- nuvoton: add expire function for generic reset (Jim)
- handle watchdogs during keyed autoboot (Rasmus)
- cyclic: Don't disable cylic function upon exceeding CPU time (Stefan)
- ulp wdog: Updates to support iMX93 and DM (Alice)
John Keeping [Thu, 8 Sep 2022 11:38:35 +0000 (12:38 +0100)]
power: regulator: fix autoset error handling
If a regulator does not support .set_suspend_enable or
.set_suspend_value then ret is set to ENOSYS early in the function.
The most serious impact of this is that when no automatic setting of
voltage is needed then the final regulator_set_enable() is skipped
because ret has not been cleared.
It seems that the error handling for regulator_set_suspend_value() is
also wrong as if this succeeds then the normal boot-on checks are still
required, and again ENOSYS needs special treatment here.
Fixes: 11406b8f7e ("dm: regulator: support regulator more state") Signed-off-by: John Keeping <john@metanate.com>
Marek Vasut [Wed, 28 Sep 2022 16:45:04 +0000 (18:45 +0200)]
cmd: sf: Handle unaligned 'update' start offset
Currently the 'sf update' command fails in case the 'start' offset is
not aligned to SPI NOR erase block size. Add the missing alignment
calculation. In case the start offset is in the middle of erase block,
round start address down to the nearest aligned one, compare only the
updated data between what is in the SPI NOR and what is being written,
copy new data at offset of the compare buffer, and write back the entire
erase block.
This is useful e.g. on i.MX6Q where the u-boot-with-spl.imx is at
offset 0x400 in the SPI NOR, while the SPI NOR may have erase block
size e.g. 0x1000 bytes.
Signed-off-by: Marek Vasut <marex@denx.de> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
Tom Rini [Tue, 25 Oct 2022 01:28:47 +0000 (21:28 -0400)]
Merge tag 'fsl-qoriq-2022-10-18' of https://source.denx.de/u-boot/custodians/u-boot-fsl-qoriq
Layerscape update
- support sysreset,
- de-select FSL_IFC when booting from SD
- disable unused parts of ICID tables
- reduce ns_dev size for csu
- enable dma snooping for ls104x
- nand driver fixups for ls1043ardb rev 7.0 boards.
- for 2022.01
- rework Kontron boards (mx6 and mx8)
- fixes for Toradex
- fixes (SPI, CAAM, )
- sync DT with Linux
- fixes for Gateworks GW7903 and GW7904 PMIC
- Engicam i.Core MX8M Plus EDIMM2.2
Marcel Ziswiler [Sat, 22 Oct 2022 21:59:45 +0000 (23:59 +0200)]
imx28: avoid num_cs and spi_max_frequency build errors
Avoid the following build errors after the device tree sync:
drivers/spi/mxs_spi.c: In function ‘mxs_spi_probe’:
drivers/spi/mxs_spi.c:327:25: error: ‘struct dtd_fsl_imx23_spi’ has no
member named ‘spi_max_frequency’
327 | priv->max_freq = dtplat->spi_max_frequency;
| ^~
drivers/spi/mxs_spi.c:328:23: error: ‘struct dtd_fsl_imx23_spi’ has no
member named ‘num_cs’
328 | plat->num_cs = dtplat->num_cs;
| ^~
Adam Ford [Sat, 22 Oct 2022 00:58:31 +0000 (19:58 -0500)]
arm: dts: imx8mn-u-boot: Create common imx8mn-u-boot.dtsi
Multiple boards create duplicate entries in their respective
-u-boot.dtsi files which all basically do the same thing.
To consolidate these and make it easier to make improvements
going forward, consolidate them all into one place.
This file creates a flash.bin image using binman, and supports
LPDDR4, DDR4 and DDR3. Since individual boards use different
peripherals and different UART ports, those entries were kept
in their respective board files, but the spba1 node was addded
which contains all UART1-3 to help facilitate SPL_DM_SERIAL.
