Siddharth Vadapalli <s-vadapalli@ti.com> says:

Hello,

This series re-enables MUX_MMIO at U-Boot proper on J721E and J7200
SoCs. The commits which introduced support for Hyperflash boot on both
of these SoCs disabled MUX_MMIO functionality at U-Boot proper, thereby
introducing a regression.

Series is based on commit
2b1c8d3b cmd: Fix Kconfig coding style
of the master branch of U-Boot.

Link: https://lore.kernel.org/r/20250201091809.1894892-1-s-vadapalli@ti.com

Commit under Fixes introduced support for Hyperflash boot but also disabled
MUX_MMIO support at U-Boot proper. MUX_MMIO is required at U-Boot proper
to setup the SERDES Lane Mapping as described by the device-tree. On J7200
SoC, PCIe functionality is broken since commit under fixes.

Fix this regression by re-enabling MUX_MMIO.

Fixes: 038f6fae ("configs: j7200_evm_*_defconfig: Add configs for Hyperflash boot")
Signed-off-by: Siddharth Vadapalli <s-vadapalli@ti.com>
Reviewed-by: Anurag Dutta <a-dutta@ti.com>

Commit under Fixes introduced support for Hyperflash boot but also disabled
MUX_MMIO support at U-Boot proper. MUX_MMIO is required at U-Boot proper
to setup the SERDES Lane Mapping as described by the device-tree.

Fix this regression by re-enabling MUX_MMIO.

Fixes: fd7fcd40 ("configs: j721e_evm_*_defconfig: Add configs for Hyperflash boot")
Signed-off-by: Siddharth Vadapalli <s-vadapalli@ti.com>

emul_rtc_probe() is not called from anywhere else that's why make it
static. Issue is reported by build with W=1.

Signed-off-by: Michal Simek <michal.simek@amd.com>

Support phy io micro voltage setting for NPCM8XX rgmii
interface.

Signed-off-by: Michael Chang <zhang971090220@gmail.com>

This patch enables switch auto-detction for mt7981 as some new mt7981
boards will use AN8855 ethernet switch.

Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>

Enable configs required for detecting and fixing up for different RAM variants.
Also resync after savedefconfig.

Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: Daniel Schultz <d.schultz@phytec.de>

Detect RAM size via EEPROM and adjust DDR size and banks accordingly.
Include necessary fixups to handle ECC-enabled configurations.

Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: Daniel Schultz <d.schultz@phytec.de>

Add bootph-all properties to I2C0 nodes to ensure the bus and EEPROM
are accessible across all stages. This enables reading the SoM
configuration at any point during the boot process.

Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: Daniel Schultz <d.schultz@phytec.de>
Reviewed-by: Dhruva Gole <d-gole@ti.com>

USB tests on ext partitions can fail with the following output

test/py/tests/test_usb.py:245: in test_usb_part
    'fstype usb %d:%d' % i, part_id
E   TypeError: not enough arguments for format string

So add brackets around the format string arguments to prevent the
error.

Fixes: a7309479 ("test/py: usb: Distinguish b/w ext2/ext4 partitions")
Signed-off-by: Andrew Goodbody <andrew.goodbody@linaro.org>
Reviewed-by: Marek Vasut <marex@denx.de>
Reviewed-by: Love Kumar <love.kumar@amd.com>

The -V version parameter is missing in the optional list of parameters
in help text. Add it.

Signed-off-by: Marek Vasut <marex@denx.de>

The change from pinctrl to pio was missing in mt7986-u-boot.dtsi and will
cause build failure. Now fix it.

Fixes: f1775996 (mediatek: mt7986: move gpio-controller up and rename pinctrl to pio)
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>

This patch add support to read register base address by name if
provided.

Also devfdt_get_addr_* is changed to dev_read_addr_* to support DT
live tree.

Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>

The reg-names in mt7988.dtsi have no _base suffix. Remove the suffix
will also make it match upstream linux format.

Fixes: 8c2cb748 (pinctrl: mediatek: mt7988: rename reg-names to upstream linux format)
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>

Without setting the shell flag to exit immediately when a command exists
with a non-zero status we can have the situation where the htmldocs
target fails with an error but the job will succeed due to infodocs
passing and being the last build target.

