platform/kernel/linux-rpi.git
18 months agodt-bindings: net: bluetooth: Add NXP bluetooth support
Neeraj Sanjay Kale [Thu, 16 Mar 2023 17:22:13 +0000 (22:52 +0530)]
dt-bindings: net: bluetooth: Add NXP bluetooth support

Add binding document for NXP bluetooth chipsets attached over UART.

Signed-off-by: Neeraj Sanjay Kale <neeraj.sanjaykale@nxp.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoserdev: Add method to assert break signal over tty UART port
Neeraj Sanjay Kale [Thu, 16 Mar 2023 17:22:12 +0000 (22:52 +0530)]
serdev: Add method to assert break signal over tty UART port

Adds serdev_device_break_ctl() and an implementation for ttyport.
This function simply calls the break_ctl in tty layer, which can
assert a break signal over UART-TX line, if the tty and the
underlying platform and UART peripheral supports this operation.

Signed-off-by: Neeraj Sanjay Kale <neeraj.sanjaykale@nxp.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoserdev: Replace all instances of ENOTSUPP with EOPNOTSUPP
Neeraj Sanjay Kale [Thu, 16 Mar 2023 17:22:11 +0000 (22:52 +0530)]
serdev: Replace all instances of ENOTSUPP with EOPNOTSUPP

This replaces all instances of ENOTSUPP with EOPNOTSUPP since ENOTSUPP
is not a standard error code. This will help maintain consistency in
error codes when new serdev API's are added.

Signed-off-by: Neeraj Sanjay Kale <neeraj.sanjaykale@nxp.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_ldisc: Fix tty_set_termios() return value assumptions
Ilpo Järvinen [Thu, 9 Mar 2023 08:20:30 +0000 (10:20 +0200)]
Bluetooth: hci_ldisc: Fix tty_set_termios() return value assumptions

tty_set_termios() never returns anything else than 0. Make the debug
prints to look directly into the new termios instead to check CRTSCTS
state.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: Improve support for Actions Semi ATS2851 based devices
Raul Cheleguini [Fri, 10 Mar 2023 15:14:10 +0000 (15:14 +0000)]
Bluetooth: Improve support for Actions Semi ATS2851 based devices

Add two more quirks to resume the device initialization and basic
operation as the device seems not to support "Read Transmit Power"
and "Set Extended Scan Parameters".

< HCI Command: LE Read Transmit Power (0x08|0x004b) plen 0
> HCI Event: Command Status (0x0f) plen 4
      LE Read Transmit Power (0x08|0x004b) ncmd 1
        Status: Unknown HCI Command (0x01)

< HCI Command: LE Set Extended Scan Parameters (0x08|0x0041) plen 8
        Own address type: Random (0x01)
        Filter policy: Accept all advertisement (0x00)
        PHYs: 0x01
        Entry 0: LE 1M
          Type: Active (0x01)
          Interval: 11.250 msec (0x0012)
          Window: 11.250 msec (0x0012)
> HCI Event: Command Status (0x0f) plen 4
      LE Set Extended Scan Parameters (0x08|0x0041) ncmd 1
        Status: Unknown HCI Command (0x01)

Signed-off-by: Raul Cheleguini <rcheleguini@google.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: btrtl: add support for the RTL8723CS
Vasily Khoruzhick [Tue, 7 Mar 2023 22:17:31 +0000 (23:17 +0100)]
Bluetooth: btrtl: add support for the RTL8723CS

The Realtek RTL8723CS is a SDIO WiFi chip. It also contains a Bluetooth
module which is connected via UART to the host.

It shares lmp subversion with 8703B, so Realtek's userspace
initialization tool (rtk_hciattach) differentiates varieties of RTL8723CS
(CG, VF, XX) with RTL8703B using vendor's command to read chip type.

Also this chip declares support for some features it doesn't support
so add a quirk to indicate that these features are broken.

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Signed-off-by: Bastian Germann <bage@debian.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: Add new quirk for broken local ext features page 2
Vasily Khoruzhick [Tue, 7 Mar 2023 22:17:30 +0000 (23:17 +0100)]
Bluetooth: Add new quirk for broken local ext features page 2

Some adapters (e.g. RTL8723CS) advertise that they have more than
2 pages for local ext features, but they don't support any features
declared in these pages. RTL8723CS reports max_page = 2 and declares
support for sync train and secure connection, but it responds with
either garbage or with error in status on corresponding commands.

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Signed-off-by: Bastian Germann <bage@debian.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: btmtkuart: mark OF related data as maybe unused
Krzysztof Kozlowski [Sat, 11 Mar 2023 11:13:54 +0000 (12:13 +0100)]
Bluetooth: btmtkuart: mark OF related data as maybe unused

The driver can be compile tested with !CONFIG_OF making certain data
unused:

  drivers/bluetooth/btmtkuart.c:971:36: error: ‘mt7668_data’ defined
  but not used [-Werror=unused-const-variable=]
  drivers/bluetooth/btmtkuart.c:966:36: error: ‘mt7663_data’ defined
  but not used [-Werror=unused-const-variable=]
  drivers/bluetooth/btmtkuart.c:962:36: error: ‘mt7622_data’ defined
  but not used [-Werror=unused-const-variable=]

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_qca: mark OF related data as maybe unused
Krzysztof Kozlowski [Sat, 11 Mar 2023 11:13:53 +0000 (12:13 +0100)]
Bluetooth: hci_qca: mark OF related data as maybe unused

The driver can be compile tested with !CONFIG_OF making certain data
unused:

  drivers/bluetooth/hci_qca.c:1869:37: error: ‘qca_soc_data_wcn6750’
  defined but not used [-Werror=unused-const-variable=]
  drivers/bluetooth/hci_qca.c:1853:37: error: ‘qca_soc_data_wcn3998’
  defined but not used [-Werror=unused-const-variable=]
  drivers/bluetooth/hci_qca.c:1841:37: error: ‘qca_soc_data_wcn3991’
  defined but not used [-Werror=unused-const-variable=]
  drivers/bluetooth/hci_qca.c:1830:37: error: ‘qca_soc_data_wcn3990’
  defined but not used [-Werror=unused-const-variable=]

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: btmrvl_sdio: mark OF related data as maybe unused
Krzysztof Kozlowski [Sat, 11 Mar 2023 11:13:52 +0000 (12:13 +0100)]
Bluetooth: btmrvl_sdio: mark OF related data as maybe unused

The driver can be compile tested with !CONFIG_OF making certain data
unused:

  drivers/bluetooth/btmrvl_sdio.c:43:34: error:
  ‘btmrvl_sdio_of_match_table’ defined but not used
  [-Werror=unused-const-variable=]

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_ll: drop of_match_ptr for ID table
Krzysztof Kozlowski [Sat, 11 Mar 2023 11:13:51 +0000 (12:13 +0100)]
Bluetooth: hci_ll: drop of_match_ptr for ID table

The driver can match only via the DT table so the table should be always
used and the of_match_ptr does not have any sense (this also allows ACPI
matching via PRP0001, even though it might not be relevant here).

  drivers/bluetooth/hci_ll.c:769:34: error: ‘hci_ti_of_match’ defined
  but not used [-Werror=unused-const-variable=]

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: L2CAP: Delay identity address updates
Luiz Augusto von Dentz [Thu, 9 Mar 2023 00:16:31 +0000 (16:16 -0800)]
Bluetooth: L2CAP: Delay identity address updates

This delays the identity address updates to give time for userspace to
process the new address otherwise there is a risk that userspace
creates a duplicated device if the MGMT event is delayed for some
reason.

Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_sync: Remove duplicate statement
Inga Stotland [Wed, 8 Mar 2023 18:32:35 +0000 (10:32 -0800)]
Bluetooth: hci_sync: Remove duplicate statement

This removes the following duplicate statement in
hci_le_ext_directed_advertising_sync():
cp.own_addr_type = own_addr_type;

Signed-off-by: Inga Stotland <inga.stotland@intel.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: Convert MSFT filter HCI cmd to hci_sync
Brian Gix [Fri, 24 Feb 2023 17:56:32 +0000 (09:56 -0800)]
Bluetooth: Convert MSFT filter HCI cmd to hci_sync

The msft_set_filter_enable() command was using the deprecated
hci_request mechanism rather than hci_sync. This caused the warning error:
hci0: HCI_REQ-0xfcf0

Signed-off-by: Brian Gix <brian.gix@intel.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_sync: Don't wait peer's reply when powering off
Archie Pusaka [Mon, 6 Mar 2023 09:07:07 +0000 (17:07 +0800)]
Bluetooth: hci_sync: Don't wait peer's reply when powering off

Currently, when we initiate disconnection, we will wait for the peer's
reply unless when we are suspending, where we fire and forget the
disconnect request.

A similar case is when adapter is powering off. However, we still wait
for the peer's reply in this case. Therefore, if the peer is
unresponsive, the command will time out and the power off sequence
will fail, causing "bluetooth powered on by itself" to users.

This patch makes the host doesn't wait for the peer's reply when the
disconnection reason is powering off.

