There is no need to perform the endian twice here.
Signed-off-by: Jacky Chou <jacky_chou@aspeedtech.com>
Reviewed-by: Dan Carpenter <dan.carpenter@linaro.org>
shrink the first argument of log_msg_ret(), add dev_xxx() functions for
error reporting.
Fixes: 9d8f78a2a79f7 ("net: add hifemac Ethernet driver for HiSilicon platform")
Signed-off-by: Yang Xiwen <forbidden405@outlook.com>
The initial commit used log_msg_ret() wrongly. Fix that by moving error
report to a separate dev_err() call and shrink the first argument of
log_msg_ret() to no more than 4 chars.
Fixes: 6b5c8d98e204 ("net: add hifemac_mdio MDIO bus driver for HiSilicon platform")
Signed-off-by: Yang Xiwen <forbidden405@outlook.com>
As with fixed-link phy device, the NC-SI phy devive does not
require an mdio bus. So, a condition is added to check the
NC-SI phy id to avoid accessing the bus pointer that is NULL.
Signed-off-by: Jacky Chou <jacky_chou@aspeedtech.com>
Booting R-Car Gen3 arm64 U-Boot with CONFIG_UBSAN=y resulted in:
=====================================================================
UBSAN: Undefined behaviour in drivers/net/phy/phy.c:728:19
left shift of 1 by 31 places cannot be represented in type 'int'
=====================================================================
Fix it by appending the UL suffix to the numeric literal. While at it,
convert the type of "addr" variable from signed to unsigned, to protect
against shifting the numeric literal by a negative value (which would
lead to yet another undefined behavior).
Fixes: 1adb406b0141 ("phy: add phy_find_by_mask/phy_connect_dev")
Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
* Using U-suffix for integer is sufficient.
* ffs() of non-zero value cannot be 0. But addr being unsigned is
* preferable.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Configure LEDs on BCM54210E so they would blink on activity
and indicate link speed. Without this the LEDs are always on
if cable is plugged in.
Signed-off-by: Marek Vasut <marex@denx.de>
The issue occurs the UAF (use-after-free) to cause double free
when do the realloc function for the pointers during the
reinitialization NC-SI process, and it will cause the memory
management occurs error.
So, nullify these pointers after free.
Signed-off-by: Jacky Chou <jacky_chou@aspeedtech.com>
The different macros use writel which is defined in asm/io.h, so let's
include the header so users of hardware.h do not need to include
asm/io.h as well.
While at it, remove asm/io.h includes wherever
asm/arch-rockchip/hardware.h is included already.
Cc: Quentin Schulz <foss+uboot@0leil.net>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Quentin Schulz <quentin.schulz@theobroma-systems.com>
In case the PHY is fully described in DT, use PHY MDIO address
from DT directly instead of always using auto-detection. This
also fixes the behavior of 'mdio list' in such DT setup, which
now prints the PHY connected to the MAC correctly.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com>
Add a new Kconfig symbol MDIO_TI_CPSW for the CPSW MDIO
driver and build it with proper DM support if enabled.
If MDIO_TI_CPSW is not enabled then we continue to
behave like before.
Clean up MDIO custom handling in am65-cpsw and use
dm_eth_phy_connect() to get the PHY.
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Tested-by: Ravi Gunasekaran <r-gunasekaran@ti.com>
The mv88e6xxx driver does not currently initialize the smi_addr field, but
instead keeps the default zero value. This leads to driver being unusable
on devices where the switch is not on address zero of the mdio bus. Fix
this problem by reading the SMI address from device tree.
Signed-off-by: Marek Mojík <marek.mojik@nic.cz>
Reviewed-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Rename RMOBILE_CPU_TYPE_* to RENESAS_CPU_TYPE_* because all
the chips are made by Renesas, while only a subset of them is
from the R-Mobile line.
Use the following command to perform the rename:
"
$ git grep -l '\<RMOBILE_CPU_TYPE_[A-Z0-9]\+\>' | \
xargs -I {} sed -i 's@\<RMOBILE\(_CPU_TYPE_[A-Z0-9]\+\)\>@RENESAS\1@g' {}
"
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com>
Rename rmobile_get_cpu_type() to renesas_get_cpu_type() because
all the chips are made by Renesas, while only a subset of them
is from the R-Mobile line.
Use the following command to perform the rename:
"
$ git grep -l '\<rmobile_get_cpu_type\>' | \
xargs -I {} sed -i 's@\<rmobile_get_cpu_type\>@renesas_get_cpu_type@g' {}
"
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com>
The actual driver does not work when there is no linked PHY. These
changes add support for fixed-link feature in the device tree.
Signed-off-by: BELOUARGA Mohamed <m.belouarga@technologyandstrategy.com>
Some platforms (such as the Lichee Pi 4A) have their dwmac device
addressable only in high memory space. Storing the node's base address
on 32 bits is not possible in such case.
Use platform's physical address type to store the base address.
Signed-off-by: Nils Le Roux <gilbsgilbert@gmail.com>
Cc: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Rename nxp_c45_tja11xx structure to nxp_c45_tja1103. The driver will
support more PHYs and nxp_c45_tja11xx is too generic.
Signed-off-by: "Radu Pirea (NXP OSS)" <radu-nicolae.pirea@oss.nxp.com>
Read PHY speed from hardware instead of assuming 100Mbps by default.
The TJA1103 works only at 100Mbps, but the driver will support more PHYs.
Signed-off-by: "Radu Pirea (NXP OSS)" <radu-nicolae.pirea@oss.nxp.com>
Use a local definition for the PHY features. PHY_100BT1_FEATURES are
not defined using the 100BaseT1 bit, so keep this workaround in the driver.
