ; is derived from this check, thus
                                ; produced verifier states differ only
                                ; in 'insn_idx'.
  8: r9 = r8                    ; Optionally make r9.id == r8.id.
  --- checkpoint ---            ; Assume is_state_visisted() creates a
                                ; checkpoint here.
  9: bpf_spin_unlock(r9)        ; (a,b) active_lock.id == 2.
                                ; (a) r9.id == 2, (b) r9.id == 1.
 10: exit(0)

Consider two verification paths:
(a) 0-10
(b) 0-7,9-10

The path (a) is verified first. If checkpoint is created at (8)
the (b) would assume that (8) is safe because regsafe() does not
compare register ids for registers of type PTR_TO_MAP_VALUE.

[1] https://lore.kernel.org/bpf/20221111202719.982118-1-memxor@gmail.com/

Reported-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Suggested-by: Edward Cree <ecree.xilinx@gmail.com>
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-6-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agoselftests/bpf: verify states_equal() maintains idmap across all frames
Eduard Zingerman [Fri, 9 Dec 2022 13:57:30 +0000 (15:57 +0200)]
selftests/bpf: verify states_equal() maintains idmap across all frames

A test case that would erroneously pass verification if
verifier.c:states_equal() maintains separate register ID mappings for
call frames.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-5-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agobpf: states_equal() must build idmap for all function frames
Eduard Zingerman [Fri, 9 Dec 2022 13:57:29 +0000 (15:57 +0200)]
bpf: states_equal() must build idmap for all function frames

verifier.c:states_equal() must maintain register ID mapping across all
function frames. Otherwise the following example might be erroneously
marked as safe:

main:
    fp[-24] = map_lookup_elem(...)  ; frame[0].fp[-24].id == 1
    fp[-32] = map_lookup_elem(...)  ; frame[0].fp[-32].id == 2
    r1 = &fp[-24]
    r2 = &fp[-32]
    call foo()
    r0 = 0
    exit

foo:
  0: r9 = r1
  1: r8 = r2
  2: r7 = ktime_get_ns()
  3: r6 = ktime_get_ns()
  4: if (r6 > r7) goto skip_assign
  5: r9 = r8

skip_assign:                ; <--- checkpoint
  6: r9 = *r9               ; (a) frame[1].r9.id == 2
                            ; (b) frame[1].r9.id == 1

  7: if r9 == 0 goto exit:  ; mark_ptr_or_null_regs() transfers != 0 info
                            ; for all regs sharing ID:
                            ;   (a) r9 != 0 => &frame[0].fp[-32] != 0
                            ;   (b) r9 != 0 => &frame[0].fp[-24] != 0

  8: r8 = *r8               ; (a) r8 == &frame[0].fp[-32]
                            ; (b) r8 == &frame[0].fp[-32]
  9: r0 = *r8               ; (a) safe
                            ; (b) unsafe

exit:
 10: exit

While processing call to foo() verifier considers the following
execution paths:

(a) 0-10
(b) 0-4,6-10
(There is also path 0-7,10 but it is not interesting for the issue at
 hand. (a) is verified first.)

Suppose that checkpoint is created at (6) when path (a) is verified,
next path (b) is verified and (6) is reached.

If states_equal() maintains separate 'idmap' for each frame the
mapping at (6) for frame[1] would be empty and
regsafe(r9)::check_ids() would add a pair 2->1 and return true,
which is an error.

If states_equal() maintains single 'idmap' for all frames the mapping
at (6) would be { 1->1, 2->2 } and regsafe(r9)::check_ids() would
return false when trying to add a pair 2->1.

