In dbsc5_read_vref_training the arrays dvw_min_byte0_table and
dvw_min_byte1_table have 128 elements per channel. The variable
vref_stop_index is limited to be a maximum of 128. This means that the
index used to access the arrays must use a test of '< vref_stop_index'
rather than '<= vref_stop_index' in order to prevent out of bounds
accesses to the arrays.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
Reviewed-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Tested-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
K3 Inline ECC mechanism can support up to 3 regions of inline ECC, add
this support for single controller.
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
The existing approach does not account for interleaving in the DDRs when
setting up regions. There is support for MSMC to calculate the regions
for each DDR, so modify k3_ddrss_probe to set the regions accordingly
for multi-DDR systems.
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
Add support for calculation of the protected regions for each DDR in
multi-DDR systems. Since MSMC is the parent node of the individual DDRs
as well as responsible for their interleaving, it only makes sense for
MSMC to contain the logic for dividing the regions.
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
In K3 multi-DDR systems, the MSMC is responsible for the interleave
mechanism across all the DDR controllers. Add support for MSMC to obtain
the number of controllers it's responsible for using the DT.
Reviewed-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
As we increase the functionalities that the K3 DDRSS sub-system support,
it is becoming more evident that the same logic cannot apply to both
single as well as multiple DDR controller devices. Add
CONFIG_K3_MULTI_DDR to be used to differentiate between the two.
Reviewed-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
Instead of defaulting to choosing the entire DDR region when enabling
inline ECC, allow picking of a range within the DDR space using DT to
enable.
It expects such a node within the memory node, in the absence of which
we resort to enabling inline ECC for the entire DDR region:
inline_ecc: protected@9e780000 {
device_type = "ecc";
reg = <0x9e780000 0x0080000>;
bootph-all;
};
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
The reserved space needed for storing the parity remains the same no
matter the size of the region that is being protected. Add this as a
comment for better code understanding.
Reviewed-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
To prevent checkpatch warning once we start using this macro more
frequently, shorten the length of it. While at it, also move the
structure k3_ddrss_ecc_region above k3_msmc so that future patches can
have it as a member of k3_msmc.
Reviewed-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
Let ecc_regions[x].start reflect the start of the ECC region in terms of
DDR addressing rather than system addressing. This will make it easier
to extend the usage of the same ecc_regions structure for multi-DDR
systems as well.
Reviewed-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
The test for the interrupt LPDDR4_INTR_BIST_DONE is using a bitwise and
but the test is simple logic so use the more appropriate logical and.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
A few ram drivers cannot build without access to some platform
specific header files. Express those requirements in Kconfig as well.
Signed-off-by: Tom Rini <trini@konsulko.com>
E Shattow <e@freeshell.de> says:
Make consistent use of lowercase hexadecimal prefix '0x' throughout U-Boot.
There are a few remaining uses of uppercase 'X' to denote hexadecimal prefix
or placeholder in documentation and error messages.
External devicetree-rebasing dts/upstream and the generated code of
xilinx/zynq are ignored for the series.
Link: https://lore.kernel.org/r/20250606224558.1117422-1-e@freeshell.de
Use consistent lowercase hex prefix style in drivers/*
Does not change hex prefix case in allcaps uppercase style error messages
Signed-off-by: E Shattow <e@freeshell.de>
RAM drivers using Device Model currently lack of their own Kconfig menu
entry, which makes Kconfig put all options of the class in the top-level
menu of device drivers. These options are also incorrectly grouped with
pinctrl options in the generated .config, which is hard to read. Let's
create a menu entry for these drivers.
Fixes: 6c51df6859f ("dm: Add support for RAM drivers")
Signed-off-by: Yao Zi <ziyao@disroot.org>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Add support for configuring DRAM controller on STM32MP13xx SoC.
The DRAM controller is basically identical to the DWC controller
on STM32MP15xx SoC, except the bus width is reduced from 32bit to
16bit and a few registers and bits are therefore not present.