Signed-off-by: "Radu Pirea (NXP OSS)" <radu-nicolae.pirea@oss.nxp.com>
Some ethernet PHY require being reset before a phy-id can be read back
on the MDIO bus. This can result in the following message being show
on e.g. a Radxa ROCK Pi E v1.21 with a RTL8211F ethernet PHY.
Could not get PHY for ethernet@ff540000: addr -1
Add support to designware ethernet driver to reset eth phy by calling
the eth phy uclass function eth_phy_set_mdio_bus(). The call use NULL
as bus parameter to not set a shared mdio bus reference that would be
freed when probe fails. Also add a eth_phy_get_addr() call to try and
get the phy addr from DT when DM_MDIO is disabled.
This help fix ethernet on Radxa ROCK Pi E v1.21:
=> mdio list
ethernet@ff540000:
1 - RealTek RTL8211F <--> ethernet@ff540000
Reported-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
This ports the pad drive strength register configuration which can be
already found in the Linux driver for this PHY.
Signed-off-by: Lukasz Tekieli <tekieli.lukasz@gmail.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
Make the "phy-handle" property optional, which allows support
for a fixed-link phy configuration.
Thus if the "phy-handle" is present in a DT, then driver will work as
before. Otherwise, phyaddr initialization will not be necessary,
as it is not needed in case of a fixed-link config.
Signed-off-by: Maksim Kiselev <bigunclemax@gmail.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
debug print delays reset of the driver. Finally I see
bunch of "rx error FFFF" errors in the screen. CI can
not handle many prints. While network works fine there
Reproduced with:
make CROSS_COMPILE=sh2-linux- r2dplus_defconfig all
qemu-system-sh4 -M r2d -nographic -serial null \
-serial mon:stdio -net user,tftp=`pwd` \
-net nic,model=rtl8139 -kernel ./u-boot.bin
Signed-off-by: Maxim Uvarov <maxim.uvarov@linaro.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
return value of smc911x_send is ignored, but on sucesseful
send we need return 0 and or error -ETIMEOUT, not opposite.
Signed-off-by: Maxim Uvarov <maxim.uvarov@linaro.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
In order to make it easier to move on to dropping common.h from code
directly, remove common.h inclusion from the rest of the header file
which had been including it.
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Port the mdio-mux-meson-gxl.c Linux driver introduced in [1],
and adapt it to U-Boot.
This driver is needed to boot U-Boot with Linux DT since v6.4,
since it switched the MDIO mux from the mmio to a proper GXL driver.
[1] 9a24e1ff4326 ("net: mdio: add amlogic gxl mdio mux support")
Link: https://lore.kernel.org/r/20231213-u-boot-gxl-mdio-mux-v2-1-c56bb02a75ea@linaro.org
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Add support for DWC EQoS MAC on i.MX93.
Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
As per the xlnx,axi-ethernet-1.00.a DT documentation in linux, the AXI
DMA registers can be obtained via the reg property or via a separate
node for the axistream DMA controller. Currently only the latter is
supported, so add support to fetch the DMA controller registers from the
"reg" property.
Signed-off-by: Mayuresh Chitale <mchitale@ventanamicro.com>
Reviewed-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/20231116164024.139934-1-mchitale@ventanamicro.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
DM_GPIO is always enable in U-Boot proper for ARCH_SUNXI, and this
driver is never enabled in SPL, so the condition is always true.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Older DesignWare Ethernet MAC versions that this driver supports can
only work with 32-bit DMA source/destination addresses. Some platforms
have no physical RAM at the lowest 4GB address space. For these
platforms the driver must translate DMA addresses to/from physical
memory addresses.
Call translation routines so that properly configured platforms can use
the DesignWare Ethernet MAC. For platforms using device-tree this
usually means adding dma-ranges property to the bus the device node is
in.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
e1000_eth_ids holds compatible strings for e1000 devices, but it
is meaningless as e1000 is a PCI device and there is no such
compatible string assigned to e1000 by the DT bindings community.
Drop it.
Signed-off-by: Bin Meng <bmeng@tinylab.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Setting the clock delay from the device tree settings
rx-internal-delay-ps and tx-internal-delay-ps was broken:
- The expected value in the device tree is suppose to be a
delay in picoseconds, but the driver only allowed an array index.
- Driver converted this array index to the actual delay in
picoseconds and tried to apply this in the device register. This
however is not a valid register value. The actual logic here was
reversed, it converted an register representation of the delay to
the device tree delay in picoseconds.
Only when the internal delays were NOT configured in the device tree
and they default value of 7 (=2000ps) was used, a valid value was
loaded in the register.
Signed-off-by: Frank de Brabander <debrabander@gmail.com>
It adds the driver for the internal MDIO bus of HIFEMAC Ethernet
controller. It's based on the mainstream linux driver.
Signed-off-by: Yang Xiwen <forbidden405@outlook.com>
It adds the driver for HIFEMAC Ethernet controller found on HiSilicon
SoCs like Hi3798MV200. It's based on the mainstream linux driver, but
quite a lot of code gets rewritten and cleaned up to adopt u-boot driver
model.
Signed-off-by: Yang Xiwen <forbidden405@outlook.com>
Marvell LinkStreet switches support Clause 45 MDIO on the internal bus.
C45 read or writes require the register address to be written first to
the SMI PHY Data register, and then a special C45 Write Address Register
OP is used on the SMI PHY Register before making a C45 Read Data Register
OP and being able to actually read the register.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Driver is currently defining the mask and bit shifting itself,
there is no need for that as U-Boot has generic bitfield macros that help
us achieve the same result but in a cleaner way.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
