The Microchip KSZ9477/KSZ9897/KSZ9567 7-Port Gigabit Ethernet Switches
support SGMII/RGMII/MII/RMII with register access via SPI, I2C, or MDIO.
This driver currently supports I2C register access but SPI or MDIO register
access can be easily added at a later time.
Tagging is not implemented and instead the active port is tracked to
avoid needing a tag to store port information.
This was tested with the imx8mm-venice-gw7901 board which has a
KSZ9897S switch with an IMX8MM FEC MAC master connected via RGMII_ID.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
If ports have their own unique MAC addrs and master has a set_promisc
function, call it so that packets will be received for ports.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Enabling promiscuous mode is necessary if FEC is the master of a DSA
switch driver where each port has their own MAC address.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Enabling promiscuous mode can be useful for DSA switches where each port
has its own MAC address.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
If the FEC is connected to a fixed-link (upstream switch port for
example) the phy_of_node should be set to the fixed-link node
so that speed and other properties can be found properly.
In addition fix a typo in the debug string.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
When using uclass_get_device* to get the FEC device we need to use
device sequence instead of index into UCLASS_ETH. In systems where for
example a I2C based DSA switch exists it will probe before the FEC
master and its ports will be registered first and have the first
indexes yet the FEC's sequence comes from the device-tree alias.
Take for example the imx8mm-venice-gw7901 board which has an i2c based
DSA switch:
u-boot=> net list
eth1 : lan1 00:0d:8d:aa:00:2f
eth2 : lan2 00:0d:8d:aa:00:30
eth3 : lan3 00:0d:8d:aa:00:31
eth4 : lan4 00:0d:8d:aa:00:32
eth0 : ethernet@30be0000 00:0d:8d:aa:00:2e active
Thus in this case uclass_get_device(UCLASS_ETH, 0, &dev) returns lan1
which is wrong but uclass_get_device_seq(UCLASS_ETH, 0, &dev) returns
ethernet@30be000 which is correct.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
The Gateworks GW7901 is an ARM based single board computer (SBC)
featuring:
- i.MX8M Mini SoC
- LPDDR4 DRAM
- eMMC FLASH
- SPI FRAM
- Gateworks System Controller (GSC)
- Atmel ATECC Crypto Authentication
- USB 2.0
- Microchip GbE Switch
- Multiple multi-protocol RS232/RS485/RS422 Serial ports
- onboard 802.11ac WiFi / BT
- microSD socket
- miniPCIe socket with PCIe, USB 2.0 and dual SIM sockets
- Wide range DC power input
- 802.3at PoE
To add support for this board:
- add dts from Linux (accepted for v5.14)
- add SPL PMIC config
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
The driver depends on DM_SPI and if it's not available (e. g. in SPL),
then we should not try to build it as this will fail.
Signed-off-by: Frieder Schrempf <frieder.schrempf@kontron.de>
Make macros actually use passed parameter instead of local variables
that happen
to be named the same as symbols in macro expansion.
Signed-off-by: Kacper Kubkowski <kkubkowski@fluence.pl>
Increase VDD_ARM to prevent timing issues as VDD_SOC is
used in OD mode. Also increase GIC clock.
Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Enable support for the fec ethernet on phyCORE-i.MX8MP.
Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
With the first redesign the debug UART had changed from
UART2 to UART1.
As the first hardware revision is considered as alpha and
will not be supported in future. The old setup will not
be preserved.
Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
This update includes eqos support and some minor changes.
Synced with kernel commit
412627f6ffe3 ("arm64: dts: imx8mp-phyboard-pollux-rdk: Add missing pinctrl entry")
Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Factor out the common node settings for dm-spl and dm-pre-reloc
and move them to imx8mp-u-boot.dtsi
Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
If reset-gpio is defined by device-tree use that if
CONFIG_PCIE_IMX_PERST_GPIO is not defined.
Note that after this the following boards which define
CONFIG_PCIE_IMX_PERST_GPIO in their board header file as well as their
device-tree should be able to remove CONFIG_PCIE_IMX_PERST_GPIO without
consequence:
- mx6sabresd
- mx6sxsabresd
- novena
- tbs2910
- vining_2000
Note that the ge_bx50v3 board uses CONFIG_PCIE_IMX_PERST_GPIO and does
not have reset-gpios defined it it's pcie node in the dt thus removing
CONFIG_PCIE_IMX_PERST_GPIO globally can't be done until that board adds
reset-gpios.
