The original code was configuring the external pins after enabling
the R_PIO clock, which meant the configuration never made it to
the pin controller the first time in SPL.
Why this was working before is uncertain. Maybe the state was left
from a previous boot sequence, or RSB just happened to be the default
configuration. However with some A33 chips, SPL failed to configure
the PMIC. This was seen by me and Maxime on the Sinlinx SinA33 dev
board.
Reordering the calls fixed this.
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Cc: Maxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Commit 487b327 ("sunxi: GPIO pin mux hardware-feature-specific function
index defines") renamed all GPIO index defines, but missed the PORT F
UART0 setup functions.
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
This reverts commit f76eba38b3.
This patch did not have a full and proper copyright/S-o-b chain.
Signed-off-by: Ian Campbell <ijc@hellion.org.uk>
Conflicts:
include/configs/sun6i.h
include/configs/sun8i.h
Some archs/boards specify their own default by pre-defining the config
which causes the Kconfig system to mix up the order of the configs in
the defconfigs... This will cause merge pain if allowed to proliferate.
Remove the configs that behave this way from the archs.
A few configs still remain, but that is because they only exist as
defaults and do not have a proper Kconfig entry. Those appear to be:
SPIFLASH
DISPLAY_BOARDINFO
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
[trini: rastaban, am43xx_evm_usbhost_boot, am43xx_evm_ethboot updates,
drop DM_USB from MSI_Primo81 as USB_MUSB_SUNXI isn't converted yet to DM]
Signed-off-by: Tom Rini <trini@konsulko.com>
DRA7/AM57xx devices can be operated in many different configurations.
When the SoC is supposed to support a configuration where low power mode
state may involve the SoC completely powered off and DDR is in self
refresh, SoC EMIF controller should not be the master of the reset
signal and an external entity might be in control of things.
The default configuration of Linux on TI evms involve not powering off
the voltage rails (due to various reasons including reliability concerns)
and must not allow DDR reset to be controlled by EMIF. On platforms
where external entity might control the reset signal, this configuration
will be a "dont care".
Fixes: 536d874708 ("ARM: DRA7: Update DDR IO registers")
Tested-by: Keerthy <j-keerthy@ti.com>
Acked-by: Brad Griffis <bgriffis@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Currently on sun6i after a "reset" the prompt returns and the user can
even type stuff until the watchdog triggers and does the actual reset.
This is somewhat unexpected behavior for the "reset" command, this
commit adds an endless loop to wait for the watchdog to trigger so that
we do not return to the prompt.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Unlike OMAP5, EMIF PHY used in DRA7 will be left in unknown state after
warm reset, emif needs to be configured to bring it back to a known
state. So configure EMIF during warm reset.
Reported-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
The recent changes for hw leveling on am33xx were not intended for
DDR2 boards, only DDR3. Update emif_sdram_type to take a sdram_config
value to check against. This lets us pass in the value we would use to
configure, when we have not yet configured the board yet. In other cases
update the call to be as functional as before and check an already
programmed value in.
Tested-by: Yan Liu <yan-liu@ti.com>
Signed-off-by: Tom Rini <trini@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
On AM57xx evm I2C5 is used to detect the LCD board by reading the
EEPROM present on the bus.
Enable i2c5 clocks to help that.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
In addition to the regular mux configuration, certain pins of DRA7
require to have "manual mode" also programmed, when predefined
delay characteristics cannot be used for the interface.
struct iodelay_cfg_entry is introduced for populating
manual mode IO timings.
For configuring manual mode, along with the normal pad
configuration do the following steps:
- Select MODESELECT field of each assocaited PAD.
CTRL_CORE_PAD_XXX[8]:MODESELECT = 1(Enable MANUAL_MODE macro along with mux)
- Populate A_DELAY, G_DELAY values that are specified in DATA MANUAL.
And pass the offset of the CFG_XXX register in iodelay_cfg_entry.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
On DRA7, in addition to the regular muxing of pins, an additional
hardware module called IODelay which is also expected to be
configured. This "IODelay" module has it's own register space that is
independent of the control module.
