Add default environment variables needed for Standard Boot enablement as
described in [1]. Also rework the eMMC partition table for Linux boot so
it only has two partitions:
1. EFI System Partition (EFI vars, GRUB efi app, firmware files)
2. rootfs partition (Debian rootfs, /boot, extlinux.conf, boot.scr)
Both partitions are made bootable so that 'bootflow' command can detect
all loader files (rootfs might contain extlinux.conf and boot.scr).
'ldfw' partition is removed too, as ldfw.bin can be loaded from ESP now
(from /EFI/firmware/ldfw.bin).
Android partitons will be added later, once Android boot is actually
enabled for E850-96.
Notes:
- $kernel_comp_addr_r uses the same address (0x88000000) as LDFW
buffer (in board/samsung/e850-96/fw.c), but that's fine, as LDFW
will be copied to another RAM location (Secure World) by SMC
command, so it's only used temporarily on startup
- addition assignment (+=) operation is used for $partitions to avoid
spaces added by newlines, so that $partitions can be used in the
shell with no quotes
Now it's possible to successfully automatically boot Debian rootfs:
=> env default -f -a
=> env save
=> gpt write mmc 0 $partitions
=> reset
[1] doc/develop/bootstd/overview.rst
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
LDFW firmware loading can fail, e.g. in case if user forgot to upload
the binary to the appropriate location (/EFI/firmware/ldfw.bin on ESP
partition). Report such errors explicitly, so that the user can notice
it early and take necessary actions. But don't return error code from
board_init() in this case, as LDFW firmware is not mandatory for board
operation and is only required for some features like TRNG.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
In case when EFI System Partition is present it can be used to store
firmware binaries, instead of keeping those on separate dedicated
partitions. That simplifies the partition table and makes it more
standard. Rework the firmware loader code to look for LDFW binary at
/EFI/firmware/ldfw.bin on ESP first, and if either the partition or the
file doesn't exist -- fallback to reading it from 'ldfw' partition. This
way backward compatibility can be kept, and Android partition tables
without ESP partition can be handled too.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Enable Standard Boot on E850-96 as documented in [1]. Along with
corresponding changes in the default environment and properly prepared
eMMC partitions (ESP and rootfs), it makes it possible to boot Debian
rootfs automatically. All boot methods were tested: efi_mgr, efi,
syslinux and script. The preferred boot method is efi_mgr, which relies
on the configured EFI variables (stored in /ubootefi.var file on ESP
partition), which boots either GRUB from /EFI/debian/grubaa64.efi, or
systemd-boot from /EFI/BOOT/BOOTAA64.EFI. Currently used boot sequence:
U-Boot -> bootcmd -> 'bootflow scan -lb' -> efi_mgr -> GRUB -> Debian
[1] doc/develop/bootstd/overview.rst
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
The basic EFI support is already enabled by EFI_LOADER. Follow
SystemReady IR recommendations [1,2] for U-Boot and enable support for
more EFI features. That includes:
- CONFIG_CMD_BOOTEFI_SELFTEST: support for "bootefi selftest" command
- CONFIG_CMD_NVEDIT_EFI: support for "env -e" to explore EFI vars
- CONFIG_CMD_EFIDEBUG: support for "efidebug" command
Also enable RTC support:
- CONFIG_EFI_SET_TIME
- CONFIG_CMD_RTC
- CONFIG_DM_RTC
- CONFIG_RTC_EMULATION
[1] https://developer.arm.com/documentation/DUI1101/1-1/Configure-U-Boot-for-SystemReady
[2] https://developer.arm.com/documentation/DUI1101/1-1/Test-SystemReady-IR
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Store U-Boot environment in BOOT2 HW area of eMMC (/dev/mmcblk0boot1),
as it's currently unused. BOOT1 area will be probably used for storing
low-level bootloaders further, so let's not touch it.