Signed-off-by: Tom Rini <trini@konsulko.com>

- e850-96 platform updates

Heinrich Schuchardt <heinrich.schuchardt@canonical.com> says:

The setexpr.s command allows to concatenate two strings.

According to the description in doc/usage/cmd/setexpr.rst the parameters
value1 and value2 can be either direct values or pointers to a
memory location holding the values.

Unfortunately `setexpr.s <value1> + <value2>` fails if any of the values
is a direct value. $? is set to false.

* Add support for direct values in setexpr.s.
* Correct the unit test for "setexpr.s fred 0".
* Add a new unit test for "setexpr.s fred '1' + '3'" giving '13'.
* Remove invalid memory leak tests

Link: https://lore.kernel.org/r/20250203151029.60265-1-heinrich.schuchardt@canonical.com

The setexpr.s command allows to concatenate two strings.

According to the description in doc/usage/cmd/setexpr.rst the parameters
value1 and value2 can be either direct values or pointers to a
memory location holding the values.

Unfortunately `setexpr.s <value1> + <value2>` fails if any of the values
is a direct value. $? is set to false.

* Add support for direct values in setexpr.s.
* Correct the unit test for "setexpr.s fred 0".
* Add a new unit test for "setexpr.s fred '1' + '3'" giving '13'.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

env_set() frees the previous value after allocating the new value.
As the free() may merge memory chunks the available memory is not
expected to stay constant.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

env_set() frees the previous value after allocating the new value.
As the free() may merge memory chunks the available memory is not
expected to stay constant.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

Assign variable buf in the sub-test where it is used.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

Heiko Schocher <hs@denx.de> says:

In linux we have the option to create the name of a led
optionally through the following properties:

- function
- color
- function-enumerator

This series adds support for parsing this properties if there
is no label property.

Link: https://lore.kernel.org/r/20250128135246.74838-1-hs@denx.de
[trini: Document name parameter in led.h]

in linux we have the option to create the name of a led
optionally through the following properties:

- function
- color
- function-enumerator

This patch adds support for parsing this properties if there
is no label property.

The led name is created in led_post_bind() and we need some
storage place for it. Currently this patch prevents to use
malloc() instead it stores the name in new member :

        char name[LED_MAX_NAME_SIZE];

of struct led_uc_plat. While at it append led tests for the
new feature.

Signed-off-by: Heiko Schocher <hs@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>

remove file doc/device-tree-bindings/leds/common.txt
as we have this now already in
dts/upstream/include/dt-bindings/leds/common.h
which is imported from linux.

Signed-off-by: Heiko Schocher <hs@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>

AMD/Xilinx changes for v2025.04-rc2

fpga:
- Cleanup help
- Show xilinx only options on Xilinx devices

ospi-versal:
- Fix alignment issue
- Fix cadence_qspi_flash_reset() prototype

zynqmp:
- Define usb_pgood_delay
- Fix bootseq number

versal:
- Fix mini_ospi configuration

versal2:
- Enable OPTEE

xilinx:
- Enable some flashes
- Clean up SYS_MALLOC_F_LEN Kconfig
- Some binman fixes
- DT updates
- Enable mkfwumdata compilation
- Enable meminfo command
- Switch to LWIP and enable HTTPS

AM62x BOOTM_LEN is too small to contain OS images. Removing this sets the
size to a default 0x4000000 (for all arm64 devices).

It is unnecessary to specifically call the default size for the other
configs.

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
Signed-off-by: Aashvij Shenai <a-shenai@ti.com>

Kconfig should not display the STATIC_MACH_TYPE option when
SUPPORT_PASSING_ATAGS is not selected. For example, when using most
platforms with menuconfig, it shows the STATIC_MACH_TYPE option causing
confusion.  As we know, this should not occur when SUPPORT_PASSING_ATAGS
is not selected

[trini: Reword commit message]
Signed-off-by: Liya Huang <1425075683@qq.com>

ARMv7-M architecture uses NVIC instead of GIC.
NVIC is an interrupt controller specially designed for
ARM Cortex-M series processors.