This issue was suggested in the following discussion:
https://lore.kernel.org/bpf/CAEf4BzbFB5g4oUfyxk9rHy-PJSLQ3h8q9mV=rVoXfr_JVm8+1Q@mail.gmail.com/

Suggested-by: Andrii Nakryiko <andrii.nakryiko@gmail.com>
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-4-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agoselftests/bpf: test cases for regsafe() bug skipping check_id()
Eduard Zingerman [Fri, 9 Dec 2022 13:57:28 +0000 (15:57 +0200)]
selftests/bpf: test cases for regsafe() bug skipping check_id()

Under certain conditions it was possible for verifier.c:regsafe() to
skip check_id() call. This commit adds negative test cases previously
errorneously accepted as safe.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-3-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agobpf: regsafe() must not skip check_ids()
Eduard Zingerman [Fri, 9 Dec 2022 13:57:27 +0000 (15:57 +0200)]
bpf: regsafe() must not skip check_ids()

The verifier.c:regsafe() has the following shortcut:

equal = memcmp(rold, rcur, offsetof(struct bpf_reg_state, parent)) == 0;
...
if (equal)
return true;

Which is executed regardless old register type. This is incorrect for
register types that might have an ID checked by check_ids(), namely:
 - PTR_TO_MAP_KEY
 - PTR_TO_MAP_VALUE
 - PTR_TO_PACKET_META
 - PTR_TO_PACKET

The following pattern could be used to exploit this:

  0: r9 = map_lookup_elem(...)  ; Returns PTR_TO_MAP_VALUE_OR_NULL id=1.
  1: r8 = map_lookup_elem(...)  ; Returns PTR_TO_MAP_VALUE_OR_NULL id=2.
  2: r7 = ktime_get_ns()        ; Unbound SCALAR_VALUE.
  3: r6 = ktime_get_ns()        ; Unbound SCALAR_VALUE.
  4: if r6 > r7 goto +1         ; No new information about the state
                                ; is derived from this check, thus
                                ; produced verifier states differ only
                                ; in 'insn_idx'.
  5: r9 = r8                    ; Optionally make r9.id == r8.id.
  --- checkpoint ---            ; Assume is_state_visisted() creates a
                                ; checkpoint here.
  6: if r9 == 0 goto <exit>     ; Nullness info is propagated to all
                                ; registers with matching ID.
  7: r1 = *(u64 *) r8           ; Not always safe.

Verifier first visits path 1-7 where r8 is verified to be not null
at (6). Later the jump from 4 to 6 is examined. The checkpoint for (6)
looks as follows:
  R8_rD=map_value_or_null(id=2,off=0,ks=4,vs=8,imm=0)
  R9_rwD=map_value_or_null(id=2,off=0,ks=4,vs=8,imm=0)
  R10=fp0

The current state is:
  R0=... R6=... R7=... fp-8=...
  R8=map_value_or_null(id=2,off=0,ks=4,vs=8,imm=0)
  R9=map_value_or_null(id=1,off=0,ks=4,vs=8,imm=0)
  R10=fp0

Note that R8 states are byte-to-byte identical, so regsafe() would
exit early and skip call to check_ids(), thus ID mapping 2->2 will not
be added to 'idmap'. Next, states for R9 are compared: these are not
identical and check_ids() is executed, but 'idmap' is empty, so
check_ids() adds mapping 2->1 to 'idmap' and returns success.

This commit pushes the 'equal' down to register types that don't need
check_ids().

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-2-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agocan: rcar_canfd: Add multi_channel_irqs to struct rcar_canfd_hw_info
Biju Das [Thu, 27 Oct 2022 08:21:57 +0000 (09:21 +0100)]
can: rcar_canfd: Add multi_channel_irqs to struct rcar_canfd_hw_info

RZ/G2L has separate IRQ lines for tx and error interrupt for each
channel whereas R-Car has a combined IRQ line for all the channel
specific tx and error interrupts.

Add multi_channel_irqs to struct rcar_canfd_hw_info to select the
driver to choose between combined and separate irq registration for
channel interrupts. This patch also removes enum rcanfd_chip_id and
chip_id from both struct rcar_canfd_hw_info, as it is unused.

Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/20221027082158.95895-6-biju.das.jz@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: rcar_canfd: Add postdiv to struct rcar_canfd_hw_info
Biju Das [Thu, 27 Oct 2022 08:21:56 +0000 (09:21 +0100)]
can: rcar_canfd: Add postdiv to struct rcar_canfd_hw_info

R-Car has a clock divider for CAN FD clock within the IP, whereas
it is not available on RZ/G2L.

