Bryan Brattlof [Tue, 21 Jun 2022 21:36:03 +0000 (16:36 -0500)]
soc: soc_ti_k3: identify j7200 SR2.0 SoCs
Anytime a new revision of a chip is produced, Texas Instruments
will increment the 4 bit VARIANT section of the CTRLMMR_WKUP_JTAGID
register by one. Typically this will be decoded as SR1.0 -> SR2.0 ...
however a few TI SoCs do not follow this convention.
Rather than defining a revision string array for each SoC, use a
default revision string array for all TI SoCs that continue to follow
the typical 1.0 -> 2.0 revision scheme.
Nishanth Menon [Fri, 17 Jun 2022 18:26:12 +0000 (13:26 -0500)]
board: ti: common: board_detect: Do 1byte address checks first.
Do 1 byte address checks first prior to doing 2 byte address checks.
When performing 2 byte addressing on 1 byte addressing eeprom, the
second byte is taken in as a write operation and ends up erasing the
eeprom region we want to preserve.
While we could have theoretically handled this by ensuring the write
protect of the eeproms are properly managed, this is not true in case
where board are updated with 1 byte eeproms to handle supply status.
Flipping the checks by checking for 1 byte addressing prior to 2 byte
addressing check prevents this problem at the minor cost of additional
overhead for boards with 2 byte addressing eeproms.
Signed-off-by: Nishanth Menon <nm@ti.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Nishanth Menon [Fri, 17 Jun 2022 18:26:11 +0000 (13:26 -0500)]
board: ti: common: Handle the legacy eeprom address width properly
Due to supply chain issues, we are starting to see a mixture of eeprom
usage including the smaller 7-bit addressing eeproms such as 24c04
used for eeproms.
These eeproms don't respond well to 2 byte addressing and fail the
read operation. We do have a check to ensure that we are reading the
alternate addressing size, however the valid failure prevents us
from checking at 1 byte anymore.
Rectify the same by falling through and depend on header data comparison
to ensure that we have valid data.
Signed-off-by: Nishanth Menon <nm@ti.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Nishanth Menon [Fri, 17 Jun 2022 18:26:10 +0000 (13:26 -0500)]
board: ti: common: Optimize boot when detecting consecutive bad records
The eeprom data area is much bigger than the data we intend to store,
however, with bad programming, we might end up reading bad records over
and over till we run out of eeprom space. instead just exit when 10
consecutive records are read.
Signed-off-by: Nishanth Menon <nm@ti.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Georgi Vlaev [Tue, 14 Jun 2022 14:45:33 +0000 (17:45 +0300)]
board: ti: am62x: Account for DDR size fixups if ECC is enabled
Call into k3-ddrss driver to fixup device tree and resize
the available amount of DDR if ECC is enabled.
A second fixup is required from A53 SPL to take the fixup
as done from R5 SPL and apply it to DT passed to A53 U-boot,
which in turn passes this to the OS.
Signed-off-by: Georgi Vlaev <g-vlaev@ti.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Georgi Vlaev [Tue, 14 Jun 2022 14:45:32 +0000 (17:45 +0300)]
board: ti: am62x: Use fdt functions for ram and bank init
Use the appropriate fdtdec_setup_mem_size_base() call in
dram_init() and fdtdec_setup_bank_size() in dram_bank_init()
to pull these values from DT, where they are already available,
instead of hardcoding them.
Signed-off-by: Georgi Vlaev <g-vlaev@ti.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Georgi Vlaev [Tue, 14 Jun 2022 14:45:30 +0000 (17:45 +0300)]
arm: mach-k3: common: Use ddr_init in spl_enable_dcache
The spl_enable_dcache() function calls dram_init_banksize()
to get the total memory size. Normally the dram_init_banksize()
setups the memory banks, while the total size is reported
by ddr_init(). This worked so far for K3 since we set the
gd->ram_size in dram_init_banksize() as well.
Signed-off-by: Georgi Vlaev <g-vlaev@ti.com> Reviewed-by: Tom Rini <trini@konsulko.com>
There are two decimal digits reserved to encode the module version and
revision. This code so far implemented A-Z which used 0-25 of this
range.
This commit extends the range to make use of all 99 numbers. After
capital letters the form with a hashtag and number (e.g. #26) is used.