Signed-off-by: Archie Pusaka <apusaka@chromium.org>
Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: Add VID/PID 0489/e0e4 for MediaTek MT7922
Meng Tang [Mon, 6 Mar 2023 01:44:59 +0000 (09:44 +0800)]
Bluetooth: Add VID/PID 0489/e0e4 for MediaTek MT7922

Tested on Acer Nitro AN515-47 Notebook

output from /sys/kernel/debug/usb/devices:

T:  Bus=03 Lev=01 Prnt=01 Port=02 Cnt=02 Dev#=  2 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=ef(misc ) Sub=02 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=0489 ProdID=e0e4 Rev= 1.00
S:  Manufacturer=MediaTek Inc.
S:  Product=Wireless_Device
S:  SerialNumber=000000000
C:* #Ifs= 3 Cfg#= 1 Atr=e0 MxPwr=100mA
A:  FirstIf#= 0 IfCount= 3 Cls=e0(wlcon) Sub=01 Prot=01
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  16 Ivl=125us
E:  Ad=82(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
I:  If#= 1 Alt= 6 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  63 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  63 Ivl=1ms
I:* If#= 2 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=(none)
E:  Ad=8a(I) Atr=03(Int.) MxPS=  64 Ivl=125us
E:  Ad=0a(O) Atr=03(Int.) MxPS=  64 Ivl=125us
I:  If#= 2 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=(none)
E:  Ad=8a(I) Atr=03(Int.) MxPS= 512 Ivl=125us
E:  Ad=0a(O) Atr=03(Int.) MxPS= 512 Ivl=125us

Signed-off-by: Meng Tang <tangmeng@uniontech.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoarm64: dts: rockchip: Update compatible for bluetooth
Chris Morgan [Tue, 28 Feb 2023 15:22:05 +0000 (09:22 -0600)]
arm64: dts: rockchip: Update compatible for bluetooth

Update the compatible for the Realtek RTL8821CS bluetooth node.

Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_h5: btrtl: Add support for RTL8821CS
Chris Morgan [Tue, 28 Feb 2023 15:22:04 +0000 (09:22 -0600)]
Bluetooth: hci_h5: btrtl: Add support for RTL8821CS

RTL8821CS is a WiFi + Bluetooth combo chip from Realtek that provides
WiFi A/B/G/N/AC over an SDIO interface and Bluetooth 4.2 over a UART
interface.

Note that the firmware this was tested with was firmware version
0x75b8f098.

Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agodt-bindings: net: realtek-bluetooth: Add RTL8821CS
Chris Morgan [Tue, 28 Feb 2023 15:22:03 +0000 (09:22 -0600)]
dt-bindings: net: realtek-bluetooth: Add RTL8821CS

Add compatible string for RTL8821CS for existing Realtek Bluetooth
driver.

Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Alistair Francis <alistair@alistair23.me>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: btusb: Add new PID/VID 04ca:3801 for MT7663
Meng Tang [Tue, 28 Feb 2023 05:55:17 +0000 (13:55 +0800)]
Bluetooth: btusb: Add new PID/VID 04ca:3801 for MT7663

This bluetooth device is found in a combo WLAN/BT card
for a MediaTek 7663.

Tested on Acer Aspire A315-24P Notebook

The device information:

T:  Bus=01 Lev=01 Prnt=01 Port=01 Cnt=01 Dev#=  2 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=ef(misc ) Sub=02 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=04ca ProdID=3801 Rev= 1.00
S:  Manufacturer=MediaTek Inc.
S:  Product=Wireless_Device
S:  SerialNumber=000000000
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=100mA
A:  FirstIf#= 0 IfCount= 2 Cls=e0(wlcon) Sub=01 Prot=01
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  16 Ivl=125us
E:  Ad=82(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
I:  If#= 1 Alt= 6 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  63 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  63 Ivl=1ms

Signed-off-by: Meng Tang <tangmeng@uniontech.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_sync: Fix smatch warning
Luiz Augusto von Dentz [Mon, 27 Feb 2023 23:50:58 +0000 (15:50 -0800)]
Bluetooth: hci_sync: Fix smatch warning

This fixes the following new warning:

net/bluetooth/hci_sync.c:2403 hci_pause_addr_resolution() warn: missing
error code? 'err'

Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <error27@gmail.com>
Link: https://lore.kernel.org/r/202302251952.xryXOegd-lkp@intel.com/
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_core: Make hci_conn_hash_add append to the list
Luiz Augusto von Dentz [Fri, 24 Feb 2023 23:27:33 +0000 (15:27 -0800)]
Bluetooth: hci_core: Make hci_conn_hash_add append to the list

This makes hci_conn_hash_add append to the tail of the conn_hash so it
matches the order they are created, this is required if the controller
attempts to match the order of ACL with CIS which uses append logic
when programming the CIS ids on the CIG.

The result of this change affects Create CIS:

Before:

< HCI Command: LE Create Connected Isochronous Stream (0x08|0x0064) plen 9
        Number of CIS: 2
        CIS Handle: 2560
        ACL Handle: 3586
        CIS Handle: 2561
        ACL Handle: 3585

After:

< HCI Command: LE Create Connected Isochronous Stream (0x08|0x0064) plen 9
        Number of CIS: 2
        CIS Handle: 2560
        ACL Handle: 3585
        CIS Handle: 2561
        ACL Handle: 3586

Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: btrtl: Add support for RTL8852BS
Victor Hassan [Wed, 22 Feb 2023 09:47:20 +0000 (17:47 +0800)]
Bluetooth: btrtl: Add support for RTL8852BS

Add the support for RTL8852BS BT controller on UART interface.
The necessary firmware file will be submitted to linux-firmware.

Signed-off-by: Victor Hassan <victor@allwinnertech.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_mrvl: Add serdev support for 88W8997
Stefan Eichenberger [Mon, 13 Feb 2023 12:09:25 +0000 (13:09 +0100)]
Bluetooth: hci_mrvl: Add serdev support for 88W8997

Add serdev support for the 88W8997 from NXP (previously Marvell). It
includes support for changing the baud rate. The command to change the
baud rate is taken from the user manual UM11483 Rev. 9 in section 7
(Bring-up of Bluetooth interfaces) from NXP.

Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: hci_mrvl: use maybe_unused macro for device tree ids
Stefan Eichenberger [Mon, 13 Feb 2023 12:09:24 +0000 (13:09 +0100)]
Bluetooth: hci_mrvl: use maybe_unused macro for device tree ids

Use the maybe_unused macro for the device tree ids instead of #ifdef
CONFIG_OF. This makes it easier to add support for new devices.

Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agodt-bindings: bluetooth: marvell: add max-speed property
Stefan Eichenberger [Mon, 13 Feb 2023 12:09:23 +0000 (13:09 +0100)]
dt-bindings: bluetooth: marvell: add max-speed property

The 88W8997 bluetooth module supports setting the max-speed property.

Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agodt-bindings: bluetooth: marvell: add 88W8997
Stefan Eichenberger [Mon, 13 Feb 2023 12:09:22 +0000 (13:09 +0100)]
dt-bindings: bluetooth: marvell: add 88W8997

Update the documentation with the device tree binding for the Marvell
88W8997 bluetooth device.

Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: btusb: Do not require hardcoded interface numbers
Tomasz Moń [Tue, 7 Feb 2023 11:57:41 +0000 (12:57 +0100)]
Bluetooth: btusb: Do not require hardcoded interface numbers

Remove hardcoded interface number check because Bluetooth specification
since version 4.0 only recommends and no longer requires specific
interface numbers.

While earlier Bluetooth versions, i.e. 2.1 + EDR and 3.0 + HS, contain
required configuration table in Volume 4 - Host Controller Interface
Part B - USB Transport Layer, Bluetooth Core Specification Addendum 2
changes the table from required to recommended configuration.

Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoBluetooth: MGMT: Use BIT macro when defining bitfields
Luiz Augusto von Dentz [Mon, 13 Feb 2023 22:28:55 +0000 (14:28 -0800)]
Bluetooth: MGMT: Use BIT macro when defining bitfields

This makes use of BIT macro when defining bitfields which makes it
clearer what bit it is toggling.

Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
18 months agoMerge branch 'act_pedit-minor-improvements'
David S. Miller [Sun, 23 Apr 2023 17:35:27 +0000 (18:35 +0100)]
Merge branch 'act_pedit-minor-improvements'

Pedro Tammela says:

====================
net/sched: act_pedit: minor improvements

This series aims to improve the code and usability of act_pedit for
netlink users.

Patches 1-2 improves error reporting for extended keys parsing with extack.

Patch 3 checks the static offsets a priori on create/update. Currently,
this is done at the datapath for both static and runtime offsets.

Patch 4 removes a check from the datapath which is redundant since the
netlink parsing validates the key types.

Patch 5 changes the 'pr_info()' calls in the datapath to rate limited
versions.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet/sched: act_pedit: rate limit datapath messages
Pedro Tammela [Fri, 21 Apr 2023 21:25:17 +0000 (18:25 -0300)]
net/sched: act_pedit: rate limit datapath messages

Unbounded info messages in the pedit datapath can flood the printk
ring buffer quite easily depending on the action created.
As these messages are informational, usually printing some, not all,
is enough to bring attention to the real issue.

Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet/sched: act_pedit: remove extra check for key type
Pedro Tammela [Fri, 21 Apr 2023 21:25:16 +0000 (18:25 -0300)]
net/sched: act_pedit: remove extra check for key type

The netlink parsing already validates the key 'htype'.
Remove the datapath check as it's redundant.

Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet/sched: act_pedit: check static offsets a priori
Pedro Tammela [Fri, 21 Apr 2023 21:25:15 +0000 (18:25 -0300)]
net/sched: act_pedit: check static offsets a priori

Static key offsets should always be on 32 bit boundaries. Validate them on
create/update time for static offsets and move the datapath validation
for runtime offsets only.

iproute2 already errors out if a given offset and data size cannot be
packed to a 32 bit boundary. This change will make sure users which
create/update pedit instances directly via netlink also error out,
instead of finding out when packets are traversing.

Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet/sched: act_pedit: use extack in 'ex' parsing errors
Pedro Tammela [Fri, 21 Apr 2023 21:25:14 +0000 (18:25 -0300)]
net/sched: act_pedit: use extack in 'ex' parsing errors

We have extack available when parsing 'ex' keys, so pass it to
tcf_pedit_keys_ex_parse and add more detailed error messages.

Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet/sched: act_pedit: use NLA_POLICY for parsing 'ex' keys
Pedro Tammela [Fri, 21 Apr 2023 21:25:13 +0000 (18:25 -0300)]
net/sched: act_pedit: use NLA_POLICY for parsing 'ex' keys

Transform two checks in the 'ex' key parsing into netlink policies
removing extra if checks.

Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: sched: Print msecs when transmit queue time out
Yajun Deng [Fri, 21 Apr 2023 08:26:06 +0000 (16:26 +0800)]
net: sched: Print msecs when transmit queue time out

The kernel will print several warnings in a short period of time
when it stalls. Like this:

First warning:
[ 7100.097547] ------------[ cut here ]------------
[ 7100.097550] NETDEV WATCHDOG: eno2 (xxx): transmit queue 8 timed out
[ 7100.097571] WARNING: CPU: 8 PID: 0 at net/sched/sch_generic.c:467
                       dev_watchdog+0x260/0x270
...

Second warning:
[ 7147.756952] rcu: INFO: rcu_preempt self-detected stall on CPU
[ 7147.756958] rcu:   24-....: (59999 ticks this GP) idle=546/1/0x400000000000000
                      softirq=367      3137/3673146 fqs=13844
[ 7147.756960]        (t=60001 jiffies g=4322709 q=133381)
[ 7147.756962] NMI backtrace for cpu 24
...

We calculate that the transmit queue start stall should occur before
7095s according to watchdog_timeo, the rcu start stall at 7087s.
These two times are close together, it is difficult to confirm which
happened first.

To let users know the exact time the stall started, print msecs when
the transmit queue time out.

Signed-off-by: Yajun Deng <yajun.deng@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge branch 'dsa-skb_mac_header'
David S. Miller [Sun, 23 Apr 2023 13:16:45 +0000 (14:16 +0100)]
Merge branch 'dsa-skb_mac_header'

Vladimir Oltean says:

====================
Remove skb_mac_header() dependency in DSA xmit path

Eric started working on removing skb_mac_header() assumptions from the
networking xmit path, and I offered to help for DSA:
https://lore.kernel.org/netdev/20230321164519.1286357-1-edumazet@google.com/

The majority of this patch set is a straightforward replacement of
skb_mac_header() with skb->data (hidden either behind skb_eth_hdr(), or
behind skb_vlan_eth_hdr()). The only patch which is more "interesting"
is 9/9.

Another potential caller of __skb_vlan_pop() on xmit (and therefore
also of skb_mac_header()) is tcf_vlan_act(), but I haven't had the time
to investigate that (enough to submit changes other than what's here).

v1->v2:
- 09/09: document the vlan_tci argument of vlan_remove_tag() in the kdoc

v1 at:
https://lore.kernel.org/netdev/20230322233823.1806736-1-vladimir.oltean@nxp.com/

Cc: Madalin Bucur <madalin.bucur@nxp.com>
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: tag_ocelot: call only the relevant portion of __skb_vlan_pop() on TX
Vladimir Oltean [Thu, 20 Apr 2023 22:56:01 +0000 (01:56 +0300)]
net: dsa: tag_ocelot: call only the relevant portion of __skb_vlan_pop() on TX

ocelot_xmit_get_vlan_info() calls __skb_vlan_pop() as the most
appropriate helper I could find which strips away a VLAN header.
That's all I need it to do, but __skb_vlan_pop() has more logic, which
will become incompatible with the future revert of commit 6d1ccff62780
("net: reset mac header in dev_start_xmit()").

Namely, it performs a sanity check on skb_mac_header(), which will stop
being set after the above revert, so it will return an error instead of
removing the VLAN tag.

ocelot_xmit_get_vlan_info() gets called in 2 circumstances:

(1) the port is under a VLAN-aware bridge and the bridge sends
    VLAN-tagged packets

(2) the port is under a VLAN-aware bridge and somebody else (an 8021q
    upper) sends VLAN-tagged packets (using a VID that isn't in the
    bridge vlan tables)

In case (1), there is actually no bug to defend against, because
br_dev_xmit() calls skb_reset_mac_header() and things continue to work.

However, in case (2), illustrated using the commands below, it can be
seen that our intervention is needed, since __skb_vlan_pop() complains:

$ ip link add br0 type bridge vlan_filtering 1 && ip link set br0 up
$ ip link set $eth master br0 && ip link set $eth up
$ ip link add link $eth name $eth.100 type vlan id 100 && ip link set $eth.100 up
$ ip addr add 192.168.100.1/24 dev $eth.100

I could fend off the checks in __skb_vlan_pop() with some
skb_mac_header_was_set() calls, but seeing how few callers of
__skb_vlan_pop() there are from TX paths, that seems rather
unproductive.

As an alternative solution, extract the bare minimum logic to strip a
VLAN header, and move it to a new helper named vlan_remove_tag(), close
to the definition of vlan_insert_tag(). Document it appropriately and
make ocelot_xmit_get_vlan_info() call this smaller helper instead.

Seeing that it doesn't appear illegal to test skb->protocol in the TX
path, I guess it would be a good for vlan_remove_tag() to also absorb
the vlan_set_encap_proto() function call.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: update TX path comments to not mention skb_mac_header()
Vladimir Oltean [Thu, 20 Apr 2023 22:56:00 +0000 (01:56 +0300)]
net: dsa: update TX path comments to not mention skb_mac_header()

Once commit 6d1ccff62780 ("net: reset mac header in dev_start_xmit()")
will be reverted, it will no longer be true that skb->data points at
skb_mac_header(skb) - since the skb->mac_header will not be set - so
stop saying that, and just say that it points to the MAC header.

I've reviewed vlan_insert_tag() and it does not *actually* depend on
skb_mac_header(), so reword that to avoid the confusion.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: tag_sja1105: replace skb_mac_header() with vlan_eth_hdr()
Vladimir Oltean [Thu, 20 Apr 2023 22:55:59 +0000 (01:55 +0300)]
net: dsa: tag_sja1105: replace skb_mac_header() with vlan_eth_hdr()

This is a cosmetic patch which consolidates the code to use the helper
function offered by if_vlan.h.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: tag_sja1105: don't rely on skb_mac_header() in TX paths
Vladimir Oltean [Thu, 20 Apr 2023 22:55:58 +0000 (01:55 +0300)]
net: dsa: tag_sja1105: don't rely on skb_mac_header() in TX paths

skb_mac_header() will no longer be available in the TX path when
reverting commit 6d1ccff62780 ("net: reset mac header in
dev_start_xmit()"). As preparation for that, let's use
skb_vlan_eth_hdr() to get to the VLAN header instead, which assumes it's
located at skb->data (assumption which holds true here).

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: tag_ksz: do not rely on skb_mac_header() in TX paths
Vladimir Oltean [Thu, 20 Apr 2023 22:55:57 +0000 (01:55 +0300)]
net: dsa: tag_ksz: do not rely on skb_mac_header() in TX paths

skb_mac_header() will no longer be available in the TX path when
reverting commit 6d1ccff62780 ("net: reset mac header in
dev_start_xmit()"). As preparation for that, let's use skb_eth_hdr() to
get to the Ethernet header's MAC DA instead, helper which assumes this
header is located at skb->data (assumption which holds true here).

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dsa: tag_ocelot: do not rely on skb_mac_header() for VLAN xmit
Vladimir Oltean [Thu, 20 Apr 2023 22:55:56 +0000 (01:55 +0300)]
net: dsa: tag_ocelot: do not rely on skb_mac_header() for VLAN xmit

skb_mac_header() will no longer be available in the TX path when
reverting commit 6d1ccff62780 ("net: reset mac header in
dev_start_xmit()"). As preparation for that, let's use
skb_vlan_eth_hdr() to get to the VLAN header instead, which assumes it's
located at skb->data (assumption which holds true here).

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: dpaa: avoid one skb_reset_mac_header() in dpaa_enable_tx_csum()
Vladimir Oltean [Thu, 20 Apr 2023 22:55:55 +0000 (01:55 +0300)]
net: dpaa: avoid one skb_reset_mac_header() in dpaa_enable_tx_csum()

It appears that dpaa_enable_tx_csum() only calls skb_reset_mac_header()
to get to the VLAN header using skb_mac_header().

We can use skb_vlan_eth_hdr() to get to the VLAN header based on
skb->data directly. This avoids spending a few cycles to set
skb->mac_header.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Madalin Bucur <madalin.bucur@oss.nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: vlan: introduce skb_vlan_eth_hdr()
Vladimir Oltean [Thu, 20 Apr 2023 22:55:54 +0000 (01:55 +0300)]
net: vlan: introduce skb_vlan_eth_hdr()

Similar to skb_eth_hdr() introduced in commit 96cc4b69581d ("macvlan: do
not assume mac_header is set in macvlan_broadcast()"), let's introduce a
skb_vlan_eth_hdr() helper which can be used in TX-only code paths to get
to the VLAN header based on skb->data rather than based on the
skb_mac_header(skb).