Add postdiv variable to struct rcar_canfd_hw_info to take care of this
difference.

Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/20221027082158.95895-5-biju.das.jz@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: rcar_canfd: Add shared_global_irqs to struct rcar_canfd_hw_info
Biju Das [Thu, 27 Oct 2022 08:21:55 +0000 (09:21 +0100)]
can: rcar_canfd: Add shared_global_irqs to struct rcar_canfd_hw_info

RZ/G2L has separate IRQ lines for receive FIFO and global error interrupt
whereas R-Car has shared IRQ line.

Add shared_global_irqs to struct rcar_canfd_hw_info to select the driver to
choose between shared and separate irq registration for global
interrupts.

Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/20221027082158.95895-4-biju.das.jz@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: rcar_canfd: Add max_channels to struct rcar_canfd_hw_info
Biju Das [Thu, 27 Oct 2022 08:21:54 +0000 (09:21 +0100)]
can: rcar_canfd: Add max_channels to struct rcar_canfd_hw_info

R-Car V3U supports a maximum of 8 channels whereas rest of the SoCs
support 2 channels.

Add max_channels variable to struct rcar_canfd_hw_info to handle this
difference.

Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/20221027082158.95895-3-biju.das.jz@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: m_can: sort header inclusion alphabetically
Vivek Yadav [Fri, 4 Nov 2022 05:16:17 +0000 (10:46 +0530)]
can: m_can: sort header inclusion alphabetically

Sort header inclusion alphabetically.

Suggested-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Vivek Yadav <vivek.2311@samsung.com>
Link: https://lore.kernel.org/all/20221104051617.21173-1-vivek.2311@samsung.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: rcar_canfd: rcar_canfd_probe: Add struct rcar_canfd_hw_info to driver data
Biju Das [Thu, 27 Oct 2022 08:21:53 +0000 (09:21 +0100)]
can: rcar_canfd: rcar_canfd_probe: Add struct rcar_canfd_hw_info to driver data

The CAN FD IP found on RZ/G2L SoC has some HW features different to that
of R-Car. For example, it has multiple resets and multiple IRQs for global
and channel interrupts. Also, it does not have ECC error flag registers
and clk post divider present on R-Car. Similarly, R-Car V3U has 8 channels
whereas other SoCs has only 2 channels.

This patch adds the struct rcar_canfd_hw_info to take care of the
HW feature differences and driver data present on both IPs. It also
replaces the driver data chip type with struct rcar_canfd_hw_info by
moving chip type to it.

Whilst started using driver data instead of chip_id for detecting
R-Car V3U SoCs.

Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/20221027082158.95895-2-biju.das.jz@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: kvaser_usb: kvaser_usb_set_{,data}bittiming(): remove empty lines in variable...
Marc Kleine-Budde [Mon, 31 Oct 2022 11:28:08 +0000 (12:28 +0100)]
can: kvaser_usb: kvaser_usb_set_{,data}bittiming(): remove empty lines in variable declaration

Fix coding style by removing empty lines in variable declaration.

Fixes: 39d3df6b0ea8 ("can: kvaser_usb: Compare requested bittiming parameters with actual parameters in do_set_{,data}_bittiming")
Cc: Jimmy Assarsson <extja@kvaser.com>
Cc: Anssi Hannula <anssi.hannula@bitwise.fi>
Link: https://lore.kernel.org/all/20221031114513.81214-2-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agocan: kvaser_usb: kvaser_usb_set_bittiming(): fix redundant initialization warning...
Marc Kleine-Budde [Mon, 31 Oct 2022 11:26:24 +0000 (12:26 +0100)]
can: kvaser_usb: kvaser_usb_set_bittiming(): fix redundant initialization warning for err

The variable err is initialized, but the initialized value is
Overwritten before it is read. Fix the warning by not initializing the
variable err at all.

