]> git.dujemihanovic.xyz Git - u-boot.git/log
u-boot.git
6 months agoarm64: zynqmp: Setup multiboot register to 0
Michal Simek [Mon, 3 Jun 2024 13:09:01 +0000 (15:09 +0200)]
arm64: zynqmp: Setup multiboot register to 0

On Kria when board starts from Image A or Image B partition multiboot
register is already setup to that location. When reset command is called
board is issuing soft reset which start SW at already used location (offset
of multiboot * 32k).
But board should continue to run from multiboot offset 0 (start of QSPI)
and call early bootloader every reboot that's why clear multiboot register
to 0 by default to go that route all the time.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/edaf714a778bdd7447533a77b3455e4fd623f9da.1717420131.git.michal.simek@amd.com
6 months agospi: versal2: Enable spi drivers for Versal Gen 2
Michal Simek [Wed, 29 May 2024 14:48:01 +0000 (16:48 +0200)]
spi: versal2: Enable spi drivers for Versal Gen 2

Enable and update OSPI/QSPI/GQSPI drivers to support Versal Gen 2 SoCs.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/691782470f56f7d49a3204f6757296f2752d4156.1716994063.git.michal.simek@amd.com
6 months agommc: versal2: Update zynq_sdhci driver to support AMD Versal Gen 2
Michal Simek [Wed, 29 May 2024 14:48:00 +0000 (16:48 +0200)]
mmc: versal2: Update zynq_sdhci driver to support AMD Versal Gen 2

Enable tap delay programming for new SoC and also enable it via defconfig.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/f07daded9704cbc393657b65a28933c34a8cec25.1716994063.git.michal.simek@amd.com
6 months agosoc: versal2: Add SoC driver for AMD Versal Gen 2
Michal Simek [Wed, 29 May 2024 14:47:59 +0000 (16:47 +0200)]
soc: versal2: Add SoC driver for AMD Versal Gen 2

Communication is happening via firmware interface (SMC) or via direct
register reading if firmware driver is not available.

Also enable it via defconfig.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/22cf9c765e47ab03dbf2b8363e6626e809113432.1716994063.git.michal.simek@amd.com
6 months agoarm64: versal2: Add support for AMD Versal Gen 2
Michal Simek [Wed, 29 May 2024 14:47:58 +0000 (16:47 +0200)]
arm64: versal2: Add support for AMD Versal Gen 2

Add support for AMD Versal Gen 2. SoC is based on Cortex-a78ae 4 cluster/2
cpu core each. A lot of IPs are shared with previous families. There are
couple of new IP blocks where the most interesting from user point of view
is UFS.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/bc2b70831ce1031bd0fac32357bff84936e1310f.1716994063.git.michal.simek@amd.com
6 months agoarm64: zynqmp: Update rproc node
Michal Simek [Thu, 30 May 2024 10:39:23 +0000 (12:39 +0200)]
arm64: zynqmp: Update rproc node

remoteproc node should be updated to be aligned with the latest dt-schema.

Reviewed-by: Tanmay Shah <tanmay.shah@amd.com>
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/d8247a46f486a612f85767de9b832ad33fa980fe.1717065556.git.michal.simek@amd.com
6 months agoxilinx: versal: Do not prioritize boot device if driver is not enabled
Venkatesh Yadav Abbarapu [Fri, 10 May 2024 06:22:38 +0000 (08:22 +0200)]
xilinx: versal: Do not prioritize boot device if driver is not enabled

SOC can boot out of the device which is not accessible from APU and running
this is detected as a warning, as the device is not accessible.For example
getting below warning when the boot mode is OSPI and OSPI is not enabled in
device tree.
Invalid bus 0 (err=-19)
Failed to initialize SPI flash at 0:0 (error -19)

Ignoring the prioritization of the boot device which driver is not enabled
and continue with the default boot_targets. Recommendation is to use custom
boot_targets via environment file as is done for example for Kria via
zynqmp_kria.env file.