We also consolidate the drivers that dereference skb->data to go through
this helper.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: vlan: don't adjust MAC header in __vlan_insert_inner_tag() unless set
Vladimir Oltean [Thu, 20 Apr 2023 22:55:53 +0000 (01:55 +0300)]
net: vlan: don't adjust MAC header in __vlan_insert_inner_tag() unless set

This is a preparatory change for the deletion of skb_reset_mac_header(skb)
from __dev_queue_xmit(). After that deletion, skb_mac_header(skb) will
no longer be set in TX paths, from which __vlan_insert_inner_tag() can
still be called (perhaps indirectly).

If we don't make this change, then an unset MAC header (equal to ~0U)
will become set after the adjustment with VLAN_HLEN.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agodrivers/net/phy: add driver for Microchip LAN867x 10BASE-T1S PHY
Ramón Nordin Rodriguez [Fri, 21 Apr 2023 16:38:54 +0000 (18:38 +0200)]
drivers/net/phy: add driver for Microchip LAN867x 10BASE-T1S PHY

This patch adds support for the Microchip LAN867x 10BASE-T1S family
(LAN8670/1/2). The driver supports P2MP with PLCA.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Ramón Nordin Rodriguez <ramon.nordin.rodriguez@ferroamp.se>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agorxrpc: Replace fake flex-array with flexible-array member
Gustavo A. R. Silva [Fri, 21 Apr 2023 13:08:53 +0000 (14:08 +0100)]
rxrpc: Replace fake flex-array with flexible-array member

Zero-length arrays as fake flexible arrays are deprecated and we are
moving towards adopting C99 flexible-array members instead.

Transform zero-length array into flexible-array member in struct
rxrpc_ackpacket.

Address the following warnings found with GCC-13 and
-fstrict-flex-arrays=3 enabled:
net/rxrpc/call_event.c:149:38: warning: array subscript i is outside array bounds of ‘uint8_t[0]’ {aka ‘unsigned char[]’} [-Warray-bounds=]

This helps with the ongoing efforts to tighten the FORTIFY_SOURCE
routines on memcpy() and help us make progress towards globally
enabling -fstrict-flex-arrays=3 [1].

Link: https://github.com/KSPP/linux/issues/21
Link: https://github.com/KSPP/linux/issues/263
Link: https://gcc.gnu.org/pipermail/gcc-patches/2022-October/602902.html
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Jeffrey Altman <jaltman@auristor.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: "David S. Miller" <davem@davemloft.net>
cc: Eric Dumazet <edumazet@google.com>
cc: Jakub Kicinski <kuba@kernel.org>
cc: Paolo Abeni <pabeni@redhat.com>
cc: linux-afs@lists.infradead.org
cc: netdev@vger.kernel.org
cc: linux-hardening@vger.kernel.org
Link: https://lore.kernel.org/r/ZAZT11n4q5bBttW0@work/
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge branch 'napi_threaded_poll-enhancements'
David S. Miller [Sun, 23 Apr 2023 12:35:08 +0000 (13:35 +0100)]
Merge branch 'napi_threaded_poll-enhancements'

Eric Dumazet says:

====================
net: give napi_threaded_poll() some love

There is interest to revert commit 4cd13c21b207
("softirq: Let ksoftirqd do its job") and use instead the
napi_threaded_poll() mode.

https://lore.kernel.org/netdev/140f61e2e1fcb8cf53619709046e312e343b53ca.camel@redhat.com/T/#m8a8f5b09844adba157ad0d22fc1233d97013de50

Before doing so, make sure napi_threaded_poll() benefits
from recent core stack improvements, to further reduce
softirq triggers.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: optimize napi_threaded_poll() vs RPS/RFS
Eric Dumazet [Fri, 21 Apr 2023 09:43:57 +0000 (09:43 +0000)]
net: optimize napi_threaded_poll() vs RPS/RFS

We use napi_threaded_poll() in order to reduce our softirq dependency.

We can add a followup of 821eba962d95 ("net: optimize napi_schedule_rps()")
to further remove the need of firing NET_RX_SOFTIRQ whenever
RPS/RFS are used.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: make napi_threaded_poll() aware of sd->defer_list
Eric Dumazet [Fri, 21 Apr 2023 09:43:56 +0000 (09:43 +0000)]
net: make napi_threaded_poll() aware of sd->defer_list

If we call skb_defer_free_flush() from napi_threaded_poll(),
we can avoid to raise IPI from skb_attempt_defer_free()
when the list becomes too big.

This allows napi_threaded_poll() to rely less on softirqs,
and lowers latency caused by a too big list.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: move skb_defer_free_flush() up
Eric Dumazet [Fri, 21 Apr 2023 09:43:55 +0000 (09:43 +0000)]
net: move skb_defer_free_flush() up

We plan using skb_defer_free_flush() from napi_threaded_poll()
in the following patch.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: do not provide hard irq safety for sd->defer_lock
Eric Dumazet [Fri, 21 Apr 2023 09:43:54 +0000 (09:43 +0000)]
net: do not provide hard irq safety for sd->defer_lock

kfree_skb() can be called from hard irq handlers,
but skb_attempt_defer_free() is meant to be used
from process or BH contexts, and skb_defer_free_flush()
is meant to be called from BH contexts.

Not having to mask hard irq can save some cycles.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: add debugging checks in skb_attempt_defer_free()
Eric Dumazet [Fri, 21 Apr 2023 09:43:53 +0000 (09:43 +0000)]
net: add debugging checks in skb_attempt_defer_free()

Make sure skbs that are stored in softnet_data.defer_list
do not have a dst attached.

Also make sure the the skb was orphaned.

Link: https://lore.kernel.org/netdev/CANn89iJuEVe72bPmEftyEJHLzzN=QNR2yueFjTxYXCEpS5S8HQ@mail.gmail.com/T/
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next...
David S. Miller [Sun, 23 Apr 2023 12:27:36 +0000 (13:27 +0100)]
Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/next-queue

Tony Nguyen says:

====================
This series lowers the CPU usage of the ice driver when using its
provided /dev/gnss*.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agonet: mtk_eth_soc: mediatek: fix ppe flow accounting for v1 hardware
Felix Fietkau [Thu, 20 Apr 2023 21:06:42 +0000 (22:06 +0100)]
net: mtk_eth_soc: mediatek: fix ppe flow accounting for v1 hardware

Older chips (like MT7622) use a different bit in ib2 to enable hardware
counter support. Add macros for both and select the appropriate bit.

Fixes: 3fbe4d8c0e53 ("net: ethernet: mtk_eth_soc: ppe: add support for flow accounting")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 months agoMerge tag 'mlx5-updates-2023-04-20' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Sat, 22 Apr 2023 03:47:04 +0000 (20:47 -0700)]
Merge tag 'mlx5-updates-2023-04-20' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2023-04-20

1) Dragos Improves RX page pool, and provides some fixes to his previous
   series:
 1.1) Fix releasing page_pool for striding RQ and legacy RQ nonlinear case
 1.2) Hook NAPIs to page pools to gain more performance.

2) From Roi, Some cleanups to TC and eswitch modules.

3) Maher migrates vnic diagnostic counters reporting from debugfs to a
    dedicated devlink health reporter

Maher Says:
===========
 net/mlx5: Expose vnic diagnostic counters using devlink

Currently, vnic diagnostic counters are exposed through the following
debugfs:

$ ls /sys/kernel/debug/mlx5/0000:08:00.0/esw/vf_0/vnic_diag/
cq_overrun
quota_exceeded_command
total_q_under_processor_handle
invalid_command
send_queue_priority_update_flow
nic_receive_steering_discard

The current design does not allow the hypervisor to view the diagnostic
counters of its VFs, in case the VFs get bound to a VM. In other words,
the counters are not exposed for representor interfaces.
Furthermore, the debugfs design is inconvenient future-wise, in case more
counters need to be reported by the driver in the future.

As these counters pertain to vNIC health, it is more appropriate to
utilize the devlink health reporter to expose them.

Thus, this patchest includes the following changes:

* Drop the current vnic diagnostic counters debugfs interface.
* Add a vnic devlink health reporter for PFs/VFs core devices, which
  when diagnosed will dump vnic diagnostic counter values that are
  queried from FW.
* Add a vnic devlink health reporter for the representor interface, which
  serves the same purpose listed in the previous point, in addition to
  allowing the hypervisor to view its VFs diagnostic counters, even when
  the VFs are bounded to external VMs.