Fixes: 39d3df6b0ea8 ("can: kvaser_usb: Compare requested bittiming parameters with actual parameters in do_set_{,data}_bittiming")
Cc: Jimmy Assarsson <extja@kvaser.com>
Cc: Anssi Hannula <anssi.hannula@bitwise.fi>
Link: https://lore.kernel.org/all/20221031114513.81214-1-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
22 months agoMerge tag 'ipsec-next-2022-12-09' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Sat, 10 Dec 2022 04:06:34 +0000 (20:06 -0800)]
Merge tag 'ipsec-next-2022-12-09' of git://git./linux/kernel/git/klassert/ipsec-next

Steffen Klassert says:

====================
ipsec-next 2022-12-09

1) Add xfrm packet offload core API.
   From Leon Romanovsky.

2) Add xfrm packet offload support for mlx5.
   From Leon Romanovsky and Raed Salem.

3) Fix a typto in a error message.
   From Colin Ian King.

* tag 'ipsec-next-2022-12-09' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next: (38 commits)
  xfrm: Fix spelling mistake "oflload" -> "offload"
  net/mlx5e: Open mlx5 driver to accept IPsec packet offload
  net/mlx5e: Handle ESN update events
  net/mlx5e: Handle hardware IPsec limits events
  net/mlx5e: Update IPsec soft and hard limits
  net/mlx5e: Store all XFRM SAs in Xarray
  net/mlx5e: Provide intermediate pointer to access IPsec struct
  net/mlx5e: Skip IPsec encryption for TX path without matching policy
  net/mlx5e: Add statistics for Rx/Tx IPsec offloaded flows
  net/mlx5e: Improve IPsec flow steering autogroup
  net/mlx5e: Configure IPsec packet offload flow steering
  net/mlx5e: Use same coding pattern for Rx and Tx flows
  net/mlx5e: Add XFRM policy offload logic
  net/mlx5e: Create IPsec policy offload tables
  net/mlx5e: Generalize creation of default IPsec miss group and rule
  net/mlx5e: Group IPsec miss handles into separate struct
  net/mlx5e: Make clear what IPsec rx_err does
  net/mlx5e: Flatten the IPsec RX add rule path
  net/mlx5e: Refactor FTE setup code to be more clear
  net/mlx5e: Move IPsec flow table creation to separate function
  ...
====================

Link: https://lore.kernel.org/r/20221209093310.4018731-1-steffen.klassert@secunet.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: devlink: Add missing error check to devlink_resource_put()
Gavrilov Ilia [Thu, 8 Dec 2022 08:31:39 +0000 (08:31 +0000)]
net: devlink: Add missing error check to devlink_resource_put()

When the resource size changes, the return value of the
'nla_put_u64_64bit' function is not checked. That has been fixed to avoid
rechecking at the next step.

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

Note that this is harmless, we'd error out at the next put().

Signed-off-by: Ilia.Gavrilov <Ilia.Gavrilov@infotecs.ru>
Link: https://lore.kernel.org/r/20221208082821.3927937-1-Ilia.Gavrilov@infotecs.ru
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoskbuff: Introduce slab_build_skb()
Kees Cook [Thu, 8 Dec 2022 06:02:59 +0000 (22:02 -0800)]
skbuff: Introduce slab_build_skb()

syzkaller reported:

  BUG: KASAN: slab-out-of-bounds in __build_skb_around+0x235/0x340 net/core/skbuff.c:294
  Write of size 32 at addr ffff88802aa172c0 by task syz-executor413/5295

For bpf_prog_test_run_skb(), which uses a kmalloc()ed buffer passed to
build_skb().

When build_skb() is passed a frag_size of 0, it means the buffer came
from kmalloc. In these cases, ksize() is used to find its actual size,
but since the allocation may not have been made to that size, actually
perform the krealloc() call so that all the associated buffer size
checking will be correctly notified (and use the "new" pointer so that
compiler hinting works correctly). Split this logic out into a new
interface, slab_build_skb(), but leave the original 0 checking for now
to catch any stragglers.