Individual users will still need to include their respective
UART and pinctrl nodes for those UARTS.
This consolidated file also supports generating a flash.bin file
which can boot from flexSPI if CONFIG_FSPI_CONF_HEADER is
enabled.
Signed-off-by: Adam Ford <aford173@gmail.com> Reviewed-by: Fabio Estevam <festevam@denx.de>
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.
verdin-imx8mm: various config additions and improvements
- enable bootcount command
- integrate bootcount using SNVS_LP general purpose register LPGPR0
- enable link-time optimisation
- explicitly set a boot delay of one second
- enable CRC32 and MD5
- enable command for low-level access to data in a partition
- enable time commands
- enable PMIC commands
- improve ETHPRIME configuration
- enable eMMC HS400 functionality
- enable fixed PHY and MDIO driver model
- remove stale PFUZE100 PMIC driver
- enable thermal management unit driver
- enable more USB host functionality
- enable hexdump
Update the distro config env memory layout for the Verdin iMX8M Mini and
Verdin iMX8M Plus again:
- loadaddr=0x48200000 allows for 128MB area for uncompressing (ie FIT
images, kernel_comp_addr_r, kernel_comp_size)
- fdt_addr_r = loadaddr + 128MB - allows for 128MB kernel
- scriptaddr = fdt_addr_r + 512KB - allows for 512KB fdt
- ramdisk_addr_r = scriptaddr + 512KB - allows for 512KB script
Memory layout taken from commit fd5c7173ade4
("imx8m{m,n}_venice: update env memory layout").
Note that for our regular BSP Layers and Reference Images for Yocto
Project an updated distro boot script is required (see
meta-toradex-bsp-common/recipes-bsp/u-boot/u-boot-distro-boot).
Note that this corrects a pre-maturely applied version 2 of the same
patch set.
Tim Harvey [Wed, 14 Sep 2022 16:02:19 +0000 (09:02 -0700)]
board: gateworks: venice: add imx8mm-gw7904 support
The GW7904 is based on the i.MX 8M Mini SoC featuring:
- LPDDR4 DRAM
- eMMC FLASH
- microSD connector with UHS support
- LIS2DE12 3-axis accelerometer
- Gateworks System Controller
- IMX8M FEC
- 2x RS232 off-board connectors
- PMIC
- 10x bi-color LED's
- 1x miniPCIe socket with PCIe and USB2.0
- 802.3at Class 4 PoE
- 10-30VDC input via barrel-jack
Genaral features:
- LCD 7" C.Touch
- microSD slot
- Ethernet 1Gb
- Wifi/BT
- 2x LVDS Full HD interfaces
- 3x USB 2.0
- 1x USB 3.0
- HDMI Out
- Plus PCIe
- MIPI CSI
- 2x CAN
- Audio Out
i.Core MX8M Plus is an EDIMM SoM based on NXP i.MX8M Plus from Engicam.
i.Core MX8M Plus needs to mount on top of this Evaluation board for
creating complete i.Core MX8M Plus EDIMM2.2 Starter Kit.
Add support for it.
Sync the i.Core MX8M Plus is an EDIMM SoM based on NXP
devicetree file from linux-next tree.
commit <aec8ad34f7f24> (arm64: dts: imx8mp: Add Engicam i.Core MX8M Plus EDIMM2.2 Starter Kit)
Signed-off-by: Manoj Sai <abbaraju.manojsai@amarulasolutions.com> Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Manoj Sai [Fri, 26 Aug 2022 12:33:35 +0000 (18:03 +0530)]
arm64: dts: imx8mp: Add Engicam i.Core MX8M Plus SoM
i.Core MX8M Plus is an EDIMM SoM based on NXP i.MX8M Plus
from Engicam.
General features:
- NXP i.MX8M Plus
- Up to 4GB LDDR4
- 8 eMMC
- Gigabit Ethernet
- USB 3.0, 2.0 Host/OTG
- PCIe 3.0 interface
- I2S
- LVDS
- rest of i.MX8M Plus features
i.Core MX8M Plus needs to mount on top of Engicam baseboards
for creating complete platform solutions.
