This implements to generate the MAC address of the device UUID.
The UUID is hashed to reduce the size to 3 bytes.
Ideas taken from eth_nxp_enet.c
Adding dependencies on: HWInfo and CRC
Signed-off-by: Adib Taraben <theadib@gmail.com>
first check if the chip has a hosh module to futher filter MAC addresses
if not then enabling the pass through for all multicast protocols
Signed-off-by: Adib Taraben <theadib@gmail.com>
This patch fixes an assert in net_buf_simple_add() function when the
lan_9250 driver would add previously added data to the net buffer
on receiving large (>128 bytes) packat. This fix was to use the frags
field of the net_buf structure instead.
Tested with a nRF54L15 DK and eth3 click board (with nordic connect sdk
2.8.0 and this patch).
Signed-off-by: Balaji Srinivasan <balaji.srinivasan@autostoresystem.com>
Currently, there is a small race window where we can miss an interrupt.
Right after we're done reading the RX buffer but just before decrementing
the RX counter to zero, the ENC28J60 may receive a packet. The chip will
raise an interrupt, but the line is still asserted. That means that the
callback will not be invoked since it is edge-triggered.
To avoid that, disable interrupts on the chip itself before processing
the RX buffer.
In fact, the ENC28J60 datasheet specifically says:
"After an interrupt occurs, the host controller should
clear the global enable bit for the interrupt pin before
servicing the interrupt. Clearing the enable bit will
cause the interrupt pin to return to the non-asserted
state (high). Doing so will prevent the host controller
from missing a falling edge should another interrupt
occur while the immediate interrupt is being serviced.
After the interrupt has been serviced, the global enable
bit may be restored. If an interrupt event occurred while
the previous interrupt was being processed, the act of
resetting the global enable bit will cause a new falling
edge on the interrupt pin to occur."
This is also what is being done in the Linux driver [1].
[1] https://elixir.bootlin.com/linux/v6.11.2/source/drivers/net/ethernet/microchip/enc28j60.c#L1126
Signed-off-by: Xavier Ruppen <xruppen@gmail.com>
The enc28j60 errata sheet says:
"The Receive Packet Pending Interrupt Flag
(EIR.PKTIF) does not reliably/accurately report
the status of pending packets."
"In the Interrupt Service Routine, if it is unknown if
a packet is pending and the source of the interrupt
is unknown, switch to Bank 1 and check the value
in EPKTCNT.
If polling to see if a packet is pending, check the
value in EPKTCNT."
A workaround has already been implemented inside of eth_enc28j60_rx().
But checking PKTIF before calling eth_enc28j60_rx() completely defeats
the purpose of the workaround. Do not check it.
Moreover, clearing ENC28J60_BIT_EIR_PKTIF is useless since it is
automatically cleared once all packets are read. So remove that check
and clarify comment.
Also please refer to the Linux driver [1].
[1] https://elixir.bootlin.com/linux/v6.11.2/source/drivers/net/ethernet/microchip/enc28j60.c#L1090
Signed-off-by: Xavier Ruppen <xruppen@gmail.com>
Add a get_config function for this driver as specified in the
Ethernet subsystem API. The implementation supports querying
the hardware checksum generation capabilities of the specified
GEM device instance. This prevents the transmission of packages
without a valid checksum for protocols such as ICMP, as the
hardware only supports IPv4/IPv6 TCP and UDP checksum generation.
Signed-off-by: Immo Birnbaum <mail@birnbaum.immo>
Compilation will fail if both adin2111 and adin1100 are used
at the same time.
Changing to define different unique names for the symbols
to avoid conflicts.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@gmail.com>
This PR adds support for LAN9250 spi ethernet controller.