Reported-by: syzbot+fda18eaa8c12534ccb3b@syzkaller.appspotmail.com
Link: https://groups.google.com/g/syzkaller-bugs/c/UnIKxTtU5-0/m/-wbXinkgAQAJ
Fixes: 38931d8989b5 ("mm: Make ksize() a reporting-only function")
Cc: Pavel Begunkov <asml.silence@gmail.com>
Cc: pepsipu <soopthegoop@gmail.com>
Cc: syzbot+fda18eaa8c12534ccb3b@syzkaller.appspotmail.com
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: kasan-dev <kasan-dev@googlegroups.com>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: ast@kernel.org
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Hao Luo <haoluo@google.com>
Cc: Jesper Dangaard Brouer <hawk@kernel.org>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: jolsa@kernel.org
Cc: KP Singh <kpsingh@kernel.org>
Cc: martin.lau@linux.dev
Cc: Stanislav Fomichev <sdf@google.com>
Cc: song@kernel.org
Cc: Yonghong Song <yhs@fb.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20221208060256.give.994-kees@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: bcmgenet: Remove the unused function
Jiapeng Chong [Fri, 9 Dec 2022 03:37:23 +0000 (11:37 +0800)]
net: bcmgenet: Remove the unused function

The function dmadesc_get_addr() is defined in the bcmgenet.c file, but
not called elsewhere, so remove this unused function.

drivers/net/ethernet/broadcom/genet/bcmgenet.c:120:26: warning: unused function 'dmadesc_get_addr'.

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3401
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Link: https://lore.kernel.org/r/20221209033723.32452-1-jiapeng.chong@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge branch 'mptcp-miscellaneous-cleanup'
Jakub Kicinski [Sat, 10 Dec 2022 03:46:14 +0000 (19:46 -0800)]
Merge branch 'mptcp-miscellaneous-cleanup'

Mat Martineau says:

====================
mptcp: Miscellaneous cleanup

Two code cleanup patches for the 6.2 merge window that don't change
behavior:

Patch 1 makes proper use of nlmsg_free(), as suggested by Jakub while
reviewing f8c9dfbd875b ("mptcp: add pm listener events").

Patch 2 clarifies success status in a few mptcp functions, which
prevents some smatch false positives.
====================

Link: https://lore.kernel.org/r/20221209004431.143701-1-mathew.j.martineau@linux.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agomptcp: return 0 instead of 'err' var
Matthieu Baerts [Fri, 9 Dec 2022 00:44:31 +0000 (16:44 -0800)]
mptcp: return 0 instead of 'err' var

When 'err' is 0, it looks clearer to return '0' instead of the variable
called 'err'.

The behaviour is then not modified, just a clearer code.

By doing this, we can also avoid false positive smatch warnings like
this one:

  net/mptcp/pm_netlink.c:1169 mptcp_pm_parse_pm_addr_attr() warn: missing error code? 'err'

Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <error27@gmail.com>
Suggested-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agomptcp: use nlmsg_free instead of kfree_skb
Geliang Tang [Fri, 9 Dec 2022 00:44:30 +0000 (16:44 -0800)]
mptcp: use nlmsg_free instead of kfree_skb

Use nlmsg_free() instead of kfree_skb() in pm_netlink.c.

The SKB's have been created by nlmsg_new(). The proper cleaning way
should then be done with nlmsg_free().

For the moment, nlmsg_free() is simply calling kfree_skb() so we don't
change the behaviour here.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge tag 'mlx5-updates-2022-12-08' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Sat, 10 Dec 2022 03:44:42 +0000 (19:44 -0800)]
Merge tag 'mlx5-updates-2022-12-08' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2022-12-08

1) Support range match action in SW steering

Yevgeny Kliteynik says:
=======================

The following patch series adds support for a range match action in
SW Steering.

SW steering is able to match only on the exact values of the packet fields,
as requested by the user: the user provides mask for the fields that are of
interest, and the exact values to be matched on when the traffic is handled.

The following patch series add new type of action - Range Match, where the
user provides a field to be matched on and a range of values (min to max)
that will be considered as hit.