Both primary and redundant environments work fine:
=> env default -f -a
## Resetting to default environment
=> env save
Saving Environment to MMC... Writing to redundant MMC(0)... OK
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
The E850-96 default environment is going to be updated soon, requiring
next config options to be enabled:
- CONFIG_DEFAULT_FDT_FILE: $fdtfile will use this value
- CONFIG_PARTITION_TYPE_GUID: $partitions will rely on "type" feature
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Update the debug() print, use __func__ to always print matching
function name, and also print bus name in case there are multiple
busses.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com>
Replace ifdeffery with plain debug() function call. No functional change.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com>
Add support for the R-Car S4 Starter Kit with R8A779F4 SoC support.
This implementation natively uses OF_UPSTREAM to pull in most recent
DT. The defconfig is derived from S4 Spider, with reduced UART baud
rate to 921600 Bdps. The DT alias to rswitch is removed as the alias
should point to rswitch ports, not to rswitch itself, see [1].
[1] https://lore.kernel.org/linux-arm-kernel/20250118111344.361617-5-marek.vasut+renesas@mailbox.org/
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Explicitly flush icache on the CR52 core before jumping to the next
stage software to make sure it does not contain any invalid content.
Explicitly flash and invalidate dcache on the CA76 core both over the
trampoline buffer and over the CR52 firmware, and then trigger full
system synchronization, to make sure the data surely land in DRAM,
from where the CR52 can surely pick them up.
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Whereas with Azure the JUnit results file is available for download,
Gitlab doesn't default to including it as an artifact to download and
only makes it available via its own JUnit parser. Fix this by listing it
as an artifact to save as well.
Signed-off-by: Tom Rini <trini@konsulko.com>
Now that we can run sandbox on arm64 hosts, have these jobs run on both
the fast arm64 and amd64 hosts to catch any issues.
Signed-off-by: Tom Rini <trini@konsulko.com>
Refer to the commit a3c101a61348 ("examples: fix building on arm64")
fix for arm32. crt0.S does not even build with:
operating system and architecture:
5.15.0-130-generic x86_64
installed version of gcc and binutils:
gcc version 9.4.0 (Ubuntu 9.4.0-1ubuntu1~20.04.3)
GNU ld (GNU Binutils for Ubuntu) 2.34
used defconfig:
stm32h750-art-pi_defconfig
used commands for building:
export CROSS_COMPILE=arm-none-eabi- ARCH=arm
make stm32h750-art-pi_defconfig
make menuconfig # to add
make
examples/api/crt0.S: Assembler messages:
examples/api/crt0.S:32: Error: lo register required -- `ldr ip,=search_hint'
examples/api/crt0.S:33: Error: lo register required -- `str sp,[ip]'
examples/api/crt0.S:39: Error: lo register required -- `ldr ip,=syscall_ptr'
examples/api/crt0.S:40: Error: lo register required -- `ldr pc,[ip]'
make[2]: *** [scripts/Makefile.build:292：examples/api/crt0.o] Error 1
Use valid register names and instructions.
Signed-off-by: Liya Huang <1425075683@qq.com>
CONFIG_STANDALONE_LOAD_ADDR has been used for examples/standalone
but not for examples/api.
The suitability of an address to load an ELF binary and run it does
not only depend on the architecture but also on the memory layout of
the board which is not reflected in the default value of
CONFIG_STANDALONE_LOAD_ADDR.
Commit 32b7e39db4d3 ("Convert CONFIG_STANDALONE_LOAD_ADDR to Kconfig")
set the default on RISC-V to 0x0 though most boards used 0x80200000
before the patch.
On most boards we can assume 8 MiB of memory available above $loadaddr.
So we can safely use $loadaddr + 4 MiB as load address for the standalone
example and eliminate CONFIG_STANDALONE_LOAD_ADDR altogether.
Fixes: 32b7e39db4d3 ("Convert CONFIG_STANDALONE_LOAD_ADDR to Kconfig")
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
The current load address for the 'demo' binary does not work for
qemu_arm_defconfig.