Add support for it.
Sync the i.Core MX8M Plus is an EDIMM SoM based on NXP i.MX8M Plus
from Engicam devicetree file from linux-next tree.
commit <eefe06b295087> (arm64: dts: imx8mp: Add Engicam i.Core MX8M Plus SoM)
Signed-off-by: Manoj Sai <abbaraju.manojsai@amarulasolutions.com> Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Alice Guo [Fri, 21 Oct 2022 08:41:18 +0000 (16:41 +0800)]
watchdog: ulp_wdog: add driver model for ulp watchdog driver
Enable driver model for ulp watchdog timer. When CONFIG_WDT=y and the
status of device node is "okay", initr_watchdog will be called and
finally calls ulp_wdt_probe() and ulp_wdt_start().
Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Ye Li <ye.li@nxp.com> Reviewed-by: Stefan Roese <sr@denx.de>
Alice Guo [Fri, 21 Oct 2022 08:41:17 +0000 (16:41 +0800)]
watchdog: ulp_wdog: enable watchdog interrupt on imx93
The reset source of the external PMIC on i.MX93 is WDOG_ANY PAD and the
source of WDOG_ANY PAD is interrupt. Therefore, using PMIC to reset
needs to enable the watchdog interrupt.
Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Stefan Roese <sr@denx.de>
Alice Guo [Fri, 21 Oct 2022 08:41:16 +0000 (16:41 +0800)]
watchdog: ulp_wdog: Update watchdog driver for imx93
The WDOG clocks are sourced from the fixed 32KHz (lpo_clk).When the
timeout period exceeds 2 seconds, the value written to the TOVAL
register is larger than 16-bit can represent. Enabling watchdog
prescaler to solve this problem.
Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Stefan Roese <sr@denx.de>
Ye Li [Fri, 21 Oct 2022 08:41:15 +0000 (16:41 +0800)]
ulp_wdog: Update ulp wdog driver for 32bits command
To use 32bits refresh and unlock command as default, check the CMD32EN
bit to select the corresponding commands.
Signed-off-by: Ye Li <ye.li@nxp.com> Signed-off-by: Alice Guo <alice.guo@nxp.com> Reviewed-by: Peng Fan <peng.fan@nxp.com> Reviewed-by: Stefan Roese <sr@denx.de>
Stefan Roese [Mon, 17 Oct 2022 07:00:58 +0000 (09:00 +0200)]
cyclic: Don't disable cylic function upon exceeding CPU time
With the migration of the watchdog infrastructure to cyclic functions
it's been noticed, that at least one watchdog driver is broken now. As
the execution time of it's watchdog reset function is quite long.
In general it's not really necessary (right now) to disable the cyclic
function upon exceeding CPU time usage. So instead of disabling the
cylic function in this case, let's just print a warning once to show
this potential problem to the user.
Signed-off-by: Stefan Roese <sr@denx.de> Suggested-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Cc: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Cc: Tom Rini <trini@konsulko.com> Cc: Pali Rohár <pali@kernel.org>
In order to test that we properly handle watchdog(s) during the "wait
for the user to interrupt autoboot" phase, we need a watchdog device
to be watching us.
In order to test that U-Boot actually maintains the watchdog device(s)
during long-running busy-loops, such as those where we wait for the
user to stop autoboot, we need a watchdog device that actually does
something during those loops; we cannot test that behaviour via the DM
test framework.
So introduce a relatively simple watchdog device which is simply based
on calling the host OS' alarm() function; that has the nice property
that a new call to alarm() simply sets a new deadline, and alarm(0)
cancels any existing alarm. These properties are precisely what we
need to implement start/reset/stop. We install our own handler so that
we get a known message printed if and when the watchdog fires, and by
just invoking that handler directly, we get expire_now for free.