There are several new notions that were implemented in order to support
Range Match:
 - MATCH_RANGES Steering Table Entry (STE): the new STE type that allows
   matching the packets' fields on the range of values instead of a specific
   value.
 - Match Definer: this is a general FW object that defines which fields
   in the packet will be referenced by the mask and tag of each STE.
   Match definer ID is part of STE fields, and it defines how the HW needs
   to interpret the STE's mask/tag values.
   Till now SW steering used the definers that were managed by FW and
   implemented the STE layout as described by the HW spec.
   Now that we're adding a new type of STE, SW steering needs to also be
   able to define this new STE's layout, and this is do

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

2) From OZ add support for meter mtu offload
   2.1: Refactor the code to allow both metering and range post actions as a
        pre-step for adding police mtu offload support.
   2.2: Instantiate mtu green/red flow tables with a single match-all rule.
        Add the green/red actions to the hit/miss table accordingly
   2.3: Initialize the meter object with the TC police mtu parameter.
        Use the hardware range match action feature.

3) From MaorD, support routes with more than 2 nexthops in multipath

4) Michael and Or, improve and extend vport representor counters.

* tag 'mlx5-updates-2022-12-08' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5: Expose steering dropped packets counter
  net/mlx5: Refactor and expand rep vport stat group
  net/mlx5e: multipath, support routes with more than 2 nexthops
  net/mlx5e: TC, add support for meter mtu offload
  net/mlx5e: meter, add mtu post meter tables
  net/mlx5e: meter, refactor to allow multiple post meter tables
  net/mlx5: DR, Add support for range match action
  net/mlx5: DR, Add function that tells if STE miss addr has been initialized
  net/mlx5: DR, Some refactoring of miss address handling
  net/mlx5: DR, Manage definers with refcounts
  net/mlx5: DR, Handle FT action in a separate function
  net/mlx5: DR, Rework is_fw_table function
  net/mlx5: DR, Add functions to create/destroy MATCH_DEFINER general object
  net/mlx5: fs, add match on ranges API
  net/mlx5: mlx5_ifc updates for MATCH_DEFINER general object
====================

Link: https://lore.kernel.org/r/20221209001420.142794-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next...
Jakub Kicinski [Sat, 10 Dec 2022 03:42:16 +0000 (19:42 -0800)]
Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/next-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2022-12-08 (ice)

Jacob Keller says:

This series of patches primarily consists of changes to fix some corner
cases that can cause Tx timestamp failures. The issues were discovered and
reported by Siddaraju DH and primarily affect E822 hardware, though this
series also includes some improvements that affect E810 hardware as well.

The primary issue is regarding the way that E822 determines when to generate
timestamp interrupts. If the driver reads timestamp indexes which do not
have a valid timestamp, the E822 interrupt tracking logic can get stuck.
This is due to the way that E822 hardware tracks timestamp index reads
internally. I was previously unaware of this behavior as it is significantly
different in E810 hardware.

Most of the fixes target refactors to ensure that the ice driver does not
read timestamp indexes which are not valid on E822 hardware. This is done by
using the Tx timestamp ready bitmap register from the PHY. This register
indicates what timestamp indexes have outstanding timestamps waiting to be
captured.

Care must be taken in all cases where we read the timestamp registers, and
thus all flows which might have read these registers are refactored. The
ice_ptp_tx_tstamp function is modified to consolidate as much of the logic
relating to these registers as possible. It now handles discarding stale
timestamps which are old or which occurred after a PHC time update. This
replaces previously standalone thread functions like the periodic work
function and the ice_ptp_flush_tx_tracker function.

In addition, some minor cleanups noticed while writing these refactors are
included.

The remaining patches refactor the E822 implementation to remove the
"bypass" mode for timestamps. The E822 hardware has the ability to provide a
more precise timestamp by making use of measurements of the precise way that
packets flow through the hardware pipeline. These measurements are known as
"Vernier" calibration. The "bypass" mode disables many of these measurements
in favor of a faster start up time for Tx and Rx timestamping. Instead, once
these measurements were captured, the driver tries to reconfigure the PHY to
enable the vernier calibrations.

