Document the image preparation and flashing techniques for
Exynos7870-based boards. This is done in a separate file in a
sub-directory, which is linked back to the main documentation.
Signed-off-by: Kaustabh Chakraborty <kauschluss@disroot.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Add basic support for the Exynos7870 SoC, this includes device tree
match logic using multiple boards, where devices use a stub dtb in
Samsung's QCDT format. S-BOOT, the previous stage bootloader, places its
cmdline arguments there, which has identifying information.
This is added with support for three devices:
* Samsung Galaxy A2 Core (codename: a2corelte)
* Samsung Galaxy J6 (codename: j6lte)
* Samsung Galaxy J7 Prime (codename: on7xelte)
Add their device trees in the defconfig, and also enable the clock and
pin controller drivers.
Signed-off-by: Kaustabh Chakraborty <kauschluss@disroot.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Add support for a generic platform which intends to support multiple
boards powered by ARMv8 Samsung Exynos SoCs. Some important features
include:
* Fastboot: This is present to provide an open alternative to Samsung's
proprietary Odin protocol. The board file configures certain features
for fastboot, such as a dynamically allocated fastboot buffer, and
standardized (lowercase) partition aliases.
* EFI: Kernel image can be loaded from an EFI partition. This
adopts a standard booting process, which multiple OS distributions
can rely on.
Signed-off-by: Kaustabh Chakraborty <kauschluss@disroot.org>
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
The problem we face currently with Azure jobs is that we're running out
of disk space on the runners as we build. There's not a good way to
split approximately 1500 configurations across 10 jobs and not be close
to or exceeding that limit. Split this in to 29 jobs instead with a goal
of averaging an hour per job. This split gets us close, but there are
still some challenging jobs to try and break up further. The list is
mostly alphabetized but with some intentional changes (catch-all are
last, mx/imx are together, SoC family splits are just grouped together).
The average build time should be close to the same, but outliers can and
will happen.
Signed-off-by: Tom Rini <trini@konsulko.com>
This pull request delivers a broad set of improvements across the
SoCFPGA family, including Agilex5, Cyclone V, SoC64, and common code.
Key updates include refined boot flows, new driver enablement, handoff
tooling enhancements, and several stability fixes.
Highlights:
* Agilex5:
- Enable FAT-based environment storage
- MMC driver restores legacy clkmgr-based clock lookup
- Cleanup of MMC raw mode enablement logic
* Cyclone V:
- SPL FAT boot support and updated bootcmd sequence
- Disable SPL SPI to prevent contention with FAT-based boot
- New board handoff script and BSP generator tooling
- Optimized Makefile support for SoCFPGA handoff workflows* New drivers:
- Cadence xSPI driver with full protocol and command support
- SPL enablement for DW APB GPIO controller
* Networking:
- xgmac MDIO now supports Clause 45 read/write operations
* NAND / SoC64:
- Enable ONFI detection in Denali NAND controller for SoC64 devices
* DTS and board updates:
- Sync common SoCFPGA U-Boot DTS with kernel sources
- Fixes for FPGA2SDRAM configuration and SoCFPGA boot stall behavior
- Vining_FPGA migrated to the modern LED framework
- Device tree relocation no longer forced off for Vining FPGA
* Tooling:
- Introduces a new Python-based Cyclone V BSP generator
covering EMIF, IOCSR, HPS, XML parsing, rendering, and documentation
to simplify board enablement and handoff regeneration workflows
Overall, this series improves boot robustness, enhances xSPI and MDIO
capabilities, modernizes board support, and introduces new tooling to
streamline SoCFPGA handoff generation.
Pipelines test passing
https://source.denx.de/u-boot/custodians/u-boot-socfpga/-/pipelines/28569
The problem we face currently with Azure jobs is that we're running out
of disk space on the runners as we build. There's not a good way to
split approximately 1500 configurations across 10 jobs and not be close
to or exceeding that limit. Split this in to 29 jobs instead with a goal
of averaging an hour per job. This split gets us close, but there are
still some challenging jobs to try and break up further. The list is
mostly alphabetized but with some intentional changes (catch-all are
last, mx/imx are together, SoC family splits are just grouped together).
The average build time should be close to the same, but outliers can and
will happen.
Signed-off-by: Tom Rini <trini@konsulko.com>
- Bump to noble-20251013
- Include tools for sage lab, build TF-A for platforms there.