Examples:
If the assembly version is between zero and 25 the numbering is as follows,
as it also has been before this commit:
0: V0.0A
1: V0.0B
...
25: V0.0Z
New numbering of assembly version:
If the number is between 26 and 99 the new assembly version name is:
26: V0.0#26
27: V0.0#27
...
99: V0.0#99
Signed-off-by: Philippe Schenker <philippe.schenker@toradex.com> Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com> Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
board: ti: j721e: Return if there is an error while configuring SerDes
While configuring SerDes, errors could be encountered, in these cases,
return instead of going ahead. This is will help in booting even if
configuration of SerDes fails.
Vaishnav Achath [Fri, 3 Jun 2022 06:02:16 +0000 (11:32 +0530)]
arm: k3: j721e: add dynamic sf bus override support for j721e
implement overrides for spl_spi_boot_bus() and spl_spi_boot_cs()
lookup functions according to bootmode selection, so as to support
both QSPI and OSPI boot using the same build.
Vaishnav Achath [Fri, 3 Jun 2022 06:02:15 +0000 (11:32 +0530)]
common: spl: spl_spi: add support for dynamic override of sf bus
Currently the SPI flash to load from is defined through the compile
time config CONFIG_SF_DEFAULT_BUS and CONFIG_SF_DEFAULT_CS, this
prevents the loading of binaries from different SPI flash using the
same build.E.g. supporting QSPI flash boot and OSPI flash boot
on J721E platform is not possible due to this limitation.
This commit adds lookup functions spl_spi_boot_bus()
and spl_spi_boot_cs for identifying the flash device based on the
selected boot device, when not overridden the lookup functions are
weakly defined in common/spl/spl_spi.c.
Tom Rini [Mon, 20 Jun 2022 12:07:54 +0000 (08:07 -0400)]
socrates: Rework CONFIG_PCI_CLK_FREQ
The symbol CONFIG_PCI_CLK_FREQ is local to this board. Provide equal
clarity in the code by referencing the numeric value directly and move
the explanatory comment to the code, just prior to use.
Tom Rini [Fri, 17 Jun 2022 20:24:34 +0000 (16:24 -0400)]
powerpc: Clean up CHAIN_OF_TRUST related options
As things stand currently, there is only one PowerPC platform that
enables the options for CHAIN_OF_TRUST. From the board header files,
remove a number of never-set options. Remove board specific values from
arch/powerpc/include/asm/fsl_secure_boot.h as well. Rework
include/config_fsl_chain_trust.h to not abuse the CONFIG namespace for
constructing CHAIN_BOOT_CMD. Migrate all of the configurable addresses
to Kconfig.
If any platforms are re-introduced with secure boot support, everything
required should still be here, but now in Kconfig, or requires migration
of an option to Kconfig.
Cc: Peng Fan <peng.fan@nxp.com> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Fri, 17 Jun 2022 20:24:33 +0000 (16:24 -0400)]
nxp: config_fsl_chain_trust.h: Clean up and remove unused portions
The way that secure boot is implemented today on NXP ARM platforms does
not reuse the elements found in include/config_fsl_chain_trust.h to
construct CONFIG_SECBOOT but instead board header files have their
environment setup as needed and then fsl_setenv_chain_of_trust() will
set secureboot in the environment. Remove a large number of unused
defines here.
Cc: Peng Fan <peng.fan@nxp.com> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Fri, 17 Jun 2022 20:24:32 +0000 (16:24 -0400)]
fsl_validate: Migrate SPL_UBOOT_KEY_HASH to Kconfig
Move setting of SPL_UBOOT_KEY_HASH to a non-NULL value to Kconfig. As
part of this, change fsl_secboot_validate(...) to check that it is
passed a non-empty string, rather than non-NULL.
Cc: Peng Fan <peng.fan@nxp.com> Cc: Priyanka Jain <priyanka.jain@nxp.com> Cc: Kshitiz Varshney <kshitiz.varshney@nxp.com> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Fri, 17 Jun 2022 20:24:31 +0000 (16:24 -0400)]
arch/Kconfig.nxp: Re-organize slightly
Make all of the CHAIN_OF_TRUST options be under a single menu and add a
comment for the rest, so the resulting config file reads more clearly.
Remove duplicate CHAIN_OF_TRUST options from
board/congatec/common/Kconfig. Remove duplicate NXP_ESBC config
questions and move to arch/Kconfig.nxp.