Unfortunately this recalibration does not work. Testing indicates that the
PHY simply remains in bypass mode without the increased timestamp precision.
Remove the attempt at recalibration and always use vernier mode. This has
one disadvantage that Tx and Rx timestamps cannot begin until after at least
one packet of that type goes through the hardware pipeline. Because of this,
further refactor the driver to separate Tx and Rx vernier calibration.
Complete the Tx and Rx independently, enabling the appropriate type of
timestamp as soon as the relevant packet has traversed the hardware
pipeline. This was reported by Milena Olech.

Note that although these might be considered "bug fixes", the required
changes in order to appropriately resolve these issues is large. Thus it
does not feel suitable to send this series to net.

* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue:
  ice: reschedule ice_ptp_wait_for_offset_valid during reset
  ice: make Tx and Rx vernier offset calibration independent
  ice: only check set bits in ice_ptp_flush_tx_tracker
  ice: handle flushing stale Tx timestamps in ice_ptp_tx_tstamp
  ice: cleanup allocations in ice_ptp_alloc_tx_tracker
  ice: protect init and calibrating check in ice_ptp_request_ts
  ice: synchronize the misc IRQ when tearing down Tx tracker
  ice: check Tx timestamp memory register for ready timestamps
  ice: handle discarding old Tx requests in ice_ptp_tx_tstamp
  ice: always call ice_ptp_link_change and make it void
  ice: fix misuse of "link err" with "link status"
  ice: Reset TS memory for all quads
  ice: Remove the E822 vernier "bypass" logic
  ice: Use more generic names for ice_ptp_tx fields
====================

Link: https://lore.kernel.org/r/20221208213932.1274143-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agodocs/bpf: Add documentation for BPF_MAP_TYPE_SK_STORAGE
Donald Hunter [Fri, 9 Dec 2022 11:24:01 +0000 (11:24 +0000)]
docs/bpf: Add documentation for BPF_MAP_TYPE_SK_STORAGE

Add documentation for the BPF_MAP_TYPE_SK_STORAGE including
kernel version introduced, usage and examples.

Signed-off-by: Donald Hunter <donald.hunter@gmail.com>
Acked-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/r/20221209112401.69319-1-donald.hunter@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
22 months agonet: openvswitch: Add support to count upcall packets
wangchuanlei [Wed, 7 Dec 2022 01:38:57 +0000 (20:38 -0500)]
net: openvswitch: Add support to count upcall packets

Add support to count upall packets, when kmod of openvswitch
upcall to count the number of packets for upcall succeed and
failed, which is a better way to see how many packets upcalled
on every interfaces.

Signed-off-by: wangchuanlei <wangchuanlei@inspur.com>
Acked-by: Eelco Chaudron <echaudro@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agorhashtable: Allow rhashtable to be used from irq-safe contexts
Tejun Heo [Tue, 6 Dec 2022 21:36:32 +0000 (11:36 -1000)]
rhashtable: Allow rhashtable to be used from irq-safe contexts

rhashtable currently only does bh-safe synchronization making it impossible
to use from irq-safe contexts. Switch it to use irq-safe synchronization to
remove the restriction.

v2: Update the lock functions to return the ulong flags value and unlock
    functions to take the value directly instead of passing around the
    pointer. Suggested by Linus.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: David Vernet <dvernet@meta.com>
Acked-by: Josh Don <joshdon@google.com>
Acked-by: Hao Luo <haoluo@google.com>
Acked-by: Barret Rhoden <brho@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoMerge branch 'net-sched-retpoline'
David S. Miller [Fri, 9 Dec 2022 09:18:08 +0000 (09:18 +0000)]
Merge branch 'net-sched-retpoline'

Pedro Tammela says:

====================
net/sched: retpoline wrappers for tc

In tc all qdics, classifiers and actions can be compiled as modules.
This results today in indirect calls in all transitions in the tc hierarchy.
Due to CONFIG_RETPOLINE, CPUs with mitigations=on might pay an extra cost on
indirect calls. For newer Intel cpus with IBRS the extra cost is
nonexistent, but AMD Zen cpus and older x86 cpus still go through the
retpoline thunk.

