The DDR calibration tool for i.MX8M currently produces 3732 MTps rate in
lpddr4_timing.c , while the PHY code expects 3733 MTps rate. Support both
variants to avoid surprises where the system fails to boot.
Signed-off-by: Marek Vasut <marex@denx.de>
This syncs drivers/ddr/marvell/a38x/ with the master branch of repository
https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell.git
up to the commit "mv_ddr: a3700: Use the right size for memset to not overflow"
d5acc10c287e40cc2feeb28710b92e45c93c702c
This patch was created by following steps:
1. Replace all a38x files in U-Boot tree by files from upstream github
Marvell mv-ddr-marvell repository.
2. Run following command to omit portions not relevant for a38x, ddr3, and ddr4:
files=drivers/ddr/marvell/a38x/*
unifdef -m -UMV_DDR -UMV_DDR_ATF -UCONFIG_APN806 \
-UCONFIG_MC_STATIC -UCONFIG_MC_STATIC_PRINT -UCONFIG_PHY_STATIC \
-UCONFIG_PHY_STATIC_PRINT -UCONFIG_CUSTOMER_BOARD_SUPPORT \
-UCONFIG_A3700 -UA3900 -UA80X0 -UA70X0 -DCONFIG_ARMADA_38X -UCONFIG_ARMADA_39X \
-UCONFIG_64BIT $files
3. Manually change license to SPDX-License-Identifier
(upstream license in upstream github repository contains long license
texts and U-Boot is using just SPDX-License-Identifier.
After applying this patch, a38x, ddr3, and ddr4 code in upstream Marvell github
repository and in U-Boot would be fully identical. So in future applying
above steps could be used to sync code again.
The only change in this patch are:
1. Some fixes with include files.
2. Some function return and basic type defines changes in
mv_ddr_plat.c (to correct Marvell bug).
3. Remove of dead code in newly copied files (as a result of the
filter script stripping out everything other than a38x, dd3, and ddr4).
Reference:
"ddr: marvell: a38x: Sync code with Marvell mv-ddr-marvell repository"
107c3391b9
Signed-off-by: Tony Dinh <mibodhi@gmail.com>
Reviewed-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
At this point, the remaining places where we have a symbol that is
defined as CONFIG_... are in fairly odd locations. While as much dead
code has been removed as possible, some of these locations are simply
less obvious at first. In other cases, this code is used, but was
defined in such a way as to have been missed by earlier checks. Perform
a rename of all such remaining symbols to be CFG_... rather than
CONFIG_...
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
We have a number of places that undef CONFIG_... while we never
reference CONFIG_... in the first place. Remove these lines.
Signed-off-by: Tom Rini <trini@konsulko.com>
The way all of the memory init code here works is that we pass
0xDEADBEEF around for the initial value (as it's a well known 'poison'
value and so easily recognized in debuggers, etc). The only point of
this CONFIG symbol was to pass in a different value for that purpose.
Drop this symbol and cleanup the code slightly.
Signed-off-by: Tom Rini <trini@konsulko.com>
The rest of the unmigrated CONFIG symbols in the CONFIG_SYS namespace do
not easily transition to Kconfig. In many cases they likely should come
from the device tree instead. Move these out of CONFIG namespace and in
to CFG namespace.
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The rest of the unmigrated CONFIG symbols in the CONFIG_SYS_SDRAM
namespace do not easily transition to Kconfig. In many cases they likely
should come from the device tree instead. Move these out of CONFIG
namespace and in to CFG namespace.
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Commit 290ffe5788 (imx8m: fix reading of DDR4 MR registers) lifted a
private definition of lpddr4_mr_read() from imx8mm-cl-iot-gate board
code to drivers/ddr/imx/imx8m/ddrphy_utils.c, because that version
actually seems to work in practice.
However, commit 99c7cc58e1 (ddr: imx: Add i.MX9 DDR controller driver)
reintroduced the broken version in drivers/ddr/imx/imx8m/ddr_init.c,
copied most of the rest of ddrphy_utils.c to
drivers/ddr/imx/phy/ddrphy_utils.c, and stopped building
drivers/ddr/imx/imx8m/ddrphy_utils.c [and that file was then finally
completely removed with 7e9bd84883 (imx8m: ddrphy_utils: Remove unused
file)].
I assume this must have broken the imx8mm-cl-iot-gate board, at least
those that have not had their eeprom programmed with the proper
information. It certainly did break our out-of-tree board which always
reads back the ID register and uses that for a sanity check.
So apply the fix from 290ffe5788 once again.