Handle the difference by factoring these parts out. Use IS_ENABLE()
as much as possible to assure code which is not enabled on builds
for a single SoC gets compiled out. Handle the different offset of
RCC_DDRITFCR register and missing DDRC2 clock the same way.
Signed-off-by: Marek Vasut <marek.vasut@mailbox.org>
The RAM_ROCKCHIP_DEBUG can be used only if DEBUG_UART is
available.
The next commit introduces changes in definition of debug
uart functions, so that DEBUG_UART is required to be defined
in order to initialize uart and use print functions.
Signed-off-by: Lukasz Czechowski <lukasz.czechowski@thaumatec.com>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
This patch cleans the vendor code of DDR initialization up, converts the
driver to fit in DM framework and use a firmware[1] packaged by binman to
ship PHY configuration.
Currently the driver is only capable of initializing the controller to
work with dual-rank 3733MHz LPDDR4, which is shipped by 16GiB variants
of LicheePi 4A boards and I could test with. Support for other
configurations could be easily added later.
Link: https://github.com/ziyao233/th1520-firmware # [1]
Signed-off-by: Yao Zi <ziyao@disroot.org>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
Add ddr driver for rk3576 to get the ram capacity.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Add support for reading DRAM size information from PMUGRF os_reg18 reg.
Compared to most Rockchip SoCs the RK3528 use os_reg18 for DRAM info,
instead of os_reg2.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Currently in j721e_init.c we check which firewalls to remove using
the board configuration (e.g CONFIG_TARGET_J721E_R5_EVM). We do this
as J721e and J7200 have different IP and firewalls but use the same
SoC definition (SOC_K3_J721E) even though they are different SoCs.
The idea was they would be similar enough that they both could use
the same SoC config to help with common code sharing. Board checks
would then be used differentiate.
This has grown far too messy to maintain any more, especially now
that there is more than one board using J721e (EVM, SK, Beagle AI64).
As differentiation is done based on board, every one of these boards
would have to have checks added for them. Instead let's split J7200
support out from J721e like how normal new SoC support is done.
This patch touches several subsystems and could not be split much better
as when we add SOC_K3_J7200 we want to make use of it in all spots that
once used the combined SOC_K3_J721E so we can turn off SOC_K3_J721E when
building for J7200 boards.
Signed-off-by: Andrew Davis <afd@ti.com>
Pass DBSC5 udevice and MODEMR0 register values to board specific
function dbsc5_get_board_data(). The board specific implementation
of dbsc5_get_board_data() can return struct renesas_dbsc5_board_config
which matches the board based on the content of MODEMR0 or content
of DT accessible via the udevice.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Extract wait for completion code from dbsc5_send_dbcmd2() into
new separate function dbsc5_wait_dbwait(). This extracted code
can be used to implement MR register read in the future.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Update dbsc5_send_dbcmd2() such that it takes multiple parameters
instead of one magic register content value. These parameters are
used to form the same resulting register value internally in the
dbsc5_send_dbcmd2() function, but from well defined input constants.
The new input constants are the operation code, channel, rank, and
operation argument. The argument is operation code specific, therefore
it is still a 16-bit magic number, but the rest of the arguments are
now split up. No functional change.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Rename dbsc5_ddr_register_read() to dbsc5_ddr_register_mr27_mr57_read()
and dbsc5_ddr_register_set() to dbsc5_ddr_register_mr28_set() to make
it clear what those functions really do. No functional change.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Make struct renesas_dbsc5_board_config {} definition public via
include/dbsc5.h, so this structure can be defined in board files
and passed into the DBSC5 DRAM driver by overriding weak function
dbsc5_get_board_data() on board level.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Add auto-detection and handling of Renesas R-Car V4H-3 and V4H-5
in addition to V4H-7 SoC variants based on OTP fuse programming.