- Switch to distro provided trace-cmd, add libengine-pkcs11-openssl
- Use mirrors for GNU projects
- Switch to QEMU 10.1.x
Signed-off-by: Tom Rini <trini@konsulko.com>
Quentin Schulz <foss+uboot@0leil.net> says:
I have to add one (1) package to the Dockerfile as a new dependency and
wanted to build the image to verify it works. I wasn't able to because
GNU servers are just not reliable at all.
The git server URL we're using doesn't seem to be a mirrored one, and
switching to mirror URLs seem to make fetching possible again.
Unfortunately, we don't have the option to do that for coreboot's
dependencies, though we can ask coreboot to download the dependencies
through its own mirror, which we do in this series.
Link: https://lore.kernel.org/r/20251127-gnu-mirror-v2-0-c86fa2e8d464@cherry.de
In the future, we'll need proper pkcs11 support so that we can validate
mkimage/binman works well with pkcs11 OpenSSL engine/provider via
softhsm2-util (already installed).
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Now that we have moved to Ubuntu 24.04 the distribution provided
trace-cmd is new enough for our needs. Switch to installing that and
stop building it from source.
Signed-off-by: Tom Rini <trini@konsulko.com>
The current release of QEMU is 10.1.2 and we should be tracking at least
that new currently, to help find and fix emulation problems. This will
make it smaller of a change when we move to 10.1.3 which will re-enable
sifive_unleashed_sdcard testing again.
Signed-off-by: Tom Rini <trini@konsulko.com>
This series adds support for Gitlab pipelines to run our pytest suite on
a limited number of hardware platforms. While better documentation and
some further enhancements will be coming soon, this can be triggered by
passing '-o ci.variable="SAGE_LAB=1"' to git push, or adding
'pushOption = ci.variable="SAGE_LAB=1"' to the .git/config file for the
project. It can also be invoked manually from the pipeline webpage on a
an existing pipeline.
Link: https://lore.kernel.org/r/20251118210015.624758-1-trini@konsulko.com
Add snmp and rsync to the normal docker image. While these tools are
only required for the lab on sage, I think it's a small enough addition
to be worth always including at this point. A higher level TODO I think
may be to see if we can / should split the resulting container up, or
if there's host tooling we can remove at a later step, after building
all the software we require.
Signed-off-by: Tom Rini <trini@konsulko.com>
This adds the Pine64+ platform to the sage lab, for both legacy and lwIP
networking stacks. In order to build this platform we need to copy
certain files that were built in the container to /tmp and then set
BINMAN_INDIRS to /tmp in order to find them when building.
For now, we disable the test_net_pxe_boot_config test on lwIP as it
leads to a crash that needs to be investigated.
Signed-off-by: Tom Rini <trini@konsulko.com>
In order to add pine64_plus to the sage lab we will need to have a build
of TF-A available for it as bl31.bin. Add this to the existing build
loop of TF-A targets.
Signed-off-by: Tom Rini <trini@konsulko.com>
This is the Gitlab side of adding support for the board lab connected to
the "konsulko-sage" runner. On the software side, this lab uses only
upstream labgrid. On the hardware side, each device under test is
connected to its own exporter (typically a Raspberry Pi 4) that must be
turned on (and cleanly turned off) as part of a given test cycle.
Add support for testing on a SolidRun Hummingboard 2 (imx6), Raspberry
Pi 3 and Raspberry Pi 4. In all cases, we enable additional options to
run more tests on the board. As we have some networking tests, we test
both the legacy network stack and lwIP. In the case of Pi platforms, we
test all of 32bit configuration, plain configuration and rpi_arm64, and
again with and without lwIP.
Signed-off-by: Tom Rini <trini@konsulko.com>
tftp_init_load_addr() always returns 0 since commit af45c84871 ("tftp:
rework the logic to validate the load address"), so we don't need to
check if it failed and can remove the error handling.
Also change tftp_init_load_addr() to static void since the return value
is now unused.
Signed-off-by: Leonard Anderweit <l.anderweit@phytec.de>
Reviewed-by: Yannic Moog <y.moog@phytec.de>
Aquantia PHYs are being used w/o SPI flash in some routers recently.
Current firmware loader only attempts to load from FS on top of MMC,
limiting the use on many devices.
Removed the old firmware loader, migrate to generic script based firmware
loader to allow a wider range and runtime override of firmware source.
(e.g., MMC, USB, UBIFS).
Tested on Buffalo WXR18000BE10P with UBIFS.
