In the RK3399 DRAM driver, the function set_ds_odt() supports operating
in two different modes, selected by the ctl_phy_reg argument: when true,
the function reads and writes directly from the DRAM registers, accessed
through "chan->pctl->denali_*"; when false, the function reads and
writes from an array, accessed through "params->pctl_regs.denali_*",
which is written to DRAM registers at a later time.
However, phy_config_io(), which is called by set_ds_odt() to do a subset
of its register operations, operates directly on DRAM registers at all
times. This means that it reads incorrect values (and writes new values
prematurely) when ctl_phy_reg in set_ds_odt() is false. Fix this by
passing in the address of the registers to work with.
This prevents an "Invalid DRV value" error in the SPL debug log and
(presumably) results in a more correct end state. See the following logs
from a RK3399 NanoPi M4 board (4GB LPDDR3):
Before:
sdram_init() Starting SDRAM initialization...
phy_io_config() Invalid DRV value.
phy_io_config() Invalid DRV value.
sdram_init() sdram_init: data trained for rank 2, ch 0
phy_io_config() Invalid DRV value.
phy_io_config() Invalid DRV value.
sdram_init() sdram_init: data trained for rank 2, ch 1
Channel 0: LPDDR3, 933MHz
BW=32 Col=10 Bk=8 CS0 Row=15 CS1 Row=15 CS=2 Die BW=16 Size=2048MB
Channel 1: LPDDR3, 933MHz
BW=32 Col=10 Bk=8 CS0 Row=15 CS1 Row=15 CS=2 Die BW=16 Size=2048MB
256B stride
256B stride
sdram_init() Finish SDRAM initialization...
After:
sdram_init() Starting SDRAM initialization...
sdram_init() sdram_init: data trained for rank 2, ch 0
sdram_init() sdram_init: data trained for rank 2, ch 1
Channel 0: LPDDR3, 933MHz
BW=32 Col=10 Bk=8 CS0 Row=15 CS1 Row=15 CS=2 Die BW=16 Size=2048MB
Channel 1: LPDDR3, 933MHz
BW=32 Col=10 Bk=8 CS0 Row=15 CS1 Row=15 CS=2 Die BW=16 Size=2048MB
256B stride
256B stride
sdram_init() Finish SDRAM initialization...
Signed-off-by: Thomas Hebb <tommyhebb@gmail.com>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
i2c changes for 2020.04
- updates the Designware I2C driver
- get timings from device tree
- handle units in nanoseconds
- make sure that the requested bus speed is not exceeded
- few smaller clean-ups
- adds enums for i2c speed and update drivers which use them
- global_data: remove unused mxc_i2c specific field
Add a driver for the regulators in the the DA9063 PMIC.
Robert Beckett: move regulator modes to header so board code can set
modes. Correct mode mask used in ldo_set_mode.
Add an option CONFIG_SPL_DM_REGULATOR_DA9063.
Signed-off-by: Martin Fuzzey <martin.fuzzey@flowbird.group>
Signed-off-by: Robert Beckett <bob.beckett@collabora.com>
This adds the basic register access operations and child regulator
binding (if a regulator driver exists).
Robert Beckett: simplify accesses by using bottom bit of address as
offset overflow. This avoids the need to track which page we are on.
Add an option CONFIG_SPL_DM_PMIC_DA9063.
Signed-off-by: Martin Fuzzey <martin.fuzzey@flowbird.group>
Signed-off-by: Robert Beckett <bob.beckett@collabora.com>
The 3-bit "command", or register, is encoded within the device address.
Configure the device accordingly, and pass command in DM I2C read/write
calls correctly.
Signed-off-by: Ian Ray <ian.ray@ge.com>
Signed-off-by: Robert Beckett <bob.beckett@collabora.com>
TI's AM654 SoC has a Cadence OSPI IP. Add a new compatible string for
the same.
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Cadence OSPI is similar to QSPI IP except that it supports Octal IO
(8 IO lines) flashes. Add support for Cadence OSPI IP with existing
driver using new compatible
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Add support for Octal flash devices. Octal flash devices use 8 IO lines
for data transfer. Currently only 1-1-8 Octal Read mode is supported.