Cc: Ian Ray <ian.ray@ge.com> (maintainer:GE BX50V3 BOARD)
Cc: Sebastian Reichel <sebastian.reichel@collabora.com> (maintainer:GE BX50V3 BOARD)
Cc: Fabio Estevam <festevam@gmail.com> (maintainer:MX6SABRESD BOARD)
Cc: Marek Vasut <marex@denx.de> (maintainer:NOVENA BOARD)
Cc: Soeren Moch <smoch@web.de> (maintainer:TBS2910 BOARD)
Cc: Silvio Fricke <open-source@softing.de> (maintainer:VINING_2000 BOARD)
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Select the CMD_UNZIP option so that the 'gzwrite' command
can be used to flash .gz image into the eMMC.
Signed-off-by: Fabio Estevam <festevam@denx.de>
On the Yocto image there is a single partition and the kernel
and dtb are present in the 'boot' directory.
Change it accordingly so that the board can boot the Yocto
image by default.
Use the generic 'load' command instead, which is able to
read from an ext4 partition.
Signed-off-by: Fabio Estevam <festevam@denx.de>
Add LVDS support for DLC0700XDP21LF 7in 1024x600 display
(equivalent to the DLC-700JMGT4 with new touch controller)
Signed-off-by: Robert Jones <rjones@gateworks.com>
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
In the default pulse-skipping mode regulators that are very lightly
loaded can fail to regulate properly. Switching them to always use
continuous mode causes only around 10mW of overall system power
difference in a lightly loaded system that isn't already operating
them in continuous mode.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
In the default 'auto' mode regulators that are very lightly loaded
can be put in PFM mode and fail to regulator properly. Switching them
to always use continuous PWM mode has a neglibable affect on system
power and garuntees proper regulation under lightly loaded circumstances.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Substitutions in EOL parts changes the VDD_2P5 voltage rail such that
the previously unused VGEN6 LDO is needed in place of the lower power
VGEN5 for the GW54xx-G.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
For some time now having USB_KEYBOARD support has caused usb to be
initialized on boot. To allow for a quicker bootup we don't want this
for Ventana and don't really need USB keyboard support so remove it.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Some IPs have their accessible address space restricted by the
interconnect. Let's make sure U-Boot only ever uses the space below
the 4G address boundary (which is 3GiB big), even when the effective
available memory is bigger.
We implement board_get_usable_ram_top() for all i.MX8M SoCs, as the
whole family is affected by this.
Signed-off-by: Frieder Schrempf <frieder.schrempf@kontron.de>
The MX25V8035F is a 8Mb SPI NOR flash and the MX25R1635F is very
similar, but has twice the size (16Mb) and supports a wider supply
voltage range.
They were tested on the Kontron Electronics i.MX6UL and i.MX8MM SoMs.
Signed-off-by: Frieder Schrempf <frieder.schrempf@kontron.de>
Aside from the usual fixes and updates one visible change is the
MMC update, which fixes some lingering bugs and gives a decent speed
increase on some boards (9->19 MB/s on H6, 21->43 MB/s on A64 eMMC).
I am keeping an watchful eye on bug reports here, to spot any correctness
regressions.
Another change is finally the enablement of the first USB host port on
many boards without micro-USB (data) sockets, like the Pine64 family.
That doubles the number of usable USB ports from 1 to 2 on those boards.
Some smaller fixes, 4GB DRAM support (on the H616) and a new board (ZeroPi)
conclude this first round of changes.
Compile-tested for all 157 sunxi boards, boot-tested on Pine H64,
Pine64-LTS, OrangePi Zero 2 and BananaPi M2 Berry.
Summary:
- DT update for H3/H5/H6
- Enable first USB port on boards without micro-USB
- ZeroPi board support
- 4GB DRAM support for H616 boards
- MMC fixes and speed improvement
- some fixes
At the moment the Allwinner MMC driver parses the bus-width and
non-removable DT properties itself, in the probe() routine.
There is actually a generic function provided by the MMC framework doing
this job, also it parses more generic properties like broken-cd and
advanced transfer modes.
Drop our own code and call mmc_of_parse() instead, to get all new
features for free.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
To avoid the complexity of DMA operations (with chained descriptors), we
use repeated MMIO reads and writes to the SD_FIFO_REG, which allows us
to drain or fill the MMC data buffer FIFO very easily.
However those MMIO accesses are somewhat costly, so this limits our MMC
performance, to between 17 and 22 MB/s, but down to 9.5 MB/s on the H6
(partly due to the lower AHB1 frequency).
As it turns out we read the FIFO status register after *every* word we
read or write, which effectively doubles the number of MMIO accesses,
thus effectively more than halving our performance.
To avoid this overhead, we can make use of the FIFO level bits, which are
in the very same FIFO status registers.
So for a read request, we now can collect as many words as the FIFO
level originally indicated, and only then need to update the status
register.