Signed-off-by: Beiyan Yun <root@infi.wang>
Split `aquantia_upload_firmware` into `aquantia_upload_firmware`
and `aquantia_do_upload_firmware` to prepare for fwloader change.
Signed-off-by: Beiyan Yun <root@infi.wang>
These two PHY drivers have some overlap of supported PHYs. A longer term
effort is needed to both remove duplication and enhance support by
dealing with some issues that downstream vendor drivers address. For
now, make both of these depend on !COMPILE_TEST so that we can enable
"allyesconfig".
Signed-off-by: Tom Rini <trini@konsulko.com>
Bind MDIO controller on Ethernet Controller load. The Airoha AN7581 SoC
have an integrated Switch based on MT7531 (or more saying MT7988).
Attach it to the mdio node in the switch node to support scanning for
MDIO devices on the BUS with DM API.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
- Introduce socfpga_g5_handoff_prepare target in U-Boot
arch/arm/mach-socfpga/config.mk
- Users can convert the handoff via make prepare.
- Detects Altera/Intel SoCFPGA boards from .config
- Combines vendor/board extraction into a single shell call
- Checks for hps_isw_handoff folder and .hiof files
- Uses ls -d instead of find for faster folder detection
- Runs BSP generator script only if files exist
- Non-blocking: continues if handoff folder or files are missing
- HANDOFF_PATH user define allows overriding auto-detected folder
- Minimizes subshells and other slow constructs for faster CI
Signed-off-by: Brian Sune <briansune@gmail.com>
Add the driver for the Airoha EN8811H 2.5 Gigabit PHY. The PHY supports
100/1000/2500 Mbps with auto negotiation only.
The driver uses two firmware files, for which updated versions are added to
linux-firmware already.
Locating the AIROHA FW within the filesystem at the designated partition
and path will trigger its automatic loading and writing to the PHY via MDIO.
If need board specific loading override,
please override the en8811h_read_fw function on board or architecture level.
Based on the Linux upstream AIROHA EN8811H driver code(air_en8811h.c),
I have modified the relevant process to align with the U-Boot boot sequence.
and have validated this on Banana Pi BPI-R3 Mini.
Signed-off-by: Lucien.Jheng <lucienzx159@gmail.com>
Reviewed-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
bcm54210e_config() configures the RGMII delays and then calls
bcm5461_config(). But the latter will do a PHY soft reset and thus
resets the delay settings again. Call bcm5461_config() first to fix it.
Fixes: cba79a1b2e ("net: phy: broadcom: add support for BCM54210E")
Signed-off-by: Michael Walle <mwalle@kernel.org>
Reviewed-by: Rafał Miłecki <rafal@milecki.pl>
According to the TI DP83869HM datasheet Revision D (June 2025), section
7.6.1.41 STRAP_STS Register, the STRAP_OPMODE bitmask is bit [11:9].
Fix this.
In case the PHY is auto-detected via PHY ID registers, or not described
in DT, or, in case the PHY is described in DT but the optional DT property
"ti,op-mode" is not present, then the driver reads out the PHY functional
mode (RGMII, SGMII, ...) from hardware straps.
Currently, all upstream users of this PHY specify both DT compatible string
"ethernet-phy-id2000.a0f1" and ti,op-mode = <DP83869_RGMII_COPPER_ETHERNET>
property, therefore it seems no upstream users are affected by this bug.
The driver currently interprets bits [2:0] of STRAP_STS register as PHY
functional mode. Those bits are controlled by ANEG_DIS, ANEGSEL_0 straps
and an always-zero reserved bit. Systems that use RGMII-to-Copper functional
mode are unlikely to disable auto-negotiation via ANEG_DIS strap, or change
auto-negotiation behavior via ANEGSEL_0 strap. Therefore, even with this bug
in place, the STRAP_STS register content is likely going to be interpreted
by the driver as RGMII-to-Copper mode.
However, for a system with PHY functional mode strapping set to other mode
than RGMII-to-Copper, the driver is likely to misinterpret the strapping
as RGMII-to-Copper and misconfigure the PHY.
For example, on a system with SGMII-to-Copper strapping, the STRAP_STS
register reads as 0x0c20, but the PHY ends up being configured for
incompatible RGMII-to-Copper mode.