It is advocated strongly in TI's official documentation considering
the existing design of the DRA7 family of processors during mux or
IODelay recalibration, there is a potential for a significant glitch
which may cause functional impairment to certain hardware. It is
hence recommended to do muxing as part of IOdelay recalibration.
IODELAY recalibration sequence:
- Complete AVS voltage change on VDD_CORE_L
- Unlock IODLAY config registers.
- Perform IO delay calibration with predefined values.
- Isolate all the IOs
- Update the delay mechanism for each IO with new calibrated values.
- Configure PAD configuration registers
- De-isolate all the IOs.
- Relock IODELAY config registers.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
do_set_mux32() is redefined in dra7xx and beagle_x15 boards.
IO delay recalibration sequence also needs this.
Making it generic to avoid duplication.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
When DLL_CALIB_INTERVAL is set, an extra delay is added
which is not required and it consumes EMIF bandwidth.
So making the DLL_CALIB_CTRL[8:0]DLL_CALIB_INTERVAL bits to 0.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
DDRIO_2 and LPDDR2CH1_1 registers are not present
for DRA7. So not configuring these registers for DRA7xx
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Updating EMIF registers to enable HW leveling
on DRA72-evm.
Also updating the timing registers.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
DRA7 EMIF supports Full leveling for DDR3.
Adding support for the Full leveling sequence.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
We plan to enable device tree in SPL by default. Before doing this,
explicitly disable it for all boards.
Signed-off-by: Simon Glass <sjg@chromium.org>
This function is used before jumping to U-Boot, but in that case we don't
always want to disable caches.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
This fixes ethernet no longer working on boards which use a gpio to enable
the phy.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Before this commit the code for determining the disconnect threshold was
checking for sun4i or sun6i assuming that those where the exception and
that newer SoCs use a disconnect threshold of 2 like sun7i does.
But it turns out that newer SoCs actually use a disconnect threshold of 3
and sun5i and sun7i are the exceptions, so check for those instead.
Here are the settings from the various Allwinner SDK sources:
sun4i-a10: USBC_Phy_Write(usbc_no, 0x2a, 3, 2);
sun5i-a13: USBC_Phy_Write(usbc_no, 0x2a, 2, 2);
sun6i-a31: USBC_Phy_Write(usbc_no, 0x2a, 3, 2);
sun7i-a20: USBC_Phy_Write(usbc_no, 0x2a, 2, 2);
sun8i-a23: USBC_Phy_Write(usbc_no, 0x2a, 3, 2);
sun8i-h3: USBC_Phy_Write(usbc_no, 0x2a, 3, 2);
sun9i-a80: USBC_Phy_Write(usbc_no, 0x2a, 3, 2);
Note this commit makes no functional changes for sun4i - sun7i, and
changes the disconnect threshold for sun8i to match what Allwinner uses.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
This updates r8a7790 QoS to revision 0.973.
This commit can changed from KConfig to fit contents of the QoS.
Signed-off-by: Kouei Abe <kouei.abe.cp@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
sun8i can share the PSCI backend with sun6i. Only difference
is sun8i does not have CPU power clamp controls.
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
This adds PSCI support for sun6i. So far it only supports
the PWR_ON method.
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The PSCI code only works for sun7i. Rename it with _sun7i suffix,
and build only if building for sun7i.
This paves the way for adding PSCI support for other platforms.
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The PSCI CPU_ON code accesses quite a few registers. Document
their names to make it easier to cross reference.
Also explain "lock cpu" and "unlock cpu" as enabling/disabling
debug access.
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
This commit adds support to the sunxi SPL to load u-boot from the internal
NAND. Note this only adds support to access the boot partitions to load
u-boot, full NAND support to load the kernel, etc. from the nand data
partition will come later.
Signed-off-by: Roy Spliet <r.spliet@ultimaker.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Add initial sun9i (A80) support, only uart + mmc are supported for now.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
The latest versions of the fel tool support loading normal u-boot builds
directly, and this is now the preferred way to use the fel boot method.