We don't know for sure the size of the FIFO (and it seems to differ
across SoCs anyway), so writing is more fragile, which is why we still
use the old method for that. If we find a minimum FIFO size available on
all SoCs, we could use that, in a later optimisation.
This patch increases the eMMC read speed on a Pine64-LTS from about
22MB/s to 44 MB/s. SD card reads don't gain that much, but with 23 MB/s
we now reach the practical limit for 3.3V SD cards.
On the H6 we double our transfer speed, from 9.5 MB/s to 19.7 MB/s.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Newer SoCs have a self calibration feature, which avoids us writing hard
coded phase delay values into the controller.
Consolidate the code by avoiding unnecessary #ifdefs, and also enabling
the feature for all those newer SoCs.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
All SoCs since the Allwinner A64 (H5, H6, R40, H616) feature the so
called "new timing mode", so enable this in Kconfig for those SoCs.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Among the SoCs using the "new timing mode", only the A83T needs to
explicitly switch to that mode.
By just defining the symbol for that one odd A83T bit to 0 for any other
SoCs, we can always OR that in, and save the confusing nested #ifdefs.
Clean up the also confusing new_mode setting on the way.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Most Allwinner SoCs which use the so called "new timing mode" in their
MMC controllers actually use the double-rate PLL6/PERIPH0 clock as their
parent input clock. This is interestingly enough compensated by a hidden
"by 2" post-divider in the mod clock, so the divider and actual output
rate stay the same.
Even though for the H6 and H616 (but only for them!) we use the doubled
input clock for the divider computation, we never accounted for the
implicit post-divider, so the clock was only half the speed on those SoCs.
This didn't really matter so far, as our slow MMIO routine limits the
transfer speed anyway, but we will fix this soon.
Clean up the code around that selection, to always use the normal PLL6
(PERIPH0(1x)) clock as an input. As the rate and divider are the same,
that makes no difference.
Explain the hardware differences in a comment.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
When enabling PHYS_64BIT on 32-bit platforms, we get two warnings about
pointer casts in sunxi_mmc.c. Those are related to MMIO addresses, which
are always below 1GB on all Allwinner SoCs, so there is no problem with
anything having more than 32 bits.
Add the proper casts to make it compile cleanly.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
The delay and bus-width setup are slightly different across the
Allwinner SoC generations, and we covered this so far with some
preprocessor conditionals.
Use the more readable IS_ENABLE() instead.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
The H616 is our first supported Allwinner SoC which goes beyond the 4GB
address space "barrier", by having more than 32 address bits.
Lift the preliminary 3GB DRAM limit for the H616, and update the page
table setup on the way, to actually map that last GB as well.
As not all devices are actually capable of dealing with more than 32
bits (the DMA in the EMAC for instance), we also limit U-Boot's own
DRAM usage to 4GB on the way.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
We hardcode the pinctrl setting for the MMC controllers in boards.c,
since we need them also in the SPL, where there is no DT yet.
Add the respective setting for the H616 SoC, to enable eMMC on boards
with this SoC as well.
Also to make diagnosing this problem easier, print a warning if a board
tries to setup MMC2 pins without a respective SoC setting being defined.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jagan Teki <jagan at amarulasolutions.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec at siol.net>
ZeroPi is a new board of high performance with low cost
designed by FriendlyElec., using the Allwinner H3 SOC.
ZeroPi features
- Allwinner H3, Quad-core Cortex-A7@1.2GHz
- 256MB/512MB DDR3 RAM
- microsd slot
- 10/100/1000Mbps Ethernet
- Debug Serial Port
- DC 5V/2A power-supply
Signed-off-by: Yu-Tung Chang <mtwget@gmail.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
OrangePi PC2 board has DRAM with ODT, so enable it.
H5 SoC is also connected to voltage regulator. It's default value is
reasonable at reset, but might be too low when rebooting with a lower
voltage programmed. In order to avoid instability, enable driver for it
and set it to appropriate voltage.
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Tested-by: Andre Przywara <andre.przywara@arm.com>
[Andre: remove original ZQ value change, adjust commit message]
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Most clock factors and dividers in the H6 PLLs use a "+1 encoding",
which we were missing on two occasions.
This fixes the MMC clock setup on the H6, which could be slightly off due
to the wrong parent frequency:
mmc 2 set mod-clk req 52000000 parent 1176000000 n 2 m 12 rate 49000000
Also the CPU frequency (PLL1) was a tad too high before.
For PLL5 (DRAM) we already accounted for this +1, but in the DRAM code
itself, not in the bit field macro. Move this there to be aligned with
what the other SoCs and other PLLs do.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jernej Skrabec <jernej.skrabec@gmail.com>