The V4H-3 and V4H-5 variants have reduced DRAM frequency options.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Update the DRAM initialization code to match DBSC5 initialization code
rev.1.10 , which is currently the latest version available. This makes
DRAM initialization operational on Renesas R-Car V4H R8A779G0 rev.3.0.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reinstate missing increment by two in DBTR11 calculation based
on the original DBSC5 initialization code rev.0.80. The original
code did ... ODTLon - (js2[JS2_tODTon_min] - 1) + 1 , which was
incorrectly converted into ODTLon - js2[JS2_tODTon_min], but
should have been converted to ODTLon - js2[JS2_tODTon_min] + 2.
Add the missing +2 .
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
The JS1 index is calculated correctly, but the limiter cannot
be the max() function because the index should be lower than
JS1_USABLEC_SPEC_HI and the max() function would unconditionally
override the JS1 index to JS1_USABLEC_SPEC_HI. Use clamp() to
limit the JS1 index instead.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Fix copy paste error in MD pin handling for 5500 Mbps and 4800 Mbps case,
each should be handled by MD[19,17] == 2 and MD[19,17] == 3 respectively.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
The SDRAM_IDX in DDRSS_V2A_CTL_REG describes the number of address bits
minus 16 that are used to determine the mask used to detect memory
rollover and prevent aliasing and false coherency issues.
Set SDRAM_IDX using the device private data, ddr_ram_size for K3 family
of SoCs.
Signed-off-by: Santhosh Kumar K <s-k6@ti.com>
Reviewed-by: Neha Malcom Francis <n-francis@ti.com>
The functionality of enabling Inline ECC is now controlled by
CONFIG_K3_INLINE_ECC. So, remove the support for 'ti,ecc-enable'
property to avoid redundancy and to ensure the Inline ECC feature is
mananged through build-time config.
Signed-off-by: Santhosh Kumar K <s-k6@ti.com>
Reviewed-by: Wadim Egorov <w.egorov@phytec.de>
Add CONFIG_K3_INLINE_ECC so that ECC functions can be compiled into R5 SPL
only when the config has been enabled.
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
Signed-off-by: Santhosh Kumar K <s-k6@ti.com>
Reviewed-by: Wadim Egorov <w.egorov@phytec.de>
Enable ECC 1-bit error, 2-bit error, multiple 1-bit error interrupts
by setting the respective bits in the DDRSS_V2A_INT_SET_REG register.
Signed-off-by: Santhosh Kumar K <s-k6@ti.com>
Setup the ECC region's start and range using the device private data,
ddrss->ddr_bank_base[0] and ddrss->ddr_ram_size. Also, move start and
range of ECC regions from 32 bits to 64 bits to accommodate for
DDR greater than or equal to 4GB.
Signed-off-by: Santhosh Kumar K <s-k6@ti.com>
As R5 is a 32 bit processor, the RAM banks' base and size calculation
is restricted to 32 bits, which results in wrong values if bank's base
is greater than 32 bits or bank's size is greater than or equal to 4GB.
So, add k3_ddrss_ddr_bank_base_size_calc() to get the base address and
size of RAM's banks from the device tree memory node, and store in a
64 bit device private data which can be used for ECC reserved memory
calculation, Setting ECC range and Fixing up bank size in device tree
when ECC is enabled.
Signed-off-by: Santhosh Kumar K <s-k6@ti.com>
Reviewed-by: Wadim Egorov <w.egorov@phytec.de>
The 1-bit inline ECC support in TI's DDRSS bridge requires
the configured memory regions to be preloaded with a pattern
before use. This is done by the k3-ddrss driver from the
R5 SPL in a 'for' loop. It takes around 10 seconds to fill
2GB of memory, for example. Memset can cut the time in half
and using DMA currently yields a similar result.