The suitability of an address to load an ELF binary and run it does
not only depend on the architecture but also on the memory layout of
the board.
On most boards we can assume that 8 MiB of memory is available above
$loadaddr. So $loadaddr + 0x400000 should work there.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
A FIT image which is NOT using -E when created by mkimage - that is with
image data within the FIT - will fail to apply FDTO if the base FDT
image node does not specify a load property (which points to an address
in DRAM). This is because we check that the FDT address we want to apply
overlay to (i.e. modify and likely increase in size) is not inside the
FIT and give up otherwise. This is assumed necessary because we may then
overwrite other data when applying in-place.
However, we can do better than giving up: relocating the FDT in another
place in DRAM where it's safe to increase its size and apply FDTOs.
While at it, do not discriminate anymore on whether the data is within
the FIT data address space - that is FIT images created with mkimage -E
- as that still may be susceptible to unintended data overwrites as
mkimage -E simply concatenates all blobs after the FIT. If the FDT blob
isn't the last, it'll result in overwriting later blobs when resizing.
The side effect is that the load property in the FIT is only
temporarily used to load the FDT but then relocated right before we
start applying overlays.
Suggested-by: Marek Vasut <marex@denx.de>
Reviewed-by: Marek Vasut <marex@denx.de>
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de>
Christian Marangi <ansuelsmth@gmail.com> says:
This is the last batch of part to push actual support of
OF_UPSTREAM for the mediatek SoC.
The plan is to make the current downstream DTS on part with
upstream implementation so we can permit a gradual transition to
it while we don't cause any regression to any user.
This is to have the same node downstream and upstream.
Mediatek is working hard upstream to also push all the remaining
nodes.
All patch are the final changes after the pinctrl patch
merged previously.
All patch pass CI tests
Link: https://github.com/u-boot/u-boot/pull/731
Link: https://lore.kernel.org/r/20250127134046.26345-1-ansuelsmth@gmail.com
Add PCIe node for MT7981 with all the required properties to make PCIe
work.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Add USB PHY nodes for MT7981. These are needed for USB support and also
for PCIe support as the u3 PHY can also be used for PHY.
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Move gpio-controller up to pinctrl node and rename label to "pio" to
match the label used in upstream kernel linux.
Update any DTS that reference the pinctrl to follow the new naming and
structure.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Move gpio-controller up to pinctrl node and rename label to "pio" to
match the label used in upstream kernel linux.
Update any DTS that reference the pinctrl to follow the new naming and
structure.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Rename reg-names to upstream linux format. Upstream linux drop the
"_base". To make use of upstream DTSI, align to the upstream naming.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Move gpio-controller up to pinctrl node and rename label to "pio" to
match the label used in upstream kernel linux.
Update any DTS that reference the pinctrl to follow the new naming and
structure.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Rename reg-names to upstream linux format. Upstream linux drop the
"_base". To make use of upstream DTSI, align to the upstream naming.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Rename reg-names to upstream linux format. Upstream linux drop the
"_base". To make use of upstream DTSI, align to the upstream naming.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Weijie Gao <weijie.gao@mediatek.com>
Tested-by: Weijie Gao <weijie.gao@mediatek.com>
Simon Glass <sjg@chromium.org> says:
This includes the VBE ABrec (A/B/recovery) implementation as well as a
number of patches needed to make it work:
- marking some code as used by SPL_RELOC
- selection of images from a FIT based on the boot phase
- removal of unwanted hash code which increases code-size too much
- a few Kconfig-related additions for VPL
Note: The goal for the next series (part H) is to enable VBE on
rk3399-generic, i.e. able to boot on multiple rk3399-based boards with
only the TPL phase being different for each board.
Link: https://lore.kernel.org/r/20250126184333.4058848-1-sjg@chromium.org/
So far only VBE-simple is implemented in U-Boot. This supports a single
image which can be updated in situ.