Example of devlink health reporter usage is:
$devlink health diagnose pci/0000:08:00.0 reporter vnic
 vNIC env counters:
    total_error_queues: 0 send_queue_priority_update_flow: 0
    comp_eq_overrun: 0 async_eq_overrun: 0 cq_overrun: 0
    invalid_command: 0 quota_exceeded_command: 0
    nic_receive_steering_discard: 0

===========

4) SW steering fixes and improvements

Yevgeny Kliteynik Says:
=======================
These short patch series are just small fixes / improvements for
SW steering:

 - Patch 1: Fix dumping of legacy modify_hdr in debug dump to
   align to what is expected by parser
 - Patch 2: Have separate threshold for ICM sync per ICM type
 - Patch 3: Add more info to the steering debug dump - Linux
   version and device name
 - Patch 4: Keep track of number of buddies that are currently
   in use per domain per buddy type

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

* tag 'mlx5-updates-2023-04-20' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5: Update op_mode to op_mod for port selection
  net/mlx5: E-Switch, Remove unused mlx5_esw_offloads_vport_metadata_set()
  net/mlx5: E-Switch, Remove redundant dev arg from mlx5_esw_vport_alloc()
  net/mlx5: Include linux/pci.h for pci_msix_can_alloc_dyn()
  net/mlx5e: RX, Hook NAPIs to page pools
  net/mlx5e: RX, Fix XDP_TX page release for legacy rq nonlinear case
  net/mlx5e: RX, Fix releasing page_pool pages twice for striding RQ
  net/mlx5e: Add vnic devlink health reporter to representors
  net/mlx5: Add vnic devlink health reporter to PFs/VFs
  Revert "net/mlx5: Expose vnic diagnostic counters for eswitch managed vports"
  Revert "net/mlx5: Expose steering dropped packets counter"
  net/mlx5: DR, Add memory statistics for domain object
  net/mlx5: DR, Add more info in domain dbg dump
  net/mlx5: DR, Calculate sync threshold of each pool according to its type
  net/mlx5: DR, Fix dumping of legacy modify_hdr in debug dump
====================

Link: https://lore.kernel.org/r/20230421013850.349646-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf...
Jakub Kicinski [Sat, 22 Apr 2023 03:32:36 +0000 (20:32 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2023-04-21

We've added 71 non-merge commits during the last 8 day(s) which contain
a total of 116 files changed, 13397 insertions(+), 8896 deletions(-).

The main changes are:

1) Add a new BPF netfilter program type and minimal support to hook
   BPF programs to netfilter hooks such as prerouting or forward,
   from Florian Westphal.

2) Fix race between btf_put and btf_idr walk which caused a deadlock,
   from Alexei Starovoitov.

3) Second big batch to migrate test_verifier unit tests into test_progs
   for ease of readability and debugging, from Eduard Zingerman.

4) Add support for refcounted local kptrs to the verifier for allowing
   shared ownership, useful for adding a node to both the BPF list and
   rbtree, from Dave Marchevsky.

5) Migrate bpf_for(), bpf_for_each() and bpf_repeat() macros from BPF
  selftests into libbpf-provided bpf_helpers.h header and improve
  kfunc handling, from Andrii Nakryiko.

6) Support 64-bit pointers to kfuncs needed for archs like s390x,
   from Ilya Leoshkevich.

7) Support BPF progs under getsockopt with a NULL optval,
   from Stanislav Fomichev.

8) Improve verifier u32 scalar equality checking in order to enable
   LLVM transformations which earlier had to be disabled specifically
   for BPF backend, from Yonghong Song.

9) Extend bpftool's struct_ops object loading to support links,
   from Kui-Feng Lee.

10) Add xsk selftest follow-up fixes for hugepage allocated umem,
    from Magnus Karlsson.

11) Support BPF redirects from tc BPF to ifb devices,
    from Daniel Borkmann.

12) Add BPF support for integer type when accessing variable length
    arrays, from Feng Zhou.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (71 commits)
  selftests/bpf: verifier/value_ptr_arith converted to inline assembly
  selftests/bpf: verifier/value_illegal_alu converted to inline assembly
  selftests/bpf: verifier/unpriv converted to inline assembly
  selftests/bpf: verifier/subreg converted to inline assembly
  selftests/bpf: verifier/spin_lock converted to inline assembly
  selftests/bpf: verifier/sock converted to inline assembly
  selftests/bpf: verifier/search_pruning converted to inline assembly
  selftests/bpf: verifier/runtime_jit converted to inline assembly
  selftests/bpf: verifier/regalloc converted to inline assembly
  selftests/bpf: verifier/ref_tracking converted to inline assembly
  selftests/bpf: verifier/map_ptr_mixing converted to inline assembly
  selftests/bpf: verifier/map_in_map converted to inline assembly
  selftests/bpf: verifier/lwt converted to inline assembly
  selftests/bpf: verifier/loops1 converted to inline assembly
  selftests/bpf: verifier/jeq_infer_not_null converted to inline assembly
  selftests/bpf: verifier/direct_packet_access converted to inline assembly
  selftests/bpf: verifier/d_path converted to inline assembly
  selftests/bpf: verifier/ctx converted to inline assembly
  selftests/bpf: verifier/btf_ctx_access converted to inline assembly
  selftests/bpf: verifier/bpf_get_stack converted to inline assembly
  ...
====================

Link: https://lore.kernel.org/r/20230421211035.9111-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: dst: fix missing initialization of rt_uncached
Maxime Bizon [Thu, 20 Apr 2023 18:25:08 +0000 (20:25 +0200)]
net: dst: fix missing initialization of rt_uncached

xfrm_alloc_dst() followed by xfrm4_dst_destroy(), without a
xfrm4_fill_dst() call in between, causes the following BUG:

 BUG: spinlock bad magic on CPU#0, fbxhostapd/732
  lock: 0x890b7668, .magic: 890b7668, .owner: <none>/-1, .owner_cpu: 0
 CPU: 0 PID: 732 Comm: fbxhostapd Not tainted 6.3.0-rc6-next-20230414-00613-ge8de66369925-dirty #9
 Hardware name: Marvell Kirkwood (Flattened Device Tree)
  unwind_backtrace from show_stack+0x10/0x14
  show_stack from dump_stack_lvl+0x28/0x30
  dump_stack_lvl from do_raw_spin_lock+0x20/0x80
  do_raw_spin_lock from rt_del_uncached_list+0x30/0x64
  rt_del_uncached_list from xfrm4_dst_destroy+0x3c/0xbc
  xfrm4_dst_destroy from dst_destroy+0x5c/0xb0
  dst_destroy from rcu_process_callbacks+0xc4/0xec
  rcu_process_callbacks from __do_softirq+0xb4/0x22c
  __do_softirq from call_with_stack+0x1c/0x24
  call_with_stack from do_softirq+0x60/0x6c
  do_softirq from __local_bh_enable_ip+0xa0/0xcc

Patch "net: dst: Prevent false sharing vs. dst_entry:: __refcnt" moved
rt_uncached and rt_uncached_list fields from rtable struct to dst
struct, so they are more zeroed by memset_after(xdst, 0, u.dst) in
xfrm_alloc_dst().

Note that rt_uncached (list_head) was never properly initialized at
alloc time, but xfrm[46]_dst_destroy() is written in such a way that
it was not an issue thanks to the memset:

if (xdst->u.rt.dst.rt_uncached_list)
rt_del_uncached_list(&xdst->u.rt);

The route code does it the other way around: rt_uncached_list is
assumed to be valid IIF rt_uncached list_head is not empty:

void rt_del_uncached_list(struct rtable *rt)
{
        if (!list_empty(&rt->dst.rt_uncached)) {
                struct uncached_list *ul = rt->dst.rt_uncached_list;

                spin_lock_bh(&ul->lock);
                list_del_init(&rt->dst.rt_uncached);
                spin_unlock_bh(&ul->lock);
        }
}

This patch adds mandatory rt_uncached list_head initialization in
generic dst_init(), and adapt xfrm[46]_dst_destroy logic to match the
rest of the code.

Fixes: d288a162dd1c ("net: dst: Prevent false sharing vs. dst_entry:: __refcnt")
Reported-by: kernel test robot <oliver.sang@intel.com>
Link: https://lore.kernel.org/oe-lkp/202304162125.18b7bcdd-oliver.sang@intel.com
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
CC: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Maxime Bizon <mbizon@freebox.fr>
Link: https://lore.kernel.org/r/20230420182508.2417582-1-mbizon@freebox.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: dsa: qca8k: fix LEDS_CLASS dependency
Arnd Bergmann [Thu, 20 Apr 2023 21:36:31 +0000 (23:36 +0200)]
net: dsa: qca8k: fix LEDS_CLASS dependency

With LEDS_CLASS=m, a built-in qca8k driver fails to link:

arm-linux-gnueabi-ld: drivers/net/dsa/qca/qca8k-leds.o: in function `qca8k_setup_led_ctrl':
qca8k-leds.c:(.text+0x1ea): undefined reference to `devm_led_classdev_register_ext'

Change the dependency to avoid the broken configuration.

Fixes: 1e264f9d2918 ("net: dsa: qca8k: add LEDs basic support")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20230420213639.2243388-1-arnd@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet/handshake: Fix section mismatch in handshake_exit
Geert Uytterhoeven [Thu, 20 Apr 2023 17:37:23 +0000 (19:37 +0200)]
net/handshake: Fix section mismatch in handshake_exit

If CONFIG_NET_NS=n (e.g. m68k/defconfig):

    WARNING: modpost: vmlinux.o: section mismatch in reference: handshake_exit (section: .exit.text) -> handshake_genl_net_ops (section: .init.data)
    ERROR: modpost: Section mismatches detected.

Fix this by dropping the __net_initdata tag from handshake_genl_net_ops.

Fixes: 3b3009ea8abb713b ("net/handshake: Create a NETLINK service for handling handshake requests")
Reported-by: noreply@ellerman.id.au
Closes: http://kisskb.ellerman.id.au/kisskb/buildresult/14912987
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Link: https://lore.kernel.org/r/20230420173723.3773434-1-geert@linux-m68k.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agonet: phy: add basic driver for NXP CBTX PHY
Vladimir Oltean [Tue, 18 Apr 2023 19:01:41 +0000 (22:01 +0300)]
net: phy: add basic driver for NXP CBTX PHY

The CBTX PHY is a Fast Ethernet PHY integrated into the SJA1110 A/B/C
automotive Ethernet switches.

It was hoped it would work with the Generic PHY driver, but alas, it
doesn't. The most important reason why is that the PHY is powered down
by default, and it needs a vendor register to power it on.