This commit removes support for the old deprecated standalone fel builds.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Unlike the A31 and the A23 the A33 actually has a SID inside the SoC again,
but sid[3] is 0 (at least on some SoCs), so it is better to use the axp221
sid.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Make DRAM_ODT_EN Kconfig setting a bool, add a separate DRAM_ODT_CORRECTION
setting for A23 SoCs and use DRAM_ODT_EN Kconfig everywhere instead of
only in dram_sun4i.c and hardcoding odt_en elsewhere.
Note this commit makes no functional changes for existing boards,
its purpose is to allow changing the odt_en value on future A33 boards.
For sun4i/sun5i/sun7i boards which set DRAM_ODT_EN=y (which no defconfigs
currently do) this patch turns on odt for both the DQ and the DQS lines,
whereas previously it was possibly (but not desirable) to turn odt on only
for one of them by setting the in DRAM_ODT_EN option to 1 or 2 instead of 3.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
When porting the allwinner dram init code to u-boot we missed some code
setting an extra bit when doing auto dram config.
This commits add this bit, fixing dram init not working on the ga10h
10" a33 tablet which I'm bringing up atm.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Add support for UART2 (2-pin version but note that RTS/CTS pins are available
pn that port for possible future use), can be selected in config
by using CONFIG_CONS_INDEX=3
Signed-off-by: Laurent Itti <laurentitti@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The MX6 has a temperature grade defined by OCOTP_MEM0[7:6] which is at 0x480
in the Fusemap Description Table in the reference manual. Return this value
as well as min/max temperature based on the value.
Note that the IMX6SDLRM and the IMX6SXRM do not indicate this in the
their Fusemap Description Table however Freescale has confirmed that these
eFUSE bits match the description within the IMX6DQRM and that they will
be added to the next revision of the respective reference manuals.
This has been tested with IMX6 Automative and Industrial parts.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
The IMX6 has four different speed grades determined by eFUSE SPEED_GRADING
indicated by OCOTP_CFG3[17:16] which is at 0x440 in the Fusemap Description
Table. Return this frequency so that it can be used elsewhere.
Note that the IMX6SDLRM and the IMX6SXRM do not indicate this in the
their Fusemap Description Table however Freescale has confirmed that these
eFUSE bits match the description within the IMX6DQRM and that they will
be added to the next revision of the respective reference manuals.
These have been tested with IMX6 Quad/Solo/Dual-light 800Mhz and 1GHz grades.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Commit fa8b7d66f49f0c7bd41467fe78f6488d8af6976a introduced fast-exit support
to the MMDC however enabling it on the DDR3 got missed. Make sure we enable
it on the DDR3 as well.
Gateworks uses Micron memory as well as Winbond in MX6. We have found in
testing that we need to enable fast-exit for Winbond stability. Gateworks
boards are currently the only boards using the MX6 SPL and enabling
fast-exit mode.
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Having bit 22 cleared in the PL310 Auxiliary Control register (shared
attribute override enable) has the side effect of transforming Normal
Shared Non-cacheable reads into Cacheable no-allocate reads.
Coherent DMA buffers in Linux always have a Cacheable alias via the
kernel linear mapping and the processor can speculatively load cache
lines into the PL310 controller. With bit 22 cleared, Non-cacheable
reads would unexpectedly hit such cache lines leading to buffer
corruption.
This was inspired by a patch from Catalin Marinas [1] and also from recent
discussions in the linux-arm-kernel list [2] where Russell King and Rob Herring
suggested that bootloaders should initialize the cache.
[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2010-November/031810.html
[2] https://lkml.org/lkml/2015/2/20/199
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
This is proposal for clamping the MMDC/DDR3 clocks to the maximum supported
frequencies as per imx6 SOC models, and for dynamically calculating valid
clock value based on mem_speed.
Currently the code uses impossible values for mem_speed (1333, 1600 MT/s) for
calculating the DDR timings, and uses fixed clock (528 or 400 MHz) which
doesn't take into account DDR3 memory limitations.
Signed-off-by: Nikolay Dimitrov <picmaster@mail.bg>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Tim Harvey <tharvey@gateworks.com>
Cc: Eric Nelson <eric.nelson@boundarydevices.com>