Tom Rini [Thu, 16 Jun 2022 18:04:40 +0000 (14:04 -0400)]
Convert CONFIG_SYS_FSL_SFP_BE et al to Kconfig
This converts the following to Kconfig:
CONFIG_KEY_REVOCATION
CONFIG_SYS_FSL_SFP_BE
CONFIG_SYS_FSL_SFP_LE
CONFIG_SYS_FSL_SFP_VER_3_0
CONFIG_SYS_FSL_SFP_VER_3_2
CONFIG_SYS_FSL_SFP_VER_3_4
CONFIG_SYS_FSL_SRK_LE
This partly means making sure to enable SYS_FSL_ERRATUM_A007186 only for
when CHAIN_OF_TRUST is enabled.
Tom Rini [Thu, 16 Jun 2022 18:04:36 +0000 (14:04 -0400)]
nxp: Rename board/freescale/common/Kconfig to arch/Kconfig.nxp
Now that board/freescale/common/Kconfig is safe to be included once,
globally, rename this to arch/Kconfig.nxp to better reflect that it
contains options that are valid on multiple architectures and SoC
families, and not specific to NXP reference platforms either.
Cc: Stefano Babic <sbabic@denx.de> Cc: Peng Fan <peng.fan@nxp.com> Signed-off-by: Tom Rini <trini@konsulko.com>
Tom Rini [Thu, 16 Jun 2022 18:04:35 +0000 (14:04 -0400)]
nxp: Make board/freescale/common/Kconfig safe to include once in arch/Kconfig
The way that we use this file currently means that we have to guard it
in every platform Kconfig. But it is also required in all NXP
platforms, including non-reference platforms. Make all options in it
have appropriate dependencies so that we can include it a single time
under arch/Kconfig
Tom Rini [Thu, 16 Jun 2022 18:04:34 +0000 (14:04 -0400)]
Convert CONFIG_HETROGENOUS_CLUSTERS et al to Kconfig
This converts the following to Kconfig:
CONFIG_HETROGENOUS_CLUSTERS
CONFIG_SYS_MAPLE
CONFIG_SYS_CPRI
CONFIG_PPC_CLUSTER_START
CONFIG_DSP_CLUSTER_START
CONFIG_SYS_CPRI_CLK
CONFIG_SYS_ULB_CLK
CONFIG_SYS_ETVPE_CLK
Tom Rini [Wed, 15 Jun 2022 16:03:49 +0000 (12:03 -0400)]
Rename CONFIG_PWM to CONFIG_PWM_S5P and move to Kconfig
We rename the S5P specific "CONFIG_PWM" to CONFIG_PWM_S5P and move it to
Kconfig. Given the usage of CONFIG_PWM_NX, we have that select this new
symbol.
Cc: Jaehoon Chung <jh80.chung@samsung.com> Cc: Minkyu Kang <mk7.kang@samsung.com> Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Tom Rini [Wed, 15 Jun 2022 16:03:46 +0000 (12:03 -0400)]
watchdog: designware: Make this depend on WDT
As this driver can dynamically determine the values set in
CONFIG_DW_WDT_BASE when using WDT, so make this depend on WDT rather
than migrate CONFIG_DW_WDT_BASE to Kconfig.
Cc: Chee Tien Fong <tien.fong.chee@intel.com> Cc: Chin-Liang See <chin.liang.see@intel.com> Cc: Dinh Nguyen <dinh.nguyen@intel.com> Cc: Holger Brunck <holger.brunck@hitachienergy.com> Cc: Ley Foon Tan <ley.foon.tan@intel.com> Cc: Marek Vasut <marex@denx.de> Cc: Siew Chin Lim <elly.siew.chin.lim@intel.com> Cc: Stefan Roese <sr@denx.de> Cc: hee Hong Ang <chee.hong.ang@intel.com> Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Stefan Roese <sr@denx.de>
Pali Rohár [Tue, 28 Jun 2022 15:54:00 +0000 (17:54 +0200)]
powerpc: mpc85xx: Simplify jump to _start_cont in flash code
After more patches code for jumping to _start_cont symbol in flash memory
involved to code with useless mathematical operations. Currently it does:
r3 := CONFIG_SYS_MONITOR_BASE + ABS(_start_cont) - CONFIG_SYS_MONITOR_BASE
jump to r3
Which is equivalent of just:
r3 := ABS(_start_cont)
jump to r3
The purpose of that code is just to jump to _start_code symbol,
independently of program counter. So branch must be done to absolute
address. Trying to write:
ba _start_cont
just cause linker error:
LD u-boot
powerpc-linux-gnuspe-ld.bfd: arch/powerpc/cpu/mpc85xx/start.o: in function `switch_as':
(.bootpg+0x4b8): relocation truncated to fit: R_PPC_ADDR24 against symbol `_start_cont' defined in .text section in arch/powerpc/cpu/mpc85xx/start.o
make: *** [Makefile:1801: u-boot] Error 1
Probably by the fact that absolute address cannot be expressed by 24-bits.