Fixes: 99c7cc58e1 (ddr: imx: Add i.MX9 DDR controller driver)
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Fabio Estevam <festevam@denx.de>
During init_dram() is called also compute_lowest_common_dimm_parameters()
function which prints multi-line detailed output. So print also re-aligning
filler after "Detected ?DIMM" line to have "DRAM: " output aligned.
Signed-off-by: Pali Rohár <pali@kernel.org>
U-Boot core code already handles the case when RAM size is bigger than
CONFIG_MAX_MEM_MAPPED. So there is no need to do duplicate check in fsl ddr
driver for CONFIG_MAX_MEM_MAPPED. Instead simplify code to just check if
RAM size can be representable in phys_size_t type. And avoid printing
warning if phys_size_t is just 1 byte smaller than RAM size, which is the
typical situation with 4 GB DDR module.
Signed-off-by: Pali Rohár <pali@kernel.org>
Function fsl_ddr_compute() always return size in unsigned long long type,
but function fsl_ddr_sdram_size() returns size in phys_size_t type.
When 36-bit addressing mode is not enabled then phys_size_t type is only
32-bit and thus it cannot store value 4GB (0x100000000). Function
fsl_ddr_sdram_size() in this case returns truncated value 0x0.
Fix this issue by returning the highest representable value, which is
0xffffffff (4GB - 1 byte).
This change fixes crashing of proper U-Boot because it detected 4 GB module
as RAM with zero size.
Signed-off-by: Pali Rohár <pali@kernel.org>
Check needs to be done against CONFIG_MAX_MEM_MAPPED macro and not fixed
size 4GB (as CONFIG_MAX_MEM_MAPPED can be lower and for example for e500
cores it is just 2GB). Also fix printf re-align, which should be applied
only for non-SPL builds, during init_dram() call.
Signed-off-by: Pali Rohár <pali@kernel.org>
The ddrphy_utils.c is now deduplicated in drivers/ddr/imx/phy/ddrphy_utils.c ,
this drivers/ddr/imx/imx8m/ddrphy_utils.c is a remnant from when the
deduplication was implemented and was not removed. Remove it as it is
unused.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@denx.de>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Ye Li <ye.li@nxp.com>
Cc: uboot-imx <uboot-imx@nxp.com>
Reviewed-by: Fabio Estevam <festevam@denx.de>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Globally replace all occurances of WATCHDOG_RESET() with schedule(),
which handles the HW_WATCHDOG functionality and the cyclic
infrastructure.
Signed-off-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Tom Rini <trini@konsulko.com> [am335x_evm, mx6cuboxi, rpi_3,dra7xx_evm, pine64_plus, am65x_evm, j721e_evm]
In both the Freescale DDR controller and the SPD spec, bank address bits
are stored as the number of bank address bits minus 2. For example, if a
chip had 8 banks (3 total bank address bits), the value of
bank_addr_bits would be 1. This is rather surprising for users
configuring their memory manually, since they can't set bank_addr_bits
to the actual number of bank address bits. Rectify this.
There is at least one example of this kind of mistake already, in
board/freescale/t102xrdb/ddr.c. The documented MT40A512M8HX has two bank
address bits, but bank_addr_bits was set to 2, implying 4 bank address
bits. Such a value is reserved in BA_BITS_CS, but I suspect the
controller simply ignores the top bit, making this kind of mistake
harmless, if misleading.
Fixes: e8a7f1c32b ("powerpc/t1023rdb: Add T1023 RDB board support")
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The interactive mode uses large several tables of options which can be
configured. However, much of the contents of these tables are
repetetive. For example, no struct is larger than half a kilobyte, so
the offset only takes up 9 bits. Similarly, the size is only ever 4 or
8, and printhex is a boolean. Reduce the size of these fields. This
reduces the size of the options tables by around 10 KiB. However, the
largest contributor to the size of the options tables is the use of a
pointer for the strings. A better approach would be to use a separate
array of strings, and store an integer index in the options tables.
However, this would require a large re-architecting of this file.
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
This converts the following to Kconfig:
CONFIG_SYS_FSL_DDR_MAIN_NUM_CTRLS
CONFIG_SYS_FSL_OTHER_DDR_NUM_CTRLS
And we remove the entries from the README for a number of already
converted items.
Signed-off-by: Tom Rini <trini@konsulko.com>
By reading binman symbols, we no need hard coded IMEM_LEN/DMEM_LEN after
we update the binman dtsi to drop 0x8000/0x4000 length for the firmware.
And that could save binary size for many KBs.