Signed-off-by: Liya Huang <1425075683@qq.com>

In U-Boot, most of mmc releated commands uses decimal value in arguments, like
"mmc dev ${bootseq}" or "bootargs=root=/dev/mmcblk${bootseq}p2". In order to
improve compatibilities, export "bootseq" number to environment variable in
decimal format instead of hex.

Signed-off-by: Benjamin Szőke <egyszeregy@freemail.hu>
Link: https://lore.kernel.org/r/20250204205617.1238-1-egyszeregy@freemail.hu


Signed-off-by: Michal Simek <michal.simek@amd.com>

U-Boot can run out of EL3, NS-EL2 and NS-EL1. Currently default
configuration is NS-EL2 with TF-A but when TF-A is not passed and
configured images can still boot just fine. That's why support this
configuration and describe it via binman.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/9b3dad80138e97005df3d033b4611c9d7e05a177.1738659214.git.michal.simek@amd.com

CONFIG_XILINX_OF_BOARD_DTB_ADDR holds DTB address which U-Boot is checking.
Currently address in binman match default value but macro can be used
directly.
Also sync node name (s/hash-1/hash/) and sync location to have the same
order load/hash/image.

All binman DTSes are compiled that's why also guard
CONFIG_XILINX_OF_BOARD_DTB_ADDR which depends on OF_BOARD || OF_SEPARATE
which is a problem for mini configurations which are using OF_EMBED.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/a22c9671b965b222bfd419f5bfaee012929f3d88.1738659214.git.michal.simek@amd.com

There is incorrect loading address listed for TEE.
CONFIG_BL32_LOAD_ADDR should be used.
Also there is missing hash for this entry which is present for other nodes.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/4e7e3a7110acc050ea7c06ac661e5b5be46e8602.1738659214.git.michal.simek@amd.com

do_fpga_loads() is not called from anywhere else that's why make it static.

Reviewed-by: Alexander Dahl <ada@thorsis.com>
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/cbe93e4eb33752aaddc943b15fd04731d4f75b68.1738574893.git.michal.simek@amd.com

The mini u-boot is hanging because of an initial stack
pointer address is used at half of the memory, when mini
u-boot is called reloc_fdt() function and doing memcpy()
for the fdt before relocation, and there is no sufficient
memory for the stack pointer. To fix, set an initial stack
pointer address to near the top of memory. The new stack
pointer address is assigned before calling the relocate
of u-boot based on the stack relocate calculation and
this new stack pointer is used while executing u-boot
from the relocated memory.

Signed-off-by: Padmarao Begari <padmarao.begari@amd.com>
Link: https://lore.kernel.org/r/20250203090409.844403-1-padmarao.begari@amd.com


Signed-off-by: Michal Simek <michal.simek@amd.com>

Enable lwip and https on our platforms to be able to use it in a boot.

Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/cb05adaf0758c2c4f1361f8665169897493638e7.1738164681.git.michal.simek@amd.com

Enable meminfo command to be able to see where things are mapped.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/b7f9f5e18e16a945277d4b8cfde5a7f057e77676.1737976295.git.michal.simek@amd.com

Build mkfwumdata tool by default for building ab mdata structure.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/6d82c9a2db30e12ea17fa125c68a555d7f070851.1737723845.git.michal.simek@amd.com

Adding new symbol for the fpga loadb command which is exclusive to
Xilinx. Default value is y for backward compatibility.

Clarify the type of file used for fpga programming commands

Signed-off-by: Ibai Erkiaga <ibai.erkiaga-elorza@amd.com>
Link: https://lore.kernel.org/r/20250121130138.1999916-6-ibai.erkiaga-elorza@amd.com


Signed-off-by: Michal Simek <michal.simek@amd.com>

Ensure all Xilinx exclusive fpga commands have a KConfig symbol and
dependency to FPGA_XILINX listed. Remove (Xilinx only) text from the
help command.

Signed-off-by: Ibai Erkiaga <ibai.erkiaga-elorza@amd.com>
Link: https://lore.kernel.org/r/20250121130138.1999916-5-ibai.erkiaga-elorza@amd.com


Signed-off-by: Michal Simek <michal.simek@amd.com>