The actual calls to the various OS functions (alarm, signal, raise)
need to be done in os.c, and since the driver code cannot get access
to the values of SIGALRM or SIG_DFL (that would require including a
host header, and that's only os.c which can do that), we cannot simply
do trivial wrappers for signal() and raise(), but instead create
specialized functions just for use by this driver.
Apart from enabling this driver for sandbox{,64}_defconfig, also
enable the wdt command which was useful for hand-testing this new
driver (especially with running u-boot under strace).
This is a companion to u-boot,noautostart. If one has a single
watchdog device that one does want to have auto-started, but several
others that one doesn't, the only way currently is to set the
CONFIG_WATCHDOG_AUTOSTART and then use the opt-out for the majority.
The main motivation for this is to add an autostarted watchdog device
to the sandbox (to test a fix) without having to set AUTOSTART in
sandbox_defconfig and add the noautostart property to the existing
devices. But it's also nice for symmetry, and the logic in
init_watchdog_dev() becomes simpler to read because we avoid all the
negations.
The SDMMC IOs can be in an IO domain, that has to be enabled.
This is done by enabling vqmmc in the driver.
This has no impact on configurations not using an IO domain, the check
can then be executed on all platforms managing regulator, and the vqmmc
regulator enabled on all platforms having it in their DT.
mmc: stm32_sdmmc2: protect against unsupported modes
The UHS modes for SD, HS200 and HS400 modes for eMMC are not supported
by the stm32_sdmmc2 driver.
Make it clear by removing the corresponding caps after parsing the DT.
To support dual data rate with STM32 sdmmc2 driver, the dedicated bit
(DDR - BIT(18)) needs to be set in the CLKRC register. Clock bypass
(no divider) is not allowed in this case. This is required for the
eMMC DDR modes.
Add Socionext F_SDH30_E51 IP support. The features of this IP includes
CMD/DAT line delay and force card insertion mode for non-removable cards.
And the IP needs to add some quirks.
Sergei Antonov [Fri, 2 Sep 2022 07:40:10 +0000 (10:40 +0300)]
mmc: ftsdc010: make command timeout 250 ms as in the comment
Get rid of discrepancy beween comment /* 250 ms */ and code
which shifts by 4 thus dividing by 16.
So change code to shift by 2 and make the timeout value 250 ms.
Signed-off-by: Sergei Antonov <saproj@gmail.com> Reviewed-by: Rick Chen <rick@andestech.com> Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
mtd: spi-nor-core: Rework s25hx_t_post_bfpt_fixup() for flash's internal address mode
The flash's internal address mode is tracked by nor->add_mode_nbytes and
it is set to 3 in BFPT parse. SEMPER multi-die package parts (>1Gb) are
3- or 4-byte address mode by default, depending on model number. We need
to make sure that 4-byte address mode is used for multi-die package parts.
For single-die package parts (<=1Gb), registers can be accessed by 3-byte
address. Read, program, and erase use the 4B opcodes that always take
4-byte address regardless of flash's internal address mode.
Signed-off-by: Takahiro Kuwano <Takahiro.Kuwano@infineon.com> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
mtd: spi-nor-core: Rework spansion_read/write_any_reg() to use addr_mode_nbytes
Read/Write Any Register commands take 3- or 4- byte address depending on
flash's internal address mode. The nor->addr_width tracks number of
address bytes used in read/program/erase ops that can be 4
(with 4B opcodes) regardless of flash's internal address mode. The
nor->addr_mode_nbytes tracks flash's internal address mode so replace
nor->addr_width by that.
Signed-off-by: Takahiro Kuwano <Takahiro.Kuwano@infineon.com> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
The nor->addr_width tracks number of address bytes used in
read/program/erase ops and eventually set to 4 for >16MB chips, regardless
of flash's internal address mode. For Infineon SEMPER flash's, we use
Read/Write Any Register commands for configuration and status check.
These commands take 3- or 4-byte address depending on flash's internal
address mode.