The BIST engine of DDR controller provides support for
initializing any memory region with a pattern. This
bypasses the DDRSS bridge, so the required inline ECC
data is not computed and populated in the memory. For
some values like zero, the computed ECC syndrome is also
zero and we can use these values to preload the memory
from the DDR controller, without the assistance of the
bridge.
The registers involved in the process are described in the
'DDR controller registers' topic in [1] AM62 and [2] J721E
reference manuals.
The patch replaces the 'for' loop memory fill function with
the BIST memory initialization procedure. This cuts the time
to preload the 2GB memory from 10 seconds down to 1 second.
The bist preload function uses the lpddr4 APIs in the k3-ddrss,
so this is compatible with devices with both 16-bit LPDDR4 and
32-bit LPDDR4 interfaces (e.g J721E).
[1] AM62x: https://www.ti.com/lit/pdf/spruiv7
[2] DRA829/TDA4VM: https://www.ti.com/lit/zip/spruil1
Signed-off-by: Georgi Vlaev <g-vlaev@ti.com>
Signed-off-by: Santhosh Kumar K <s-k6@ti.com>
Add Renesas R-Car Gen4 DBSC5 DRAM controller driver. This driver is currently
capable of bringing LPDDR5 DRAM on Renesas R-Car V4H Whitehawk board. Further
boards can be supported by supplying board specific DRAM configuration data
via dbsc5_get_board_data(). Support for R-Car V4M is not implemented, however
the driver is already mostly prepared to support this SoC.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Those debug() calls might be useful, but beware. They can cause the DDR
controller to hang if we do not run the sequence quickly enough.
They usually are not an issue with upstream U-Boot and the default DDR
config, but they have become troublesome with custom DDR configs.
Drop those debug() statements that shouldn't be present in
time-sensitive code, to avoid anyone else falling into the trap.
Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
Simon Glass <sjg@chromium.org> says:
When the SPL build-phase was first created it was designed to solve a
particular problem (the need to init SDRAM so that U-Boot proper could
be loaded). It has since expanded to become an important part of U-Boot,
with three phases now present: TPL, VPL and SPL
Due to this history, the term 'SPL' is used to mean both a particular
phase (the one before U-Boot proper) and all the non-proper phases.
This has become confusing.
For a similar reason CONFIG_SPL_BUILD is set to 'y' for all 'SPL'
phases, not just SPL. So code which can only be compiled for actual SPL,
for example, must use something like this:
#if defined(CONFIG_SPL_BUILD) && !defined(CONFIG_TPL_BUILD)
In Makefiles we have similar issues. SPL_ has been used as a variable
which expands to either SPL_ or nothing, to chose between options like
CONFIG_BLK and CONFIG_SPL_BLK. When TPL appeared, a new SPL_TPL variable
was created which expanded to 'SPL_', 'TPL_' or nothing. Later it was
updated to support 'VPL_' as well.
This series starts a change in terminology and usage to resolve the
above issues:
- The word 'xPL' is used instead of 'SPL' to mean a non-proper build
- A new CONFIG_XPL_BUILD define indicates that the current build is an
'xPL' build
- The existing CONFIG_SPL_BUILD is changed to mean SPL; it is not now
defined for TPL and VPL phases
- The existing SPL_ Makefile variable is renamed to SPL_
- The existing SPL_TPL Makefile variable is renamed to PHASE_
It should be noted that xpl_phase() can generally be used instead of
the above CONFIGs without a code-space or run-time penalty.
This series does not attempt to convert all of U-Boot to use this new
terminology but it makes a start. In particular, renaming spl.h and
common/spl seems like a bridge too far at this point.
The series is fully bisectable. It has also been checked to ensure there
are no code-size changes on any commit.
Use PHASE_ as the symbol to select a particular XPL build. This means
that SPL_TPL_ is no-longer set.
Update the comment in bootstage to refer to this symbol, instead of
SPL_
Signed-off-by: Simon Glass <sjg@chromium.org>