Tested-by: Tim Harvey <tharvey@gateworks.com> #imx8m[m,n,p]-venice
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
[Alper: Check BINMAN_SYMS_OK instead]
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Add Kconfig for enabling reference events counter in DDRC performance
monitor by default
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Since i.MX9 uses same DDR PHY with i.MX8M, split the DDRPHY to a common
directory under imx, then use dedicated ddr controller driver for each
iMX9 and iMX8M.
The DDRPHY registers are space compressed, so it needs conversion to
access the DDRPHY address. Introduce a common PHY address remap function
for both iMX8M and iMX9 for all PHY registers accessing.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
- Drop the emulator CONFIG test from include/configs/ls1088ardb.h
- Migrate CONFIG_SYS_FSL_DDR_EMU to a select'able option in
drivers/ddr/fsl/Kconfig
Signed-off-by: Tom Rini <trini@konsulko.com>
Convert the constant integer to 'phys_size_t' to avoid overflow
when calculating the SDRAM size.
Signed-off-by: Dinesh Maniyam <dinesh.maniyam@intel.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@intel.com>
Replace with phys_size_t for all memory size variables declaration
for the sake of scalability. phys_size_t is defined in
/arch/arm/include/asm/types.h.
Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
Bit[7-4] for both register seq2core and core2seq handshake in HPS are not
required for triggering DDR re-calibration or resetting EMIF. So, ignoring
these bits just for playing it safe.
Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
All the source code of sdram_n5x.c are from Intel, update the license to
use both GPL2.0 and BSD-3 Clause because this copy of code may used for
open source and internal project.
Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
I was trying to employ lpddr4_mr_read() to something similar to what
the imx8mm-cl-iot-gate board is doing for auto-detecting the RAM
type. However, the version in drivers/ddr/imx/imx8m/ddrphy_utils.c
differs from the private one used by that board in how it extracts the
byte value, and I was only getting zeroes. Adding a bit of debug
printf'ing gives me
tmp = 0x00ffff00
tmp = 0x00070700
tmp = 0x00000000
tmp = 0x00101000
and indeed I was expecting a (combined) value of 0xff070010 (0xff
being Manufacturer ID for Micron). I can't find any documentation that
says how the values are supposed to be read, but clearly the iot-gate
definition is the right one, both for its use case as well as my
imx8mp-based board.
So lift the private definition of lpddr4_mr_read() from the
imx8mm-cl-iot-gate board code to ddrphy_utils.c, and add a declaration
in the ddr.h header where e.g. get_trained_CDD() is already declared.
This has only been compile-tested for the imx8mm-cl-iot-gate
board (since I don't have the hardware), but since I've merely moved
its definition of lpddr4_mr_read(), I'd be surprised if it changed
anything for that board.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Tested-by: Ying-Chun Liu (PaulLiu) <paul.liu@linaro.org>
Reviewed-by: Fabio Estevam <festevam@denx.de>
When env support is disabled then usage of env_get_f() generates linker
errors. So do not compile env_get_f() when env support is disabled (for
example when disabled only in SPL).
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
Add entry for 3732 MT/s mode of operation of the LPDDR4, in
which case the DDR PLL has to be configured in 933 MHz mode.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@denx.de>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
Reviewed-by: Fabio Estevam <festevam@denx.de>
The value CONFIG_DB_784MP_GP is only used in the DDR code to refer to
CONFIG_TARGET_DB_MV784MP_GP so just use that second value directly.
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Stefan Roese <sr@denx.de>
This converts the following to Kconfig:
CONFIG_CHIP_SELECTS_PER_CTRL
Cc: Alison Wang <alison.wang@nxp.com>
Cc: Pramod Kumar <pramod.kumar_1@nxp.com>
Cc: Priyanka Jain <priyanka.jain@nxp.com>
Cc: Rajesh Bhagat <rajesh.bhagat@nxp.com>
Cc: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
In commit 3fc92a215b ("ddr: marvell: a38x: fix SPLIT_OUT_MIX state
decision") I ported a cleaned up and changed version of patch
mv_ddr: a380: fix SPLIT_OUT_MIX state decision
In the port we removed checking for BYTE_HOMOGENEOUS_SPLIT_OUT bit,
because:
- the fix seemed to work without it
- the bit was checked for only at one place out of two, while the second
bit, BYTE_SPLIT_OUT_MIX, was checked for in both cases
- without the removal it didn't work on Allied Telesis' x530 board
We recently had a chance to test on more boards, and it seems that the
change needs to be opposite: instead of removing the check for
BYTE_HOMOGENEOUS_SPLIT_OUT from the first if() statement, the check
needs to be added also to the second one - it needs to be at both
places.