Signed-off-by: Takahiro Kuwano <Takahiro.Kuwano@infineon.com> Reviewed-by: Jagan Teki <jagan@amarulasolutions.com>
- NC-SI handling support and enable on evb-ast2[56]00, gpio driver for
ADP5585, improve qfw support, print more sysresets info, gw_ventana
and gcc-12 bugfix, improve BCB support, fix a few typos and remove an
unused keymile CONFIG symbol.
For some blk operations, it's possible that a different hw partition
gets selected via blk_dselect_hwpart().
In that case, only the region of the device covered by that partition
is accessible.
This breaks "bcb load" which attempts to read the gpt and assumes it's
on the user(0) hw partition:
board: gateworks: gw_ventana: fix building with GCC 12.2
Building with GCC 12.2 results in an error
board/gateworks/gw_ventana/gw_ventana.c:636:68: error: the comparison
will always evaluate as 'true' for the address of 'pwm_padmux' will
never be NULL [-Werror=address]
636 | } else if (hwconfig_subarg_cmp(arg, "mode", "pwm") &&
| ^~
Remove the superfluous check.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com> Reviewed-By: Tim Harvey <tharvey@gateworks.com> Reviewed-by: Fabio Estevam <festevam@denx.de>
Michal Suchanek [Mon, 10 Oct 2022 18:29:40 +0000 (20:29 +0200)]
common: board_f: Print information for all sysresets
Boards can have multiple sysresets, iterate all when printing sysreset
info.
Fixes: 23471aed5c ("board_f: Add reset status printing") Signed-off-by: Michal Suchanek <msuchanek@suse.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Andre Przywara [Mon, 10 Oct 2022 17:01:22 +0000 (18:01 +0100)]
qemu: Try to automatically boot from the QEMU firmware device (qfw)
At the moment the QEMU boot sequence tries various (storage) devices
when trying to find a payload to boot.
To simplify starting a specific kernel and initrd, there is also the qfw
command, which can use the files specified on the QEMU command line, via
the -kernel and -initrd options.
Add this command to the list of boot options to try. Since users
specifying those options on the command line probably explicitly want
to run them, let's place the new command first. Without those options,
the qfw command will just gracefully fail, and we continue with the
existing order.
This allows auto-booting of specific kernels in QEMU, for instance in CI
systems.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Andre Przywara [Mon, 10 Oct 2022 17:01:21 +0000 (18:01 +0100)]
qfw: return failure when no kernel could be loaded
When we try to load a kernel via the QEMU firmware device, we currently
"return -1;" if no kernel was specified on the QEMU command line. This
leads to the usage output, which is confusing (since nothing on the
command line was really wrong), but also somewhat hides the actual error
message.
Return CMD_RET_FAILURE (1), as it's a proper error, and make the message
more clear that this is not only a "warning".
This helps to call this command in boot scripts, and to gracefully
continue if this doesn't work.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Andre Przywara [Mon, 10 Oct 2022 17:01:20 +0000 (18:01 +0100)]
qfw: store loaded file size in environment variable
At the moment the QEMU firmware command just prints the size of the
loaded binaries on the console.
To go with all the other load methods, and many boot scripts'
expectations, also store the size of the file loaded last in the
environment variable "filesize".
We first put the kernel size in there, but overwrite this with the
initrd size, should we have one, because this is probably the more
prominent user of $filesize (in the booti or bootz command).
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Alice Guo [Sun, 9 Oct 2022 03:19:22 +0000 (11:19 +0800)]
gpio: adp5585: add gpio driver for ADP5585 I/O Expander Controller
Add gpio driver for ADP5585 I/O Expander Controller. The ADP5585 is a 10
input/output port expander and can be used to increase the number of
I/Os available to a processor.
Update the ftgmac100 driver to support NC-SI instead of an mdio phy
where available. This is a common setup for Aspeed AST2x00 platforms.
NC-SI mode is determined from the device-tree if either phy-mode sets it
or the use-ncsi property exists. If set then normal mdio setup is
skipped in favour of the NC-SI phy.
Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com> Signed-off-by: Joel Stanley <joel@jms.id.au> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Add the handling of NC-SI ethernet frames, and add a check at the start
of net_loop() to configure NC-SI before starting other network commands.
Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com> Signed-off-by: Joel Stanley <joel@jms.id.au> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
- Enforce CONFIG_DM being enabled (which has been the case for all
boards for a bit now) and remove non-DM_KEYBOARD options as they're also
unused for some time now.
Fabio Estevam [Mon, 3 Oct 2022 13:59:24 +0000 (10:59 -0300)]
imx8mn-venice-u-boot: Fix broken boot
When the imx8mm.dtsi file was pulled in from Linux, the UARTs
were moved into an spba sub-node which wasn't being included
in the SPL device tree. This meant the references to the UART
weren't being handled properly and when booting the system would
constantly reboot. Fix this by adding the spba node to the spl
device tree to restore normal booting.
Based on the patch from Adam Ford for the imx8mn-beacon-kit-u-boot
board.
Fixes: 4e5114daf9eb ("imx8mn: synchronise device tree with linux") Signed-off-by: Fabio Estevam <festevam@denx.de> Reviewed-by: Michael Trimarchi <michael@amarulasolutions.com>
Fabio Estevam [Mon, 3 Oct 2022 13:39:21 +0000 (10:39 -0300)]
imx8mn-ddr4-evk-u-boot: Fix broken boot
When the imx8mm.dtsi file was pulled in from Linux, the UARTs
were moved into an spba sub-node which wasn't being included
in the SPL device tree. This meant the references to the UART
weren't being handled properly and when booting the system would
constantly reboot. Fix this by adding the spba node to the spl
device tree to restore normal booting.
Based on the patch from Adam Ford for the imx8mn-beacon-kit-u-boot
board.
Fixes: 4e5114daf9eb ("imx8mn: synchronise device tree with linux") Signed-off-by: Fabio Estevam <festevam@denx.de> Reviewed-by: Michael Trimarchi <michael@amarulasolutions.com>
Marek Vasut [Wed, 28 Sep 2022 16:45:18 +0000 (18:45 +0200)]
ARM: imx6: dh-imx6: Use sf update instead of sf erase and write
With sf update fixed to support unaligned start offset, use plain
sf update to update the bootloader in SPI NOR. This also helps
avoid the case where not enough SPI NOR has been erased and the
bootloader has been written to unerased area, and thus corrupted.
Signed-off-by: Marek Vasut <marex@denx.de> Reviewed-by: Peng Fan <peng.fan@nxp.com>
Marek Vasut [Wed, 28 Sep 2022 15:40:11 +0000 (17:40 +0200)]
ARM: imx: Enable btrfs support on DHSOM
The btrfs filesystem provides advanced functionality like copy-on-write
and snapshots, as well as metadata and data duplication and checksumming.
Enable btrfs in U-Boot to permit even the primary partition to be btrfs
and let system boot from it.
Signed-off-by: Marek Vasut <marex@denx.de> Reviewed-by: Peng Fan <peng.fan@nxp.com>
Update the distro config env memory layout for the Apalis iMX6,
Colibri iMX6, Colibri iMX6ULL and Colibri iMX7:
- loadaddr=0x84200000 (resp. 0x14200000 on them i.MX 6) allows for 64MB
area for uncompressing (ie FIT images)
- fdt_addr_r = loadaddr + 64MB : allows for 64MB kernel
- scriptaddr = fdt_addr_r + 512KB : allows for 512KB fdt
- pxefile_addr_r = scriptaddr + 512KB : allows for 512KB script
- ramdisk_addr_r = pxefile_addr_r + 512KB : allows for 1MB extlinux.conf
Memory layout analogous to 64-bit one from commit fd5c7173ade4
("imx8m{m,n}_venice: update env memory layout") but left pxefile_addr_r
updated according to doc/develop/distro.rst.
This fixes a potential issue caused by the compressed kernel being
relocated on top of the ramdisk causing its corruption.