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Add support for Direct Access Controller mode of Cadence QSPI. This
allows MMIO access to SPI NOR flash providing better read performance.
Direct mode is only exercised if AHB window size is greater than 8MB.
Support for flash address remapping is also not supported at the moment
and can be added in future.
For better performance, driver uses DMA to copy data from flash in
direct mode using dma_memcpy().
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Tested-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Acked-by: Jagan Teki <jagan@amarulasolutions.com>
Current Cadence QSPI driver has few limitations. It assumes all read
operations to be in Quad mode and thus does not support SFDP parsing.
Also, adding support for new mode such as Octal mode would not be
possible with current configuration. Therefore move the driver over to spi-mem
framework. This has added advantage that driver can be used to support
SPI NAND memories too.
Hence, move driver over to new spi-mem APIs.
Please note that this gets rid of mode bit setting done when
CONFIG_SPL_SPI_XIP is defined as there does not seem to be any user to
that config option.
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Tested-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Acked-by: Jagan Teki <jagan@amarulasolutions.com>
Linux has supported W25N01GV for a long time, so lets import it.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Make sure corresponding setup registers are updated depending on CS.
This ensures that driver can support QSPI flashes on ChipSelects other
than on CS0
Reported-by: Andreas Dannenberg <dannenberg@ti.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
The SPI stack relies on a proper bus speed/mode configuration
by calling dm_spi_claim_bus(). However the hitherto code
allowed to accidentally override those settings in
the spi_get_bus_and_cs() routine.
The initially established speed could be discarded by using
the slave platdata, which turned out to be an issue on
the platforms whose slave maximum supported frequency
is not on par with the maximum frequency of the bus controller.
This patch fixes above issue by configuring the bus from
spi_get_bus_and_cs() only in case it was not done before.
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Jagan Teki <jagan@amarulasolutions.com>
Add chip select number check in spi_find_chip_select().
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Jagan Teki <jagan@amarulasolutions.com> # SoPine
This is a port of the kernel's spi-nxp-fspi driver. It uses the new
spi-mem interface and does not expose the more generic spi-xfer
interface. The source was taken from the v5.3-rc3 tag.
The port was straightforward:
- remove the interrupt handling and the completion by busy polling the
controller
- remove locks
- move the setup of the memory windows into claim_bus()
- move the setup of the speed into set_speed()
- port the device tree bindings from the original fspi_probe() to
ofdata_to_platdata()
There were only some style change fixes, no change in any logic. For
example, there are busy loops where the return code is not handled
correctly, eg. only prints a warning with WARN_ON(). This port
intentionally left most functions unchanged to ease future bugfixes.
This was tested on a custom LS1028A board. Because the LS1028A doesn't
have proper clock framework support, changing the clock speed was not
tested. This also means that it is not possible to change the SPI
speed on LS1028A for now (neither is it possible in the linux driver).
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Kuldeep Singh <kuldeep.singh@nxp.com>
Move some of the code currently in the ofdata_to_platdata() method to
probe() so that it is not executed when generating ACPI tables.
Signed-off-by: Simon Glass <sjg@chromium.org>
We want to be able to calculate the speed separately from actually setting
the speed, so we can generate the required ACPI tables. Split out the
calculation into its own function.
Drop the double underscore on __dw_i2c_set_bus_speed while we are here.
That is reserved for compiler internals.
Signed-off-by: Simon Glass <sjg@chromium.org>
This is used to store the speed information for a bus. We want to provide
this to ACPI so that it can tell the kernel. Move this struct to the
header file so it can be accessed by the ACPI i2c implementation being
added later.
Signed-off-by: Simon Glass <sjg@chromium.org>
Convert the obvious uses of i2c bus speeds to use the enum.
Use livetree access for code changes.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Update this driver to use the new standard enums for speed.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Tested-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Update this driver to use the new standard enums for speed.
Note: This driver needs to move to driver model.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Update this driver to use the new standard enums for speed.