So write the code via mtlr+blr pattern as it was before and load general
purpose register with absolute address of the symbol:
lis r3,_start_cont@h
ori r3,r3,_start_cont@l
mtlr r3
blr
Seems that gcc and gnu ld linker support symbol@h and symbol@l syntax like
number@h and number@l without any problem. And disassembling of compiler
u-boot binary proved that lis+ori instructions are called with numbers
which represent halves of absolute address of _start_cont symbol.
Pali Rohár [Thu, 23 Jun 2022 16:22:21 +0000 (18:22 +0200)]
board: freescale: p1_p2_rdb_pc: Remove mapping for TDM-PMC card
From whole P1/P2 family of RDB boards is TDM-PMC card (PCI Mezzanine Card,
Freescale PQ-MDS-T1) available only on P1021RDB and P1025RDB boards.
So address mapping for TDM-PMC card on LBC should not be enabled on any
other P1/P2 RDB board as there is no device at that TDM-PMC address.
Support for P1021RDB and P1025RDB boards was already removed from mainline
U-Boot in commits 51370d733845 ("board/freescale: Remove P1021RDB board
support") and 0617c9fc9abf ("board/freescale: Remove P1025RDB board
support").
So do not enable TDM-PMC address mapping on remaining P1/P2 RDB boards and
remove all macros related to TDM-PMC address mappings.
DT node pic@40000 is defined explicitly in p2020-post.dtsi file and also
transitionally via include file pq3-mpic.dtsi. Remove duplicate definition
from p2020-post.dtsi.
Pali Rohár [Mon, 20 Jun 2022 11:07:03 +0000 (13:07 +0200)]
mtd: rawnand: fsl_elbc: Fix detection when nand_scan_ident() has not selected ecc.mode
ecc.mode is set to 0 (aliased to NAND_ECC_NONE) either when function
nand_scan_ident() has not selected ecc.mode or when it selected it to none
ecc mode.
Distinguish between these two states by checking of node property
"nand-ecc-mode" which function nand_scan_ident() uses for filling ecc.mode.
This change fixes usage of none ecc mode if it is specified in DTS file.
Fixes: 0fe929c4f70b ("mtd: rawnand: fsl_elbc: Use ECC configuration from device tree") Signed-off-by: Pali Rohár <pali@kernel.org>
Tom Rini [Thu, 30 Jun 2022 19:21:52 +0000 (15:21 -0400)]
Merge tag 'u-boot-at91-2022.10-a' of https://source.denx.de/u-boot/custodians/u-boot-at91 into next
First set of u-boot-at91 features for the 2022.10 cycle:
This feature set includes mostly fixes and alignments: DT alignment with
Linux for sama7g5, removal of invalid eeprom compatibles, removal of
extra debug_uart_init calls for all at91 boards, support for pio4 driver
pioE bank, and other minor fixes and enhancements for sam9x60 and
sama5d2_icp boards.
Tom Rini [Wed, 29 Jun 2022 13:54:02 +0000 (09:54 -0400)]
Merge branch '2022-06-28-Kconfig-migrations' into next
- Convert a large number of CONFIG symbols to Kconfig. Of note is a
large chunk of USB symbols (and dead code removal), ensuring all
SPL/TPL/VPL symbols have an appropriate dependency, largely (but not
entirely) removing the testing of CONFIG_SPL_BUILD in board headers,
and allowing CONFIG_EXTRA_ENV_TEXT and CONFIG_EXTRA_ENV_SETTINGS to
co-exist as this facilities migration of many platforms.
