Pin 23 on South bridge does not belong to the rgmii group. It belongs to
a separate group which can have 3 functions.
Due to this the fix also have to update the way the functions are
managed. Until now each groups used NB_FUNCS(which was 2) functions. For
the mpp23, 3 functions are available but it is the only group which needs
it, so on the loop involving NB_FUNCS an extra test was added to handle
only the functions added.
The bug was visible when the gpio regulator used the gpio 23, the whole
rgmii group was setup to gpio which broke the Ethernet support on the
Armada 3720 DB board. Thanks to this patch, the UHS SD cards (which need
the vqmmc) _and_ the Ethernet work again.
Reviewed-on: http://vgitil04.il.marvell.com:8080/43284
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
On the south bridge we have pin from 0 to 29, so it gives 30 pins (and
not 29).
Reviewed-on: http://vgitil04.il.marvell.com:8080/43285
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Currently, ATU (address translation unit) implementation doesn't
support translate addresses > 32 bits.
This patch allows to configure ATU correctly for different
memory accesses (memory, configuration and IO).
The same approach is used in Linux Kernel.
Signed-off-by: Igal Liberman <igall@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Use dm_pci_map_bar function for BAR mapping. This has the advantage
of clearing BAR flags and and only accepting mapped memory.
Signed-off-by: Bernhard Messerklinger <bernhard.messerklinger@br-automation.com>
Reviewed-by: Hannes Schmelzer <hannes.schmelzer@br-automation.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
PCI enumeration may happen very early on an x86 board. The board
information pointer should have been checked in decode_regions()
as its space may not be allocated yet.
With this commit, Intel Galileo board boots again.
Fixes: 664758c ("pci: Fix decode regions for memory banks")
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
nsd32_mmc was created to support ftsdc010 dm.
It is not necessary to separate both, so merge it
to ftsdc010.
Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>
Convert CONFIG_FTSDC010_SDIO to Kconfig.
So CONFIG_FTSDC010_SDIO can also be
removed from config_whitelist.txt.
Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>
Only three defconfig(adp-ag101p_defconfig,
adp-ae3xx_defconfig, nx25-ae250_defconfig)
set CONFIG_FTSDC010=y. And they all also
enable CONFIG_DM_MMC. So the non-dm code
of ftsdc010 can be dropped now.
Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>
ftsdc010 dm driver has been disable High-Speed mode
as default to work around Andes AE3XX platform's problem,
because of it does not support High-Speed mode in
commit id 73cd56b2df.
But other platforms or SoCs maybe support this function.
So High-Speed mode can be enabled from dts with
cap-mmc-highspeed or cap-sd-highspeed property.
Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
The Rockchip-specific SDHCI wrapper does not process the 'bus-width'
property in the SDHCI node. Consequently, the bus is always kept in
4bit mode, even if 8bit wide operation is available, supported and
requested in the DTS.
This change adds processing of the 'bus-width' property and sets the
host capability flag for an 8bit wide bus, if set to 8. As the logic
in sdhci.c does not support clearing the 4bit capability, we assume
that 4bit operation is always supported.
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
The patch set dpll settings for 300MHz to values used by binary
blob[1]. With new values dpll still generate 300MHz clock, but
EMAC work. Probably with new values dpll generate more stable clock.
dpll on rk3188 provide clocks to DDR and EMAC. With current
dpll settings EMAC doesn't work on radxa rock. EMAC sends packets
to network, but it doesn't receive anything. ifconfig shows a lot
of framing errors.
[1] https://github.com/linux-rockchip/u-boot-rockchip/blob/u-boot-rk3288/
tools/rk_tools/3188_LPDDR2_300MHz_DDR3_300MHz_20130830.bin
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
The error checking should also catch the case that no range has beeen
defined.
syscon_get_first_range() returns NULL if no range is defined.
cf. rk3399_mipi.c
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Pointers are never negative. Use macro IS_ERR() for checking.
cf. rk3288_mipi.c
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Assigning f_rkusb->reboot_flag twice doesn't make sense.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
If we want to check the duration we need to increment the counter.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Clean the iomux definitions at grf_rk3188.h, and move them into
pinctrl-driver for resolving the compiling error of redefinition.
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Clean the iomux definitions at grf_rk3036.h, and move them into
pinctrl-driver for resolving the compiling error of redefinition.