This driver is tested on the Mikroe ETH Click 3
https://www.mikroe.com/eth-3-click
Signed-off-by: Mario Paja <mariopaja@hotmail.com>
On the interrupt handling, one thread per driver instance is involved
into monitoring the semaphor, sends inside the gpio callback triggered
by the gpio interrupt. Each time, when the link parameters are change,
the DM8806 is generating the gpio interrupt. After getting semaphor,
the application callback function which was linked during initialization
process is called to get the new link parameters with standard API
calls
Signed-off-by: Robert Slawinski <robert.slawinski1@gmail.com>
New driver for Davicom DM8806 PHY. Driver is using standar mdio API
to manage the DM8806 switch controller. Register access needs the
PHY addres or switch address to be one of five possible values, since
DM8806 has built-in five PHY's. These values should be defined in the
application .dts file. One DM8806 ethernet port must corresponds with
one ethernet PHY node with two properties for ethernet port: one for
PHY address and one for switch address - <reg> for register access from
Internal PHY Register area and <reg-switch> for register access from
Switch Per-Port Registers area. Device tree example below:
example device-tree:
dm8806_phy: ethernet-phy@0 {
reg = <2>;
reg-switch = <8>;
compatible = "davicom,dm8806-phy";
status = "okay";
davicom,interface-type = "rmii";
reset-gpio = <&gpiod 2 GPIO_ACTIVE_LOW>;
interrupt-gpio = <&gpioc 1 GPIO_ACTIVE_HIGH>;
};
Signed-off-by: Robert Slawinski <robert.slawinski1@gmail.com>
This commit is to enable Ethernet drivers support on Renesas RA
MCU, first target support is the Renesas RA8 series
Signed-off-by: Duy Nguyen <duy.nguyen.xa@renesas.com>
The ICS1894 phy AN_COMPLETE bit is latched high, this make the
BMSR first read return incorrect status of the AN state, update
one more BMSR read to ensure all latched bit is clear and BMSR
return actual status of the phy chip
Signed-off-by: Duy Nguyen <duy.nguyen.xa@renesas.com>
The _nxp_enet_dma_buffer_section is configured to __nocache area.
That makes the RxPkt performance very low (<= 50Mbps). By defining
it to the cacheable area, the RxPkt performance is >10x better.
Example with Zperf:
DUT command - zperf udp download 5001
PC command - iperf -u -c 192.0.2.1 -p 5001 -b 800M
Signed-off-by: Shrek Wang <shrek.wang@nxp.com>
- Move selection of CONFIG_PINCTRL from soc to individual
drivers
- in accordance with issue #78619
Signed-off-by: McAtee Maxwell <maxwell.mcatee@infineon.com>
Not all STM32 series support Zephyr MDIO API yet, while the API is enabled
by default.
To preserve compatibility, put MDIO API related code under the condition
of "st,stm32-mdio" compatible enablement.
Signed-off-by: Erwan Gouriou <erwan.gouriou@st.com>
Previously, ipv6 being disabled was a dependency of the hw acceleration
of the checksums for the nxp enet driver, because this ethernet has an
errata causing icmpv6 checksum to not be supported. Now, there is a new
config type in ethernet api for checksum types, so we can re-enable
hardware acceleration for ipv6 by implementing this type in the
get_config api in this driver.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
To remove CONFIG_PINCTRL from board side for numaker boards.
The Drivers using Pinctrl should be turning Pinctrl on
instead of the responsibility of the board.
Fixes#78619
Signed-off-by: cyliang tw <cyliang@nuvoton.com>
Improve type safety and consistency by adjusting variable and
parameter types to avoid signed/unsigned comparisons and implicit casts.
Moreover, explicit casts were applied when converting from
`size_t` to `uint16_t`.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
This change marks `w5500_api_funcs`, an instance of
the `ethernet_api`, as `const`.
By using `const`, we ensure immutability, leading to usage of only
`.rodata` and a reduction in the `.data` area.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
The newest AN1760 application note - Revision F (DS60001760G - June 2024)
is recommending to not write the node count to PLCA_CTRL1 register when
the node is not the PLCA coordinator (i.e. its ID is not zero).
Signed-off-by: Lukasz Majewski <lukma@denx.de>
This patch brings update of the procedure to initially configure the
LAN865x devices. It follows setup guidelines from newest AN1760 [*].
The values from "TABLE1" on the [*] must be written to the device in the
indicated order with recommended values.
This was not the case previously, as first values from in-flash allocated
(const) table were written and only afterwards calculated configuration
parameters (cfgparams) were updated.
With this patch the lan865x_conf[] table is allocated in-RAM, so
placeholder values can be updated and it can be written at once at the
end of configuration process.