When the requested flash speed is 0, the baudrate division for the
requested speed causing drop in the performance. So set the ospi flash
to operate at max frequency when requested speed is zero.
T Karthik Reddy [Thu, 12 May 2022 10:05:34 +0000 (04:05 -0600)]
spi: cadence_qspi: Enable apb linear mode for apb read & write operations
On versal platform, enable apb linear mode for apb read and write
execute operations amd disable it when using dma reads. This is done by
xilinx_pm_request() secure calls when CONFIG_ZYNQMP_FIRMWARE is enabled,
else we use direct raw reads and writes in case of mini U-Boot.
T Karthik Reddy [Thu, 12 May 2022 10:05:33 +0000 (04:05 -0600)]
spi: cadence-qspi: reset qspi flash for versal platform
When flash operated at non default mode like DDR, flash need to be reset
to operate in SDR mode to read flash ids by spi-nor framework. Reset the
flash to the default state before using the flash. This reset is handled
by a gpio driver, in case of mini U-Boot as gpio driver is disabled, we
do raw read and write access by the registers.
Versal platform utilizes spi calibration for read delay programming, so
incase by default read delay property is set in DT. We make sure not to
use read delay from DT by overwriting read_delay with -1.
T Karthik Reddy [Thu, 12 May 2022 10:05:32 +0000 (04:05 -0600)]
arm64: versal: Add versal specific cadence ospi driver
Add support for cadence ospi driver for Versal platform. This driver
provides support for DMA read operation which utilizes cadence qspi
driver.
If "cdns,is-dma" DT property is specified use dma for read operation
from cadence_qspi driver. As cadence_qspi_apb_dma_read() is defined in
cadence_ospi_versal driver add a weak function defination in
cadence_qspi driver.
Tom Rini [Tue, 14 Jun 2022 02:57:37 +0000 (22:57 -0400)]
gw_ventana: Migrate to using CONFIG_EXTRA_ENV_TEXT
Move the environment text over from being set via
CONFIG_EXTRA_ENV_SETTINGS in include/configs/gw_ventana.h and over
to plain text in board/gateworks/gw_ventana/gw_ventana.env. This lets
us drop CONFIG_EXTRA_ENV_SETTINGS_COMMON as everything resides in a
single environment file now.
Cc: Tim Harvey <tharvey@gateworks.com> Signed-off-by: Tom Rini <trini@konsulko.com> Acked-by: Tim Harvey <tharvey@gateworks.com>
Tom Rini [Tue, 14 Jun 2022 02:57:36 +0000 (22:57 -0400)]
Complete migration of MTDPARTS_DEFAULT / MTDIDS_DEFAULT, include in environment
- Ensure that everyone setting mtdids= and mtdparts= is doing so via the
CONFIG options.
- If the CONFIG options are set, ensure that the default environment
sets mtdparts / mtdids.
Tom Rini [Tue, 14 Jun 2022 02:57:35 +0000 (22:57 -0400)]
Rename CONFIG_SYS_AUTOLAOD to CONFIG_SYS_DISABLE_AUTOLOAD
The "autoload" environment variable is always checked with env_get_yesno
as it can be set to any form of no. The default behavior of
env_get_yesno is to return -1 on variables that are not set, which acts
as true in general (we test for non-zero return). To convert
CONFIG_SYS_AUTOLOAD to Kconfig, given that it was almost always used to
set autoload to no, first rename to CONFIG_SYS_DISABLE_AUTOLOAD for
consistency sake. Then, make it so that if enabled we set autoload=0 in
the default environment. Migrate all platforms which set
CONFIG_SYS_AUTOLOAD to non-true or that set autoload to false in their
default environment to using CONFIG_SYS_DISABLE_AUTOLOAD
Tom Rini [Tue, 14 Jun 2022 02:57:34 +0000 (22:57 -0400)]
opos6uldev: Migrate to using CONFIG_EXTRA_ENV_TEXT
Move the environment text over from being set via
CONFIG_EXTRA_ENV_SETTINGS in include/configs/opos6uldev.h and over to
plain text in board/armadeus/opos6uldev/opos6uldev.env. This lets us
manage env_version without a CONFIG variable.
Cc: Sébastien Szymanski <sebastien.szymanski@armadeus.com> Signed-off-by: Tom Rini <trini@konsulko.com>