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
This adds support for the (to date unsupported) I2C controllers 1~4
and 6~7 (i.e. now all controllers except I2C5, which is not accessible
on the RK3399-Q7, are supported by pinctrl).
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Klaus Goger <klaus.goger@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
The ti,pindir-d0-out-d1-in property is not expected to have a value
according to the device-tree binding, so treat it as a boolean not a
uint property.
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed-by: Jagan Teki <jagan@openedev.com>
This patch print pl clocks (pl0...pl3) and watchdog
clock using clk dump.
Signed-off-by: Vipul Kumar <vipulk@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
NAND erase was not happening for size 1GiB or more. Erase
command was executing successfully but in actual, it was not
erasing.
This patch fixed erase issue for 1 GiB or more size nand.
Signed-off-by: Vipul Kumar <vipulk@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
This IP can be found on Zynq and ZynqMP devices.
The driver was tested with reset-on-timeout; feature.
Also adding WATCHDOG symbol to Kconfig because it is required.
Signed-off-by: Shreenidhi Shedi <imshedi@gmail.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
watchdog clock is also connected to cpu 1X clocksource.
Zynq> clk dump
...
Before:
swdt 4294967290
After:
swdt 111111110
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
This patch bypasses phy detection logic for GMII interface
and just depend on phy address received from DT. This patch
is required as phy detection logic is different for some phys
like xilinx phy which can be connected over SGMII and GMII
interface.
This fixes the issue of ethernet failures when xilinx phy is
connected over GMII interface.
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Check !desc earlier to simplify code.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
Reviewed-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
With DM enabled in SPL, DM_FLAG_PRE_RELOC is required for
the omap_gpio driver to be bound to the gpio devices.
Therefore, add DM_FLAG_PRE_RELOC flag to the omap_gpio driver.
Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Last user of this driver went away in October 2014 in
commit d58a9451e7 ("ppc/arm: zap EMK boards").
Signed-off-by: Tuomas Tynkkynen <tuomas@tuxera.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
In
if (a > =0) {...}
else (a < 0) {...}
the second logical constraint is superfluous.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
In
int ret = A;
ret = B;
the first assignment has not effect.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
The lan75xx and lan78xx drivers need to drive their phy via the generic
phylib framework. Let's reflect that dependency in Kconfig, so that we
don't get build errors when phylib does not get selected.
Signed-off-by: Alexander Graf <agraf@suse.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Current Cortina phy driver assumes that firmware upload
is required during initialization and is dependent
on presence of corresponding macros like CONFIG_CORTINA_FW_ADDR
for compilation.
But Cortina phy has provision to store phy firmware in
attached dedicated EEPROM. And boards designed with such
EEPROM does not require firmware upload.
Add CORTINA_NO_FW_UPLOAD option in cortina.c to support
such boards.
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Update get_phy_id() implementation in cortina.c to check
for Cortina_phy by comparing device phy_id with cortina phy_id
instead of relying on presence of CORTINA macros.
This will allow get_phy_id to work with non-cortina phy devices
which might have same phy address as Cortina device but on
different bus.
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
pfe_command provides command line support for several features that
support pfe, like starting or stopping the pfe, checking the health
of the processor engines and checking status of different units inside
pfe.
Signed-off-by: Calvin Johnson <calvin.johnson@nxp.com>
Signed-off-by: Anjaneyulu Jagarlmudi <anji.jagarlmudi@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
This patch adds PFE driver to U-Boot
Following are the main driver files:-
pfe_hw.c: provides low level helper functions to initialize PFE
internal processor engines and other hardware blocks
pfe_driver.c: provides initialization functions
and packet send and receive functions
pfe_eth.c: provides high level gemac initialization functions
pfe_firmware.c: provides functions to load firmware into PFE
internal processor engines.
pfe_mdio.c: provides functions to initialize phy and mdio.
Signed-off-by: Calvin Johnson <calvin.johnson@nxp.com>
Signed-off-by: Anjaneyulu Jagarlmudi <anji.jagarlmudi@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
With CONFIG_DM_ETH enabled, aquantia driver compilation fails with
below error. This patch fixes the issue by including dm.h.
drivers/net/phy/aquantia.c: In function ‘aquantia_startup’:
drivers/net/phy/aquantia.c:73:21: error: dereferencing pointer to
incomplete
type ‘struct udevice’
phydev->dev->name);
^~
Signed-off-by: Calvin Johnson <calvin.johnson@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
After allocating plat the pointer is checked.