With this change all the Turris Omnia boards I have had available to
test seem to work, I didn't encounter not even one failed DDR training.
As last time, I am noting that I do not understand what this code is
actually doing, I haven't studied the DDR training algorithm and
I suspect that no one will be able to explain it to U-Boot contributors,
so we are left with this blind poking in the code with testing whether
it works on several boards and hoping it doesn't break anything for
anyone :-(.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Tested-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
After acking the requested frequency, should wait the ack bit clear
by DDR controller and check the DFS interrupt for next request polling.
Otherwise, the next polling of request bit will get previous value
that DDR controller have not cleared it, so a wrong request frequency
is used.
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
On i.MX8ULP, The dram config timing need to be saved into sram for
ddr retention when APD enter PD mode, so add this support on i.MX8ULP.
Reviewed-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
This is a cleaned up and fixed version of a patch
mv_ddr: a380: fix SPLIT_OUT_MIX state decision
in each pattern cycle the bus state can be changed
in order to avoide it, need to back to the same bus state on each
pattern cycle
by
Moti Boskula <motib@marvell.com>
The original patch is not in Marvell's mv-ddr-marvell repository. It was
gives to us by Marvell to fix an issues with DDR training on some
boards, but it cannot be applied as is to mv-ddr-marvell, because it is
a very dirty draft patch that would certainly break other things, mainly
DDR4 training code in mv-ddr-marvell, since it changes common functions.
I have cleaned up the patch and removed stuff that seemed unnecessary
(when removed, it still fixed things). Note that I don't understand
completely what the code does exactly, since I haven't studied the DDR
training code extensively (and I suspect that no one besides some few
people in Marvell understand the code completely).
Anyway after the cleanup the patch still fixes isssues with DDR training
on the failing boards.
There was also a problem with the original patch on some of the Allied
Telesis' x530 boards, reported by Chris Packham. I have asked Chris to
send me some logs, and managed to fix it:
- if you look at the change, you'll notice that it introduces
subtraction of cur_start_win[] and cur_end_win[] members, depending on
a bit set in the current_byte_status variable
- the original patch subtracted cur_start_win[] if either
BYTE_SPLIT_OUT_MIX or BYTE_HOMOGENEOUS_SPLIT_OUT bits were set, but
subtracted cur_end_win[] only if the first one (BYTE_SPLIT_OUT_MIX)
was set
- from Chris Packham logs I discovered that the x530 board where the
original patch introduced DDR training failure, only the
BYTE_HOMOGENEOUS_SPLIT_OUT bit was set, and on our boards where the
patch is needed only the BYTE_SPLIT_OUT_MIX is set in the
current_byte_status variable
- this led me to the hypothesis that both cur_start_win[] and
cur_end_win[] should be subtracted only if BYTE_SPLIT_OUT_MIX bit is
set, the BYTE_HOMOGENEOUS_SPLIT_OUT bit shouldn't be considered at all
- this hypothesis also gains credibility when considering the commit
title ("fix SPLIT_OUT_MIX state decision")
Hopefully this will fix things without breaking anything else.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Chris Packham <judge.packham@gmail.com>
Before commit 4c289425752f ("mv_ddr: a38x: add support for ddr async
mode"), Asynchornous Mode was only used when the CPU Subsystem Clock
Options[4:0] field in the SAR1 register was set to value 0x13: CPU at
2 GHz and DDR at 933 MHz.
Then commit 4c289425752f ("mv_ddr: a38x: add support for ddr async
mode") added support for Asynchornous Modes with frequencies other than
933 MHz (but at least 467 MHz), but the code it added to check for
whether Asynchornous Mode should be used is wrong: it checks whether the
frequency setting in board DDR topology map is set to value other than
MV_DDR_FREQ_SAR.
Thus boards which define a specific value, greater than 400 MHz, for DDR
frequency in their board topology (e.g. Turris Omnia defines
MV_DDR_FREQ_800), are incorrectly put into Asynchornous Mode after that
commit.
The A38x Functional Specification, section 10.12 DRAM Clocking, says:
In Synchornous mode, the DRAM and CPU clocks are edge aligned and run
in 1:2 or 1:3 CPU to DRAM frequency ratios.
Change the check for whether Asynchornous Mode should be used according
to this explanation in Functional Specification.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Tested-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
No functional change intended. This patch switches from the legacy I2C
API to the DM I2C API, so that this code can be used with DM I2C
enabled.
Signed-off-by: Stefan Roese <sr@denx.de>