It has a linear memory map that is accessed over SPI by the SJA1110
switch driver, which exposes a fake MDIO controller. It has the
following (and only the following) standard clause 22 registers:

0x0: MII_BMCR
0x1: MII_BMSR
0x2: MII_PHYSID1
0x3: MII_PHYSID2
0x4: MII_ADVERTISE
0x5: MII_LPA
0x6: MII_EXPANSION
0x7: the missing MII_NPAGE for Next Page Transmit Register

Every other register is vendor-defined.

The register map expands the standard clause 22 5-bit address space of
0x20 registers, however the driver does not need to access the extra
registers for now (and hopefully never). If it ever needs to do that, it
is possible to implement a fake (software) page switching mechanism
between the PHY driver and the SJA1110 MDIO controller driver.

Also, Auto-MDIX is turned off by default in hardware, the driver turns
it on by default and reports the current status. I've tested this with a
VSC8514 link partner and a crossover cable, by forcing the mode on the
link partner, and seeing that the CBTX PHY always sees the reverse of
the mode forced on the VSC8514 (and that traffic works). The link
doesn't come up (as expected) if MDI modes are forced on both ends in
the same way (with the cross-over cable, that is).

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20230418190141.1040562-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoselftests/bpf: verifier/value_ptr_arith converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:34 +0000 (20:42 +0300)]
selftests/bpf: verifier/value_ptr_arith converted to inline assembly

Test verifier/value_ptr_arith automatically converted to use inline assembly.

Test cases "sanitation: alu with different scalars 2" and
"sanitation: alu with different scalars 3" are updated to
avoid -ENOENT as return value, as __retval() annotation
only supports numeric literals.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-25-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/value_illegal_alu converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:33 +0000 (20:42 +0300)]
selftests/bpf: verifier/value_illegal_alu converted to inline assembly

Test verifier/value_illegal_alu automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-24-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/unpriv converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:32 +0000 (20:42 +0300)]
selftests/bpf: verifier/unpriv converted to inline assembly

Test verifier/unpriv semi-automatically converted to use inline assembly.

The verifier/unpriv.c had to be split in two parts:
- the bulk of the tests is in the progs/verifier_unpriv.c;
- the single test that needs `struct bpf_perf_event_data`
  definition is in the progs/verifier_unpriv_perf.c.

The tests above can't be in a single file because:
- first requires inclusion of the filter.h header
  (to get access to BPF_ST_MEM macro, inline assembler does
   not support this isntruction);
- the second requires vmlinux.h, which contains definitions
  conflicting with filter.h.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-23-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/subreg converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:31 +0000 (20:42 +0300)]
selftests/bpf: verifier/subreg converted to inline assembly

Test verifier/subreg automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-22-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/spin_lock converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:30 +0000 (20:42 +0300)]
selftests/bpf: verifier/spin_lock converted to inline assembly

Test verifier/spin_lock automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-21-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/sock converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:29 +0000 (20:42 +0300)]
selftests/bpf: verifier/sock converted to inline assembly

Test verifier/sock automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-20-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/search_pruning converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:28 +0000 (20:42 +0300)]
selftests/bpf: verifier/search_pruning converted to inline assembly

Test verifier/search_pruning automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-19-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/runtime_jit converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:27 +0000 (20:42 +0300)]
selftests/bpf: verifier/runtime_jit converted to inline assembly

Test verifier/runtime_jit automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-18-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/regalloc converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:26 +0000 (20:42 +0300)]
selftests/bpf: verifier/regalloc converted to inline assembly

Test verifier/regalloc automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-17-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/ref_tracking converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:25 +0000 (20:42 +0300)]
selftests/bpf: verifier/ref_tracking converted to inline assembly

Test verifier/ref_tracking automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-16-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/map_ptr_mixing converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:22 +0000 (20:42 +0300)]
selftests/bpf: verifier/map_ptr_mixing converted to inline assembly

Test verifier/map_ptr_mixing automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-13-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/map_in_map converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:21 +0000 (20:42 +0300)]
selftests/bpf: verifier/map_in_map converted to inline assembly

Test verifier/map_in_map automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-12-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/lwt converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:20 +0000 (20:42 +0300)]
selftests/bpf: verifier/lwt converted to inline assembly

Test verifier/lwt automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-11-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/loops1 converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:19 +0000 (20:42 +0300)]
selftests/bpf: verifier/loops1 converted to inline assembly

Test verifier/loops1 automatically converted to use inline assembly.

There are a few modifications for the converted tests.
"tracepoint" programs do not support test execution, change program
type to "xdp" (which supports test execution) for the following tests
that have __retval tags:
- bounded loop, count to 4
- bonded loop containing forward jump

Also, remove the __retval tag for test:
- bounded loop, count from positive unknown to 4

As it's return value is a random number.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-10-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/jeq_infer_not_null converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:18 +0000 (20:42 +0300)]
selftests/bpf: verifier/jeq_infer_not_null converted to inline assembly

Test verifier/jeq_infer_not_null automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-9-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/direct_packet_access converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:17 +0000 (20:42 +0300)]
selftests/bpf: verifier/direct_packet_access converted to inline assembly

Test verifier/direct_packet_access automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-8-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/d_path converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:16 +0000 (20:42 +0300)]
selftests/bpf: verifier/d_path converted to inline assembly

Test verifier/d_path automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-7-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/ctx converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:15 +0000 (20:42 +0300)]
selftests/bpf: verifier/ctx converted to inline assembly

Test verifier/ctx automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-6-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/btf_ctx_access converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:14 +0000 (20:42 +0300)]
selftests/bpf: verifier/btf_ctx_access converted to inline assembly

Test verifier/btf_ctx_access automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-5-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/bpf_get_stack converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:13 +0000 (20:42 +0300)]
selftests/bpf: verifier/bpf_get_stack converted to inline assembly

Test verifier/bpf_get_stack automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-4-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: verifier/bounds converted to inline assembly
Eduard Zingerman [Fri, 21 Apr 2023 17:42:12 +0000 (20:42 +0300)]
selftests/bpf: verifier/bounds converted to inline assembly

Test verifier/bounds automatically converted to use inline assembly.

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-3-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: Add notion of auxiliary programs for test_loader
Eduard Zingerman [Fri, 21 Apr 2023 17:42:11 +0000 (20:42 +0300)]
selftests/bpf: Add notion of auxiliary programs for test_loader

In order to express test cases that use bpf_tail_call() intrinsic it
is necessary to have several programs to be loaded at a time.
This commit adds __auxiliary annotation to the set of annotations
supported by test_loader.c. Programs marked as auxiliary are always
loaded but are not treated as a separate test.

For example:

    void dummy_prog1(void);

    struct {
            __uint(type, BPF_MAP_TYPE_PROG_ARRAY);
            __uint(max_entries, 4);
            __uint(key_size, sizeof(int));
            __array(values, void (void));
    } prog_map SEC(".maps") = {
            .values = {
                    [0] = (void *) &dummy_prog1,
            },
    };

    SEC("tc")
    __auxiliary
    __naked void dummy_prog1(void) {
            asm volatile ("r0 = 42; exit;");
    }

    SEC("tc")
    __description("reference tracking: check reference or tail call")
    __success __retval(0)
    __naked void check_reference_or_tail_call(void)
    {
            asm volatile (
            "r2 = %[prog_map] ll;"
            "r3 = 0;"
            "call %[bpf_tail_call];"
            "r0 = 0;"
            "exit;"
            :: __imm(bpf_tail_call),
            :  __clobber_all);
    }

Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20230421174234.2391278-2-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoMerge branch 'bpf: add netfilter program type'
Alexei Starovoitov [Fri, 21 Apr 2023 18:35:51 +0000 (11:35 -0700)]
Merge branch 'bpf: add netfilter program type'

Florian Westphal says:

====================
Changes since last version:
- rework test case in last patch wrt. ctx->skb dereference etc (Alexei)
- pacify bpf ci tests, netfilter program type missed string translation
  in libbpf helper.

This still uses runtime btf walk rather than extending
the btf trace array as Alexei suggested, I would do this later (or someone else can).

v1 cover letter:

Add minimal support to hook bpf programs to netfilter hooks, e.g.
PREROUTING or FORWARD.

For this the most relevant parts for registering a netfilter
hook via the in-kernel api are exposed to userspace via bpf_link.

The new program type is 'tracing style', i.e. there is no context
access rewrite done by verifier, the function argument (struct bpf_nf_ctx)
isn't stable.
There is no support for direct packet access, dynptr api should be used
instead.

With this its possible to build a small test program such as:

 #include "vmlinux.h"
extern int bpf_dynptr_from_skb(struct __sk_buff *skb, __u64 flags,
                               struct bpf_dynptr *ptr__uninit) __ksym;
extern void *bpf_dynptr_slice(const struct bpf_dynptr *ptr, uint32_t offset,
                                   void *buffer, uint32_t buffer__sz) __ksym;
SEC("netfilter")
int nf_test(struct bpf_nf_ctx *ctx)
{
struct nf_hook_state *state = ctx->state;
struct sk_buff *skb = ctx->skb;
const struct iphdr *iph, _iph;
const struct tcphdr *th, _th;
struct bpf_dynptr ptr;

if (bpf_dynptr_from_skb(skb, 0, &ptr))
return NF_DROP;

iph = bpf_dynptr_slice(&ptr, 0, &_iph, sizeof(_iph));
if (!iph)
return NF_DROP;

th = bpf_dynptr_slice(&ptr, iph->ihl << 2, &_th, sizeof(_th));
if (!th)
return NF_DROP;

bpf_printk("accept %x:%d->%x:%d, hook %d ifin %d\n",
   iph->saddr, bpf_ntohs(th->source), iph->daddr,
   bpf_ntohs(th->dest), state->hook, state->in->ifindex);
        return NF_ACCEPT;
}

Then, tail /sys/kernel/tracing/trace_pipe.