Afterwards name is allocated and not checked.
Add the missing check to avoid a possible NULL dereference.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Setting PHY_RTL8211E_PINE64_GIGABIT_FIX forces internal rx/tx delays off
on the PHY, as well as flipping some magical undocumented bits. The
magic number comes from the Pine64 engineering team, presumably as a
proxy from Realtek. This configuration fixes the throughput on some
Pine64 models. Packet loss of up to 60-70% has been observed without
this.
Signed-off-by: Kyle Evans <kevans@FreeBSD.org>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Current DW SPI driver uses 32 bit access for some registers and
16 bit access for others. So if DW SPI IP is connected via bus
which doesn't support 16 bit access we will get bus error.
Fix that by switching to 32 bit access only instead of 16 and 32 bit mix
Additional Documentation to Support this Change:
The DW_apb_ssi databook states:
"All registers in the DW_apb_ssi are addressed at 32-bit boundaries
to remain consistent with the AHB bus. Where the physical size of
any register is less than 32-bits wide, the upper unused bits of
the 32-bit boundary are reserved. Writing to these bits has no
effect; reading from these bits returns 0." [1]
[1] Section 6.1 of dw_apb_ssi.pdf (version 3.22a)
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
DW SPI internal chip select management has limitation:
it hold CS line in active state only when the FIFO is not
empty. If the FIFO freed before we add new data the SPI transaction will
be broken.
So add option to use external gpio for chip select. Gpio can be added
via device tree using standard gpio bindings.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
There is no sense in waiting for RX data in dw_reader function:
there is no chance that RX data will appear in RX FIFO if
RX FIFO is empty after previous TX write in dw_writer function.
So get rid of this waiting. After that we can get rid of dw_reader
return value and make it returning void. After that we can get rid
of dw_reader return value check in poll_transfer function.
With these changes we're getting closer to Linux DW SPI driver.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
In current implementation we get -ETIMEDOUT error when we try to use
transmit only mode (SPI_TMOD_TO)
This happens because in transmit only mode input FIFO never gets any data
which breaks our logic in dw_reader(): we are waiting until RX data will be
ready in dw_reader, but this newer happens, so we return with error.
Fix that by using SPI_TMOD_TR instead of SPI_TMOD_TO which allows to use
RX FIFO.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
In current implementation if some data still exists in Tx FIFO it
can be silently flushed, i.e. dropped on disabling of the controller,
which happens when writing 0 to DW_SPI_SSIENR (it happens in the
beginning of new transfer)
So add wait for current transmit operation to complete to be sure
that current transmit operation is finished before new one.
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
Make sure the user is notified instead of silently returning an error.
Signed-off-by: Liam Beguin <liambeguin@gmail.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
ppc4xx support was removed some time ago. Lets remove the now unused
NAND driver and all its references for this platform as well.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: Scott Wood <oss@buserror.net>
add RCC clock driver for STMP32MP157
- base on driver model = UCLASS_CLK
- support ops to enable, disable and get rate
of all SOC clock needed by U-Boot
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
- add the 2 new compatible used by STM32MP157
"st,stm32mp157-pinctrl"
"st,stm32mp157-z-pinctrl"
- update the mask for the port
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
This driver implements register read/write operations for STPMU1.
The STPMU1 PMIC provides 4 BUCKs, 6 LDOs, 1 VREF
and 2 power switches. It is accessed via an I2C interface.
This device is used with STM32MP1 SoCs.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
- add new arch stm32mp for STM32 MPU/Soc based on Cortex A
- support for stm32mp157 SOC
- SPL is used as first boot stage loader
- using driver model for all the drivers, even in SPL
- all security feature are deactivated (ETZC and TZC)
- reused STM32 MCU drivers when it is possible
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Kmalloc is using memalign allocation function. It is not necessary to
align this structure so to save bytes, we move to calloc.
And kmalloc function can't be used in SPL early stage (in board_init_f())
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Update the GPIO driver to support a live device tree.
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Add option to include RESET driver and uclass in SPL.