Fixes: f3e22eea81 ("net: phy: add TI DP83869HM ethernet driver")
Signed-off-by: Thanh Quan <thanh.quan.xn@renesas.com>
Signed-off-by: Hai Pham <hai.pham.ud@renesas.com>
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org> # Use FIELD_GET()
Disable support for using SPI in SPL to solve Cyclone V storage issue as
the OCRAM is only 64kb. The SPI configurations are only
required during uboot proper only.
Signed-off-by: Alif Zakuan Yuslaimi <alif.zakuan.yuslaimi@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
Enable support for FAT filesystem with SPL for CycloneV instead of raw
mode.
Recent changes breaks CycloneV MMC boot from raw mode, and so we are taking
this opportunity to migrate MMC boot mode to FAT as smaller OCRAM size is
required, as well as aligning MMC boot flow with our other devices.
Fixes: 2a00d73d08 ("spl: mmc: Try to clean up raw-mode options")
Signed-off-by: Alif Zakuan Yuslaimi <alif.zakuan.yuslaimi@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
We no longer use raw mode to boot from MMC for our devices in favor
of FAT filesystem.
Maintaining this config for legacy gen5 devices as to not risk breaking
any configurations still utilizing raw mode.
Signed-off-by: Alif Zakuan Yuslaimi <alif.zakuan.yuslaimi@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
Restore legacy implementation of retrieving clkmgr base address from
mach-socfpga/misc.c driver for our legacy devices.
Excluding Agilex7/7M from this implementation as these devices' clock
driver is already following clock driver model and is supporting
enable/disable APIs.
The legacy devices' clock driver will have to be refactored to support
driver model which enables us to support enable/disable APIs for these
devices.
Fixes: ab27182cac ("mmc: socfpga_dw_mmc: Enable/disable SDMMC clock via API")
Signed-off-by: Alif Zakuan Yuslaimi <alif.zakuan.yuslaimi@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
mkeficapsule tool will be built by default if EFI_LOADER is set due to
commit b7a625b1ce ("tools: Build mkeficapsule tool by default if
EFI_LOADER is set").
This will cause compilation error on all our SoCFPGA devices, hence we will
be disabling this config as we do not utilize this tool.
Signed-off-by: Alif Zakuan Yuslaimi <alif.zakuan.yuslaimi@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
This migrates from the legacy LED API to use the modern LED framework
which makes use of the FDT.
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de>
Acked-by: Tien Fong Chee <tien.fong.chee@altera.com>
Remove setting of fdt_high to ~0, which disables device tree relocation,
from the default environment. Doing so prevents U-Boot from correcting
problems such as having an unaligned device tree and leads to various
failure modes in the OS.
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Marek Vasut <marek.vasut@mailbox.org>
Since U-Boot 2025.07 pure SD Card
boot no longer works. Now Altera released 2025.07
shows the different on the u-boot files.
After testing, the major root case is
get_managers_addr. And this patch fix the
SD boot stall via pulling from offical.
Signed-off-by: Brian Sune <briansune@gmail.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
The dtsi for socfpga common should
turn on L2 and memory and no reason not
to do so
Signed-off-by: Brian Sune <briansune@gmail.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
Enable ONFI parameter page detection for SoCFPGA SoC64 devices by
selecting SYS_NAND_ONFI_DETECTION in the NAND_DENALI Kconfig entry.
This allows SoCFPGA SoC64 platforms using the Denali NAND controller
to automatically detect NAND parameters via the ONFI interface instead
of relying on hardcoded configuration values.
The selection is limited to TARGET_SOCFPGA_SOC64 to avoid affecting
non-SoC64 platforms that use legacy NAND handling.
Signed-off-by: Dinesh Maniyam <dinesh.maniyam@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
Since turning from old build flow.
New Altera SoCFPGA requires converting handsoff
conversion via the python script. This is from
official provided, and now sync to U-Boot with
better location at tools/cv_xxxx. Meantime,
requirement.txt is also provided to further
explain the libraries require for these scripts.
Signed-off-by: Brian Sune <briansune@gmail.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
This patch ports the Cadence xSPI controller driver from
the Linux kernel. The controller supports three operating modes:
1. ACMD (Auto Command) mode
- Includes PIO and CDMA submodes.
- CDMA mode uses linked descriptors for high-performance,
low-overhead operation.
- PIO mode is suitable for simple, single-command transactions.
2. STIG (Software Triggered Instruction Generator) mode
- Issues low-level 128-bit instructions to memory.
- Uses the Slave DMA interface for data transfers.