Changes since v3:
- uapi: remove 'reserved' struct member, s/prio/priority (Alexei)
- add ctx access test cases (Alexei, see last patch)
- some arm32 can only handle cmpxchg on u32 (build bot)
- Fix kdoc annotations (Simon Horman)
- bpftool: prefer p_err, not fprintf (Quentin)
- add test cases in separate patch

Changes since v2:
1. don't WARN when user calls 'bpftool loink detach' twice
   restrict attachment to ip+ip6 families, lets relax this
   later in case arp/bridge/netdev are needed too.
2. show netfilter links in 'bpftool net' output as well.

Changes since v1:
1. Don't fail to link when CONFIG_NETFILTER=n (build bot)
2. Use test_progs instead of test_verifier (Alexei)

Changes since last RFC version:
1. extend 'bpftool link show' to print prio/hooknum etc
2. extend 'nft list hooks' so it can print the bpf program id
3. Add an extra patch to artificially restrict bpf progs with
   same priority.  Its fine from a technical pov but it will
   cause ordering issues (most recent one comes first).
   Can be removed later.
4. Add test_run support for netfilter prog type and a small
   extension to verifier tests to make sure we can't return
   verdicts like NF_STOLEN.
5. Alter the netfilter part of the bpf_link uapi struct:
   - add flags/reserved members.
  Not used here except returning errors when they are nonzero.
  Plan is to allow the bpf_link users to enable netfilter
  defrag or conntrack engine by setting feature flags at
  link create time in the future.
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: add missing netfilter return value and ctx access tests
Florian Westphal [Fri, 21 Apr 2023 17:03:00 +0000 (19:03 +0200)]
selftests/bpf: add missing netfilter return value and ctx access tests

Extend prog_tests with two test cases:

 # ./test_progs --allow=verifier_netfilter_retcode
 #278/1   verifier_netfilter_retcode/bpf_exit with invalid return code. test1:OK
 #278/2   verifier_netfilter_retcode/bpf_exit with valid return code. test2:OK
 #278/3   verifier_netfilter_retcode/bpf_exit with valid return code. test3:OK
 #278/4   verifier_netfilter_retcode/bpf_exit with invalid return code. test4:OK
 #278     verifier_netfilter_retcode:OK

This checks that only accept and drop (0,1) are permitted.

NF_QUEUE could be implemented later if we can guarantee that attachment
of such programs can be rejected if they get attached to a pf/hook that
doesn't support async reinjection.

NF_STOLEN could be implemented via trusted helpers that can guarantee
that the skb will eventually be free'd.

v4: test case for bpf_nf_ctx access checks, requested by Alexei Starovoitov.
v5: also check ctx->{state,skb} can be dereferenced (Alexei).

 # ./test_progs --allow=verifier_netfilter_ctx
 #281/1   verifier_netfilter_ctx/netfilter invalid context access, size too short:OK
 #281/2   verifier_netfilter_ctx/netfilter invalid context access, size too short:OK
 #281/3   verifier_netfilter_ctx/netfilter invalid context access, past end of ctx:OK
 #281/4   verifier_netfilter_ctx/netfilter invalid context, write:OK
 #281/5   verifier_netfilter_ctx/netfilter valid context read and invalid write:OK
 #281/6   verifier_netfilter_ctx/netfilter test prog with skb and state read access:OK
 #281/7   verifier_netfilter_ctx/netfilter test prog with skb and state read access @unpriv:OK
 #281     verifier_netfilter_ctx:OK
Summary: 1/7 PASSED, 0 SKIPPED, 0 FAILED

This checks:
1/2: partial reads of ctx->{skb,state} are rejected
3. read access past sizeof(ctx) is rejected
4. write to ctx content, e.g. 'ctx->skb = NULL;' is rejected
5. ctx->state content cannot be altered
6. ctx->state and ctx->skb can be dereferenced
7. ... same program fails for unpriv (CAP_NET_ADMIN needed).

Link: https://lore.kernel.org/bpf/20230419021152.sjq4gttphzzy6b5f@dhcp-172-26-102-232.dhcp.thefacebook.com/
Link: https://lore.kernel.org/bpf/20230420201655.77kkgi3dh7fesoll@MacBook-Pro-6.local/
Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20230421170300.24115-8-fw@strlen.de
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agobpf: add test_run support for netfilter program type
Florian Westphal [Fri, 21 Apr 2023 17:02:59 +0000 (19:02 +0200)]
bpf: add test_run support for netfilter program type

add glue code so a bpf program can be run using userspace-provided
netfilter state and packet/skb.

Default is to use ipv4:output hook point, but this can be overridden by
userspace.  Userspace provided netfilter state is restricted, only hook and
protocol families can be overridden and only to ipv4/ipv6.

Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20230421170300.24115-7-fw@strlen.de
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agotools: bpftool: print netfilter link info
Florian Westphal [Fri, 21 Apr 2023 17:02:58 +0000 (19:02 +0200)]
tools: bpftool: print netfilter link info

Dump protocol family, hook and priority value:
$ bpftool link
2: netfilter  prog 14
        ip input prio -128
        pids install(3264)
5: netfilter  prog 14
        ip6 forward prio 21
        pids a.out(3387)
9: netfilter  prog 14
        ip prerouting prio 123
        pids a.out(5700)
10: netfilter  prog 14
        ip input prio 21
        pids test2(5701)

v2: Quentin Monnet suggested to also add 'bpftool net' support:

$ bpftool net
xdp:

tc:

flow_dissector:

netfilter:

        ip prerouting prio 21 prog_id 14
        ip input prio -128 prog_id 14
        ip input prio 21 prog_id 14
        ip forward prio 21 prog_id 14
        ip output prio 21 prog_id 14
        ip postrouting prio 21 prog_id 14

'bpftool net' only dumps netfilter link type, links are sorted by protocol
family, hook and priority.

v5: fix bpf ci failure: libbpf needs small update to prog_type_name[]
    and probe_prog_load helper.
v4: don't fail with -EOPNOTSUPP in libbpf probe_prog_load, update
    prog_type_name[] with "netfilter" entry (bpf ci)
v3: fix bpf.h copy, 'reserved' member was removed (Alexei)
    use p_err, not fprintf (Quentin)

Suggested-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/bpf/eeeaac99-9053-90c2-aa33-cc1ecb1ae9ca@isovalent.com/
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20230421170300.24115-6-fw@strlen.de
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agonetfilter: disallow bpf hook attachment at same priority
Florian Westphal [Fri, 21 Apr 2023 17:02:57 +0000 (19:02 +0200)]
netfilter: disallow bpf hook attachment at same priority

This is just to avoid ordering issues between multiple bpf programs,
this could be removed later in case it turns out to be too cautious.

bpf prog could still be shared with non-bpf hook, otherwise we'd have to
make conntrack hook registration fail just because a bpf program has
same priority.

Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20230421170300.24115-5-fw@strlen.de
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agonetfilter: nfnetlink hook: dump bpf prog id
Florian Westphal [Fri, 21 Apr 2023 17:02:56 +0000 (19:02 +0200)]
netfilter: nfnetlink hook: dump bpf prog id

This allows userspace ("nft list hooks") to show which bpf program
is attached to which hook.

Without this, user only knows bpf prog is attached at prio
x, y, z at INPUT and FORWARD, but can't tell which program is where.

v4: kdoc fixups (Simon Horman)

Link: https://lore.kernel.org/bpf/ZEELzpNCnYJuZyod@corigine.com/
Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20230421170300.24115-4-fw@strlen.de
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agobpf: minimal support for programs hooked into netfilter framework
Florian Westphal [Fri, 21 Apr 2023 17:02:55 +0000 (19:02 +0200)]
bpf: minimal support for programs hooked into netfilter framework

This adds minimal support for BPF_PROG_TYPE_NETFILTER bpf programs
that will be invoked via the NF_HOOK() points in the ip stack.

Invocation incurs an indirect call.  This is not a necessity: Its
possible to add 'DEFINE_BPF_DISPATCHER(nf_progs)' and handle the
program invocation with the same method already done for xdp progs.

This isn't done here to keep the size of this chunk down.

Verifier restricts verdicts to either DROP or ACCEPT.

Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20230421170300.24115-3-fw@strlen.de
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agobpf: add bpf_link support for BPF_NETFILTER programs
Florian Westphal [Fri, 21 Apr 2023 17:02:54 +0000 (19:02 +0200)]
bpf: add bpf_link support for BPF_NETFILTER programs

Add bpf_link support skeleton.  To keep this reviewable, no bpf program
can be invoked yet, if a program is attached only a c-stub is called and
not the actual bpf program.

Defaults to 'y' if both netfilter and bpf syscall are enabled in kconfig.

Uapi example usage:
union bpf_attr attr = { };

attr.link_create.prog_fd = progfd;
attr.link_create.attach_type = 0; /* unused */
attr.link_create.netfilter.pf = PF_INET;
attr.link_create.netfilter.hooknum = NF_INET_LOCAL_IN;
attr.link_create.netfilter.priority = -128;

err = bpf(BPF_LINK_CREATE, &attr, sizeof(attr));

... this would attach progfd to ipv4:input hook.