It is often necessary to support two images (A and B) so that the board
is not bricked if the update is interrupted or is bad.
In some cases, a non-updatable recovery image is desirable, so that the
board can be returned to a known-good state in the event of a serious
failure.
Introduce ABrec which provides these features. It supports three
independent images and the logic to select the desired one on boot.
While we are here, fix a debug message to indicate the function it
called. Provide a maintainers entry for VBE.
Note that fwupdated only supports VBE-simple so far, but supports for
ABrec will appear in time.
Signed-off-by: Simon Glass <sjg@chromium.org>
When VBE operates within VPL it does not want the FDT to be changed.
Provide a way to disable this feature.
Move the FIT_IMAGE_TINY condition out of spl_fit_record_loadable() so
that both conditions are together. This makes the code easier to
understand.
Replace the existing fit_loaded member, which is no-longer used.
Signed-off-by: Simon Glass <sjg@chromium.org>
Join the comment block for the fit_image_load() call back to where it
should be. Also fix a debug statement.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add support for moving from TPL->VPL->SPL so that the VPL build can fit
properly into the boot flow.
Use #ifdefs to avoid creating unwanted symbols which Binman would then
try (and perhaps fail) to provide.
Add debugging to indicate the next phase.
Signed-off-by: Simon Glass <sjg@chromium.org>
Mark these functions as needed by relocation. These functions are used
to copy data while relocating the next-phase image.
Drop the 'safe' versions from SPL as they are not needed. Change the
static array to a local one, to avoid link errors when trying to access
the data.
Signed-off-by: Simon Glass <sjg@chromium.org>
Mark the lz4 decompression code as needed by relocation. This is used to
decompress the next-phase image.
Drop the 'safe' versions from SPL as they are not needed. Change the
static array to a local one, to avoid a crash errors when trying to
access the data from relocated code. Make this conditional to avoid a
code-size increase when SPL_RELOC is not used/
Signed-off-by: Simon Glass <sjg@chromium.org>
Mark the crc8 code as needed by relocation. This is used as a simple
check against corruption of the code when copying.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add a linker symbol which can be used to mark relocation code, so it can
be collected by the linker and copied into a suitable place and executed
when needed.
Signed-off-by: Simon Glass <sjg@chromium.org>
Some phases may wish to use full FIT-loading and others not, so allow
this to be controlled.
Add some debugging while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
With SPL we want to specify the phase of the image to be loaded. Add
support for this.
This is the implementation of a FIT feature added to the spec a few
years ago and entails a small code-size increase, about 70 bytes on
Thumb2.
Signed-off-by: Simon Glass <sjg@chromium.org>
Link: https://docs.u-boot.org/en/latest/usage/fit/index.html
At present the simple FIT-loader always loads images, ignoring whether
they are intended for the next phase or not.
VBE packages up several images in the same FIT, some destined for VPL
and some for SPL. Add logic to check the phase before loading the
image. Return -EPERM in that case and handle it gracefully.
Fix a unnecessary re-computation of read_offset while here.
Signed-off-by: Simon Glass <sjg@chromium.org>
Update the build rule so that hash algorithms are only included in an
SPL build if they are requested. This helps to reduce code size.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
The size of some malloc() fields has reduced on 64-bit machines, but
the spl_reloc code was not updated. Fix this to avoid a compiler
warning.
Also update for the new xPL naming.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
This code is not necessarily needed in VPL, even if SPL uses it, so
adjust the rules to allow it to be dropped.
Do the same for the hash API.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
VPL may want to use driver model for MMC even if TPL does not. Update
the rule in this driver to support that.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
JH7110 SoC requires no specific handling before entering Linux kernel.
Let's drop the specific implementation to avoid duplication.
Signed-off-by: Yao Zi <ziyao@disroot.org>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
The current implementation is equivalent to the fallback one, so
this shouldn't change any behaviour but cleans the code up only.
Signed-off-by: Yao Zi <ziyao@disroot.org>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