Note: This driver needs to move to driver model.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Some versions of this peripheral include a spike-suppression phase of the
bus. Add support for this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
At present the driver can end up with timing parameters which are slightly
faster than those expected. It is possible to optimise the parameters to
get the best possible result.
Create a new function to handle the timing calculation. This uses a table
of defaults for each speed mode rather than writing it in code.
The function works by calculating the 'period' of each bit on the bus in
terms of the input clock to the controller (IC_CLK). It makes sure that
the constraints are met and that the different components of that period
add up correctly.
This code was taken from coreboot which has ended up with this same
driver, but now in a much-different form.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Create a struct to hold the three timing parameters. This will make it
easier to move these calculations into a separate function in a later
patch.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Instead of passing this parameter into __dw_i2c_set_bus_speed(), pass in
the driver's private data, from which the function can obtain that
information. This allows the function to have access to the full state of
the driver.
Signed-off-by: Sicomp_param1mon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Signed-off-by: Simon Glass <sjg@chromium.org>
The i2c controller defines a few timing properties. Read these in and
store them for use by the driver.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
At present the driver uses an approximation for the bus clock, e.g. 166MHz
instead of 166 2/3 MHz.
This can result in small errors in the resulting I2C speed, perhaps 0.5%
or so.
Adjust the existing code to start from the accurate figure, even if later
rounding reduces this accuracy.
Update the bus speed code to work in KHz instead of MHz, which removes
most of the error.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Group these #defines into an enum to make it easier to understand the
relationship between them.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Jun Chen <ptchentw@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Some SoCs support a higher speed than what is currently called 'max' in
this driver. Rename it to 'high' speed, which is the official name of the
3.4MHz speed.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Jun Chen <ptchentw@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
We use struct clk here so really should include this header file to avoid
build errors. Also switch the order of clk.h in the C file to match the
required code style.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Ley Foon Tan <ley.foon.tan@intel.com>
Reviewed-by: Jun Chen <ptchentw@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
If a different input clock is required then the correct way to do this is
with a clock driver. Don't allow boards to override IC_CLK.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
Some versions of this peripherals provide more control of the bus
behaviour. Add definitions for these registers.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Ley Foon Tan <ley.foon.tan@intel.com>
Reviewed-by: Jun Chen <ptchentw@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Guard 'parent_rate==0' to prevent 'divide by zero' issue in
clk_pplv3_av_set_rate(). If it is 0, let's return with -EINVAL.
Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Guard 'mfd==0' to prevent 'divide by zero' issue in
clk_pplv3_av_get_rate(). If it is 0, let's return with EIO since mfd
should never be 0 at all.
Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Guard 'parent_rate==0' to prevent 'divide by zero' issue in
clk_pplv3_sys_get_rate(). If it is 0, let's return with -EINVAL.
Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
m68k needs block cache list initialized after relocation.
Other architectures must not be involved.
Fixing regression related to:
commit 1526bcce0f
("common: add blkcache init")
Signed-off-by: Angelo Durgehello <angelo.dureghello@timesys.com>
This function parameter usage of load_addr was incorrectly caught in
the clarifying renames of commit bb872dd930, which results in boot
failures on Marvell A38x.
Signed-off-by: Joel Johnson <mrjoel@lixil.net>
Patch-to: Simon Glass <sjg@chromium.org>
clk_get_by_name() requires clk_get_by_id() that is not available if
CONFIG_OF_PLATDATA is defined, so move clk_get_by_name() into #else
condition of #if CONFIG_IS_ENABLED(OF_PLATDATA).
Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
mtk_clk_find_parent_rate is calling clk_get_rate to know the rate
of a parent clock. clk_get_rate returns a ulong, while
mtk_clk_find_parent_rate returns an int. This implicit cast creates
an issue for clock rates big enough to need the full 32 bits to
store its data. When that happen the clk rate will become incorrect
because of the implicit cast between ulong -> int -> ulong.
This commit change the return type of mtk_clk_find_parent_rate to
ulong.
Signed-off-by: Fabien Parent <fparent@baylibre.com>