That can be useful to handle IP reset with same driver
in U-Boot and in SPL.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Add simple and meaningful kconfig option for pmic_bus.c
instead of using MACH type on Makefile.
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
This patch increases timeout to 2s.
It was seen on 2 USB devices (Verbatim STORE N GO 070B4AED0FB22358 and
USB DISK 2.0 9000729BA41DDF40) that the request sense command takes
between 1.3s and and 1.5s.
Signed-off-by: Christophe Kerello <christophe.kerello@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
This patch adds an interface to disable the power in dwc2 driver.
This new interface is called when the device is removed.
Signed-off-by: Christophe Kerello <christophe.kerello@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Factorize PHY get/init/poweron and PHY poweroff/exit operations
into separate function, it simplify the error path.
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Add generic_phy_power_on() and generic_phy_power_off()
calls to switch ON/OFF phy during probe and remove functions.
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Factorize PHY get/init/poweron and PHY poweroff/exit operations
into separate function, it simplify the error path.
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
ppc4xx support was removed some time ago. Lets remove the now unused
EHCI driver and all its references for this platform as well.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: Marek Vasut <marex@denx.de>
DWC3 USB3 controllers will need USB3 PHY to be enabled, in addition to
USB2 PHY, to be functional. Therefore enable USB3 PHY when available.
Signed-off-by: Vignesh R <vigneshr@ti.com>
Refractor PHY get/init/poweron and PHY poweroff/exit operations into
separate function so that its easy to support multiple PHYs.
Signed-off-by: Vignesh R <vigneshr@ti.com>
It is wrong that expect .phy_init() to also power on the PHY. Therefore,
explicitly, call generic_phy_power_on() after generic_phy_power_init() in
order to power on PHY before using it.
Signed-off-by: Vignesh R <vigneshr@ti.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
If we use hardware with very small RAM (let's consider just a couple
of hundreds of kB but not megabytes) it is not super convenient to lose
64kB for statically allocated bufer which most probably won't be used
as big as it is. Typically we'll have much shorter data packages to
excahnge and in the worst case longer packets will be split on separate
transactions.
For those corner-cases user will be able to set his buffer size of
choice via USB_DWC2_BUFFER_SIZE option in menuconfig.
By default we'll use 64 kB as it was hard-coeded before so existing
users shouldn't be affected at all.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Marek Vasut <marex@denx.de>
This converts the following to Kconfig:
CONFIG_CONS_INDEX
We have existing entries for this option in a number of places, with
different guards on them. They're also sometimes used for things not
directly inside of the serial driver. First, introduce a new symbol to
guard the use of CONFIG_CONS_INDEX, so that in the case where we don't
need this for the serial driver, but for some other use, we can still do
it. Next, consolidate all of these into the single entry in
drivers/serial/Kconfig. Finally, introduce CONS_INDEX_[023456] so that
we can imply a correct value here to make the defconfig side of this
smaller.
Signed-off-by: Adam Ford <aford173@gmail.com>
[trini: Rework a lot of the logic here, such that I took authorship from
Adam, but kept his S-o-B line]
Signed-off-by: Tom Rini <trini@konsulko.com>
Remove busy looping during watchdog reset.
Each polling of W_PEND_WTGR bit ("finish posted
write") after watchdog reset takes 120-140us
on BeagleBone Black board. Current U-Boot code
has watchdog resets in random places and often
there is situation when watchdog is reset
few times in a row in nested functions.
This adds extra delays and slows the whole system.
Instead of polling W_PEND_WTGR bit, we skip
watchdog reset if the bit is set. Anyway, watchdog
is in the middle of reset *right now*, so we can
just return.
This noticeably increases performance of the
system. Below are some measurements on BBB:
- DFU upload over USB 15% faster
- fastboot image upload 3x times faster
- USB ep0 transfers with 4k packets 20% faster
Signed-off-by: Ruslan Bilovol <ruslan.bilovol@gmail.com>
Tested-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Tested-by: Alex Kiernan <alex.kiernan@gmail.com>
As said in the SAMA5D2 datasheet, the PLLA clock must be divided
by 2 by writing the PLLADIV2 bit in PMC_MCKR, if the ratio between
PCK and MCK is 3 (MDIV = 3). This is the purpose of the driver.
Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