Such hook gets removed automatically if the calling program exits.

BPF_NETFILTER program invocation is added in followup change.

NF_HOOK_OP_BPF enum will eventually be read from nfnetlink_hook, it
allows to tell userspace which program is attached at the given hook
when user runs 'nft hook list' command rather than just the priority
and not-very-helpful 'this hook runs a bpf prog but I can't tell which
one'.

Will also be used to disallow registration of two bpf programs with
same priority in a followup patch.

v4: arm32 cmpxchg only supports 32bit operand
    s/prio/priority/
v3: restrict prog attachment to ip/ip6 for now, lets lift restrictions if
    more use cases pop up (arptables, ebtables, netdev ingress/egress etc).

Signed-off-by: Florian Westphal <fw@strlen.de>
Link: https://lore.kernel.org/r/20230421170300.24115-2-fw@strlen.de
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agobpftool: Update doc to explain struct_ops register subcommand.
Kui-Feng Lee [Thu, 20 Apr 2023 00:28:22 +0000 (17:28 -0700)]
bpftool: Update doc to explain struct_ops register subcommand.

The "struct_ops register" subcommand now allows for an optional *LINK_DIR*
to be included. This specifies the directory path where bpftool will pin
struct_ops links with the same name as their corresponding map names.

Signed-off-by: Kui-Feng Lee <kuifeng@meta.com>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/r/20230420002822.345222-2-kuifeng@meta.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agobpftool: Register struct_ops with a link.
Kui-Feng Lee [Thu, 20 Apr 2023 00:28:21 +0000 (17:28 -0700)]
bpftool: Register struct_ops with a link.

You can include an optional path after specifying the object name for the
'struct_ops register' subcommand.

Since the commit 226bc6ae6405 ("Merge branch 'Transit between BPF TCP
congestion controls.'") has been accepted, it is now possible to create a
link for a struct_ops. This can be done by defining a struct_ops in
SEC(".struct_ops.link") to make libbpf returns a real link. If we don't pin
the links before leaving bpftool, they will disappear. To instruct bpftool
to pin the links in a directory with the names of the maps, we need to
provide the path of that directory.

Signed-off-by: Kui-Feng Lee <kuifeng@meta.com>
Reviewed-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/r/20230420002822.345222-1-kuifeng@meta.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
18 months agoselftests/bpf: Verify optval=NULL case
Stanislav Fomichev [Tue, 18 Apr 2023 22:53:39 +0000 (15:53 -0700)]
selftests/bpf: Verify optval=NULL case

Make sure we get optlen exported instead of getting EFAULT.

Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230418225343.553806-3-sdf@google.com
18 months agobpf: Don't EFAULT for getsockopt with optval=NULL
Stanislav Fomichev [Tue, 18 Apr 2023 22:53:38 +0000 (15:53 -0700)]
bpf: Don't EFAULT for getsockopt with optval=NULL

Some socket options do getsockopt with optval=NULL to estimate the size
of the final buffer (which is returned via optlen). This breaks BPF
getsockopt assumptions about permitted optval buffer size. Let's enforce
these assumptions only when non-NULL optval is provided.

Fixes: 0d01da6afc54 ("bpf: implement getsockopt and setsockopt hooks")
Reported-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/ZD7Js4fj5YyI2oLd@google.com/T/#mb68daf700f87a9244a15d01d00c3f0e5b08f49f7
Link: https://lore.kernel.org/bpf/20230418225343.553806-2-sdf@google.com
18 months agoMerge tag 'wireless-next-2023-04-21' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Fri, 21 Apr 2023 14:35:51 +0000 (07:35 -0700)]
Merge tag 'wireless-next-2023-04-21' of git://git./linux/kernel/git/wireless/wireless-next

Kalle Valo says:

====================
wireless-next patches for v6.4

Most likely the last -next pull request for v6.4. We have changes all
over. rtw88 now supports SDIO bus and iwlwifi continues to work on
Wi-Fi 7 support. Not much stack changes this time.

Major changes:

cfg80211/mac80211
 - fix some Fine Time Measurement (FTM) frames not being bufferable
 - flush frames before key removal to avoid potential unencrypted
   transmission depending on the hardware design

iwlwifi
 - preparation for Wi-Fi 7 EHT and multi-link support

rtw88
 - SDIO bus support
 - RTL8822BS, RTL8822CS and RTL8821CS SDIO chipset support

rtw89
 - framework firmware backwards compatibility

brcmfmac
 - Cypress 43439 SDIO support

mt76
 - mt7921 P2P support
 - mt7996 mesh A-MSDU support
 - mt7996 EHT support
 - mt7996 coredump support

wcn36xx
 - support for pronto v3 hardware

ath11k
 - PCIe DeviceTree bindings
 - WCN6750: enable SAR support

ath10k
 - convert DeviceTree bindings to YAML

* tag 'wireless-next-2023-04-21' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (261 commits)
  wifi: rtw88: Update spelling in main.h
  wifi: airo: remove ISA_DMA_API dependency
  wifi: rtl8xxxu: Simplify setting the initial gain
  wifi: rtl8xxxu: Add rtl8xxxu_write{8,16,32}_{set,clear}
  wifi: rtl8xxxu: Don't print the vendor/product/serial
  wifi: rtw88: Fix memory leak in rtw88_usb
  wifi: rtw88: call rtw8821c_switch_rf_set() according to chip variant
  wifi: rtw88: set pkg_type correctly for specific rtw8821c variants
  wifi: rtw88: rtw8821c: Fix rfe_option field width
  wifi: rtw88: usb: fix priority queue to endpoint mapping
  wifi: rtw88: 8822c: add iface combination
  wifi: rtw88: handle station mode concurrent scan with AP mode
  wifi: rtw88: prevent scan abort with other VIFs
  wifi: rtw88: refine reserved page flow for AP mode
  wifi: rtw88: disallow PS during AP mode
  wifi: rtw88: 8822c: extend reserved page number
  wifi: rtw88: add port switch for AP mode
  wifi: rtw88: add bitmap for dynamic port settings
  wifi: rtw89: mac: use regular int as return type of DLE buffer request
  wifi: mac80211: remove return value check of debugfs_create_dir()
  ...
====================

Link: https://lore.kernel.org/r/20230421104726.800BCC433D2@smtp.kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
18 months agoselftests/xsk: Put MAP_HUGE_2MB in correct argument
Magnus Karlsson [Fri, 21 Apr 2023 06:22:08 +0000 (08:22 +0200)]
selftests/xsk: Put MAP_HUGE_2MB in correct argument

Put the flag MAP_HUGE_2MB in the correct flags argument instead of the
wrong offset argument.

Fixes: 2ddade322925 ("selftests/xsk: Fix munmap for hugepage allocated umem")
Reported-by: Kal Cutter Conley <kal.conley@dectris.com>
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230421062208.3772-1-magnus.karlsson@gmail.com
18 months agobpf: Fix bpf_refcount_acquire's refcount_t address calculation
Dave Marchevsky [Fri, 21 Apr 2023 07:44:31 +0000 (00:44 -0700)]
bpf: Fix bpf_refcount_acquire's refcount_t address calculation

When calculating the address of the refcount_t struct within a local
kptr, bpf_refcount_acquire_impl should add refcount_off bytes to the
address of the local kptr. Due to some missing parens, the function is
incorrectly adding sizeof(refcount_t) * refcount_off bytes. This patch
fixes the calculation.

Due to the incorrect calculation, bpf_refcount_acquire_impl was trying
to refcount_inc some memory well past the end of local kptrs, resulting
in kasan and refcount complaints, as reported in [0]. In that thread,
Florian and Eduard discovered that bpf selftests written in the new
style - with __success and an expected __retval, specifically - were
not actually being run. As a result, selftests added in bpf_refcount
series weren't really exercising this behavior, and thus didn't unearth
the bug.

With this fixed behavior it's safe to revert commit 7c4b96c00043
("selftests/bpf: disable program test run for progs/refcounted_kptr.c"),
this patch does so.

  [0] https://lore.kernel.org/bpf/ZEEp+j22imoN6rn9@strlen.de/

Fixes: 7c50b1cb76ac ("bpf: Add bpf_refcount_acquire kfunc")
Reported-by: Florian Westphal <fw@strlen.de>
Reported-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20230421074431.3548349-1-davemarchevsky@fb.com
18 months agobpf: Fix race between btf_put and btf_idr walk.
Alexei Starovoitov [Fri, 21 Apr 2023 01:49:01 +0000 (18:49 -0700)]
bpf: Fix race between btf_put and btf_idr walk.

Florian and Eduard reported hard dead lock:
[   58.433327]  _raw_spin_lock_irqsave+0x40/0x50
[   58.433334]  btf_put+0x43/0x90
[   58.433338]  bpf_find_btf_id+0x157/0x240
[   58.433353]  btf_parse_fields+0x921/0x11c0

This happens since btf->refcount can be 1 at the time of btf_put() and
btf_put() will call btf_free_id() which will try to grab btf_idr_lock
and will dead lock.
Avoid the issue by doing btf_put() without locking.

Fixes: 3d78417b60fb ("bpf: Add bpf_btf_find_by_name_kind() helper.")
Fixes: 1e89106da253 ("bpf: Add bpf_core_add_cands() and wire it into bpf_core_apply_relo_insn().")
Reported-by: Florian Westphal <fw@strlen.de>
Reported-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20230421014901.70908-1-alexei.starovoitov@gmail.com