Known built-in symbols can be optimized into direct calls, thus
avoiding the retpoline thunk. So far, tc has not been leveraging this
build information and leaving out a performance optimization for some
CPUs. In this series we wire up 'tcf_classify()' and 'tcf_action_exec()'
with direct calls when known modules are compiled as built-in as an
opt-in optimization.

We measured these changes in one AMD Zen 4 cpu (Retpoline), one AMD Zen 3 cpu (Retpoline),
one Intel 10th Gen CPU (IBRS), one Intel 3rd Gen cpu (Retpoline) and one
Intel Xeon CPU (IBRS) using pktgen with 64b udp packets. Our test setup is a
dummy device with clsact and matchall in a kernel compiled with every
tc module as built-in.  We observed a 3-8% speed up on the retpoline CPUs,
when going through 1 tc filter, and a 60-100% speed up when going through 100 filters.
For the IBRS cpus we observed a 1-2% degradation in both scenarios, we believe
the extra branches check introduced a small overhead therefore we added
a static key that bypasses the wrapper on kernels not using the retpoline mitigation,
but compiled with CONFIG_RETPOLINE.

1 filter:
CPU        | before (pps) | after (pps) | diff
R9 7950X   | 5914980      | 6380227     | +7.8%
R9 5950X   | 4237838      | 4412241     | +4.1%
R9 5950X   | 4265287      | 4413757     | +3.4%   [*]
i5-3337U   | 1580565      | 1682406     | +6.4%
i5-10210U  | 3006074      | 3006857     | +0.0%
i5-10210U  | 3160245      | 3179945     | +0.6%   [*]
Xeon 6230R | 3196906      | 3197059     | +0.0%
Xeon 6230R | 3190392      | 3196153     | +0.01%  [*]

100 filters:
CPU        | before (pps) | after (pps) | diff
R9 7950X   | 373598       | 820396      | +119.59%
R9 5950X   | 313469       | 633303      | +102.03%
R9 5950X   | 313797       | 633150      | +101.77% [*]
i5-3337U   | 127454       | 211210      | +65.71%
i5-10210U  | 389259       | 381765      | -1.9%
i5-10210U  | 408812       | 412730      | +0.9%    [*]
Xeon 6230R | 415420       | 406612      | -2.1%
Xeon 6230R | 416705       | 405869      | -2.6%    [*]

[*] In these tests we ran pktgen with clone set to 1000.

On the 7950x system we also tested the impact of filters if iteration order
placement varied, first by compiling a kernel with the filter under test being
the first one in the static iteration and then repeating it with being last (of 15 classifiers existing today).
We saw a difference of +0.5-1% in pps between being the first in the iteration vs being the last.
Therefore we order the classifiers and actions according to relevance per our current thinking.

v5->v6:
- Address Eric Dumazet suggestions

v4->v5:
- Rebase

v3->v4:
- Address Eric Dumazet suggestions

v2->v3:
- Address suggestions by Jakub, Paolo and Eric
- Dropped RFC tag (I forgot to add it on v2)

v1->v2:
- Fix build errors found by the bots
- Address Kuniyuki Iwashima suggestions

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

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet/sched: avoid indirect classify functions on retpoline kernels
Pedro Tammela [Tue, 6 Dec 2022 13:55:13 +0000 (10:55 -0300)]
net/sched: avoid indirect classify functions on retpoline kernels

Expose the necessary tc classifier functions and wire up cls_api to use
direct calls in retpoline kernels.

Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Victor Nogueira <victor@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet/sched: avoid indirect act functions on retpoline kernels
Pedro Tammela [Tue, 6 Dec 2022 13:55:12 +0000 (10:55 -0300)]
net/sched: avoid indirect act functions on retpoline kernels

Expose the necessary tc act functions and wire up act_api to use
direct calls in retpoline kernels.

Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Victor Nogueira <victor@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>