Its single entry has been reduced from 8B to only 4B. Moreover, moving
it out of flash saves 512B of flash memory.
Note:
[*] - AN1760 Revision F (DS60001760G - June 2024)
Signed-off-by: Lukasz Majewski <lukma@denx.de>
This reverts commit 0036b8bf21.
The reverted commit causes a compile error with the STM32F2 because
there are some variables used in the file eth_stm32_hal that are
not defined in the HAL module of the STM32F2 series,
such as 'ETH_RX_DESC_CNT' and 'HAL_ETH_MII_MODE'
Signed-off-by: IBEN EL HADJ MESSAOUD Marwa <marwa.ibenelhadjmessaoud-ext@st.com>
Still mark the iface as down after ethernet_init, but then actually
check the link state and initialize carrier appropriately
This fixes the case where, the phy driver doesn't give a callback after
iface init due to the link already being up, there was no change from
the phy driver perspective, so callback wouldn't happen, and therefore
the interface could remain marked as down after boot even if carrier is up.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
This commit moves one function before another, to make the diff of the
next commit clearer of what it's doing.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
Use net_if_carrier_off during iface init instead of net_eth_carrier_off,
to immediately mark net if as down
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
In some circumstances the struct eth_stm32_tx_context object that was
allocated on eth_tx's stack is still referenced after the function exits.
This usually happens when the network is disconnected, depending on the
PHY hardware.
When the network is reconnected there will eventually be a call to
HAL_ETH_ReleaseTxPacket, which calls HAL_ETH_TxFreeCallback with the
(now invalid) pointer to the tx context. When HAL_ETH_TxFreeCallback
tries to dereference that pointer we get a bus error.
Fix this by allocating struct eth_stm32_tx_context objects from a
static array, similarly to how the buffers are allocated. This ensures
that they remain valid until the HAL is finished with them.
Fixes: #79037
Signed-off-by: Kevin ORourke <kevin.orourke@ferroamp.se>
This Kconfig has wrongly been added to defconfig files. It is not the
right place for it. It has never been the right place for it. Drivers
that need it should select the symbol in their Kconfig entries. Drop
PINCTL from Kconfig.defconfig and add proper select at Kconfig.sam*.
Fixes#78619
Signed-off-by: Gerson Fernando Budke <nandojve@gmail.com>
I for the life of me do not know what is going on here with the RNB chip
but it seems this override must be set in order for the chip to work,
regardless of strap-in configuration, and if not set explicitly, the
value after a reset for these two bits will be seemingly random and
inconsistent. And it was working before by luck before removing a second
redundant reset in a recent commit, because apparently the register
was getting the opposite of the reset value according to the datasheet
which makes it work. The result of these bits after reset seem to vary
depending on host mcu, board, debugger, number of times reset, type of
reset, and with a pinch of random chance after keeping all variables
seemingly the same, so let's just set it to the value that works
explicitly, even if it doesn't make sense. The bit here doesn't have
clear documentation but it seems it's for using RMII regardless of the
strap in option, which is what we want to do anyways if we know the
interface type from DT, so I think it's fine, considering it is making
this driver work again.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
500 ms reset time is only for software reset and comes from IEEE spec.
Datasheet mentions for hardware reset the assertion of the signal should
only need to be 500 us, and 100 us after deassert to wait to access
programming interface.
Also remove an unused macro.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
Replace "CONFIG_SOC_SERIES_STM32H7X || CONFIG_SOC_SERIES_STM32H5X"
by the compatble st_stm32h7_ethernet
Signed-off-by: IBEN EL HADJ MESSAOUD Marwa <marwa.ibenelhadjmessaoud-ext@st.com>
If OA read fails, then the essential thread is terminated.
Correct the behavior so that the driver re-tries the read and
continues to work.
Refactor offload thread OA and Generic SPI parts into
if/else statement.
Add missing `is_adin2111` to OA port 2 RX status check.
Signed-off-by: Georgij Cernysiov <geo.cgv@gmail.com>
The Drivers using Pinctrl should be turning Pinctrl on
this should not be the responsibility of the board. This
commit removes CONFIG_PINCTRL from the boards side for nxp boards.
Signed-off-by: Emilio Benavente <emilio.benavente@nxp.com>