3. Direct mode
- Enables direct data access through the slave interface
without commands.
Currently, only the STIG work mode is enabled. Additional modes will be
supported in future updates.
At the same time, also enabling the kconfig option for xSPI driver.
This driver has been ported and functionally verified on the Intel Simics
platform. It is intended for evaluation and experimental use at this stage.
Link: https://lore.kernel.org/all/1632038734-23999-1-git-send-email-pthombar@cadence.com/
Signed-off-by: Boon Khai Ng <boon.khai.ng@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
Enable storing the U-Boot environment in a FAT filesystem for Agilex5.
This allows the board to read and write environment variables from the
first partition of the first device formatted with FAT, in addition to the
existing UBI-based environment configuration.
Signed-off-by: Alif Zakuan Yuslaimi <alif.zakuan.yuslaimi@altera.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
After testing, w/o proper setup
the FPGA2SDRAM bridge will not work and stall.
Pulling from official fix and w/o this initialization,
both 2025.07 and 2025.10 also suffer stall on U-Boot
and distro. Any FPGA to HPS-SDRAM action will immediate
stall the CPU. As such, this patch fix the issue.
Signed-off-by: Brian Sune <briansune@gmail.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@altera.com>
Add SPL_DWAPB_GPIO configuration option to enable the Designware APB
GPIO driver in SPL builds.
Changes:
- Add SPL_DWAPB_GPIO Kconfig option with SPL_DM_GPIO dependency
- Update Makefile to use CONFIG_$(PHASE_)DWAPB_GPIO pattern for
conditional compilation in both SPL and main U-Boot builds
Signed-off-by: Tanmay Kathpalia <tanmay.kathpalia@altera.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
coreboot buildgcc script downloads GNU dependencies from GNU FTP server.
For some reason, this is also as unreliable as their git main server.
There's no option to use a GNU mirror (and I'm not even sure if there's
one for FTP), so we simply pass --mirror to the buildgcc script via the
BUILDGCC_OPTIONS variable so that it makes use of coreboot's mirror.
Hopefully, this proves more reliable than GNU original FTP server.
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de>
grub bootstrap script downloads gnulib from a non-mirror URL and thus is
unreliable.
One can specify the URL to fetch from with GNULIB_URL environment
variable, so let's make this variable point at a mirror URL.
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de>
GNU announced they activated mirrors for git servers[1] in May this
year.
The main git server keeps being very unreliable and switching to those
mirrors seems to improve reliability (albeit somewhat slow).
Yes, the URL in this patch has nothing in common with the URL in the
linked mail, it was extracted from the Clone section in the mirrored
cgit page[2].
While at it, switch to the HTTPS clone which is "more secure" than git
protocol.
[1] https://lists.gnu.org/archive/html/savannah-users/2025-05/msg00002.html
[2] https://cgit.git.savannah.gnu.org/cgit/grub.git/
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Quentin Schulz <quentin.schulz@cherry.de>
CI: https://source.denx.de/u-boot/custodians/u-boot-imx/-/pipelines/28551
- Stop disabling device tree relocation on i.MX boards.
- Add support for imx6ulz_smm_m2d.
- Add support for Kontron i.MX93 OSM-S SoM and BL carrier board
- Fix probe of i.MX FEC MDIO bus for DM_MDIO.
- Cleanup for tqma6 boards.
- Initialize the ELE RNG context on imx95_evk
- Disable EFI partition, MP, sysboot, bind commands on tbs2910.
fdtoverlay_addr_r is the correct variable name according to standardboot
documentation and already used for phycore_imx8mp.
Signed-off-by: Benjamin Hahn <B.Hahn@phytec.de>
We want to change the bootdevice dynamically, so that when booting
U-Boot from sdcard, kernel is also preferably booted from sdcard by
default. If the user decides to set their own bootorder, this should not
be overwritten. This was already implemented for imx8mp-libra-fpsc
board, but as we set the default value for boot_targets in devicetree
now, the env_get_default will return NULL. Also env_get might return
NULL when boot_targets is not set. A string compare with NULL is unsafe
and should be avoided. To fix this we only change the env value if the
variable was not set before (NULL), as this is the new default value.
In any other case the user has changed the value so it will not be
overwritten.
Also let standardboot set bootcmd automatically. For this
CONFIG_USE_BOOTCOMMAND needs to be activated.
Signed-off-by: Benjamin Hahn <B.Hahn@phytec.de>