Signed-off-by: Venkatesh Yadav Abbarapu <venkatesh.abbarapu@amd.com>
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/8b7cca5c7b84cb4854104e0c48f8aa63c4ec5ace.1715322156.git.michal.simek@amd.com
6 months agomtd: spi-nor: Add SPI_NOR_OCTAL_READ flag for mx66uw2g345gx0 flash part
Prasad Kummari [Wed, 8 May 2024 05:27:50 +0000 (10:57 +0530)]
mtd: spi-nor: Add SPI_NOR_OCTAL_READ flag for mx66uw2g345gx0 flash part

Added SPI_NOR_OCTAL_READ flag for Macronix mx66uw2g345gx0 2Gb(256MB)
NOR Flash memory. Initial testing was conducted on the Versal NET board
using SDR mode, which included basic erase, write, and read-back
operations.

Signed-off-by: Prasad Kummari <prasad.kummari@amd.com>
Link: https://lore.kernel.org/r/20240508052749.214286-1-prasad.kummari@amd.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
6 months agoarm64: zynq(mp): Rename spl_board_init() to spl_soc_init()
Lukas Funke [Wed, 27 Mar 2024 12:11:53 +0000 (13:11 +0100)]
arm64: zynq(mp): Rename spl_board_init() to spl_soc_init()

Rename spl_board_init() to spl_soc_init(). SoC specific
implementation should be separated from board specific implementation
in order to be extended by board developers.

Signed-off-by: Lukas Funke <lukas.funke@weidmueller.com>
Acked-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/20240327121153.2455126-3-lukas.funke-oss@weidmueller.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
6 months agospl: Introduce SoC specific init function
Lukas Funke [Wed, 27 Mar 2024 12:11:52 +0000 (13:11 +0100)]
spl: Introduce SoC specific init function

Some architectures use spl_board_init() in their SoC specific
implementation. Board developers should be able to add board specific
implementation via spl_board_init(). Hence, introduce a spl_soc_init()
method which is called right before spl_board_init() for SoC
specific implementation.

Signed-off-by: Lukas Funke <lukas.funke@weidmueller.com>
Reviewed-by: Devarsh Thakkar <devarsht@ti.com>
Link: https://lore.kernel.org/r/20240327121153.2455126-2-lukas.funke-oss@weidmueller.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
6 months agoxilinx: zynqmp: Allow multiboot environment write even in saved environment
Kory Maincent [Wed, 29 May 2024 10:01:06 +0000 (12:01 +0200)]
xilinx: zynqmp: Allow multiboot environment write even in saved environment

Once the environment was saved, the current multiboot image information
became unreachable. When dealing with firmware updates, this information
is necessary alongside the saved environment to know the booted image.

Move the multiboot environment set operation before the saved environment
check to ensure this information is always available.

Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Link: https://lore.kernel.org/r/20240529100107.137159-1-kory.maincent@bootlin.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
6 months agosdhci: zynq: Fix tap delay for SD on Versal NET
Simek, Michal [Thu, 18 Apr 2024 08:06:13 +0000 (20:06 -1200)]
sdhci: zynq: Fix tap delay for SD on Versal NET

I can't see any way how tap delays are setup on Versal NET platform because
xlnx,versal-8.9a compatible string is also used there but driver is not
letting to setup tap delays. Not sure if versal_iclk_phases[] is also valid
for Versal NET but the patch is made to investigate it.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/e535cfc1a59b5146a5c9a3ab389dc770de80440c.1713427490.git.michal.simek@amd.com
6 months agoboard: zynqmp: Move zynqmp commands from board/ to arch/
Charlie Johnston [Wed, 10 Apr 2024 19:50:08 +0000 (12:50 -0700)]
board: zynqmp: Move zynqmp commands from board/ to arch/

The zynqmp cmds.c is currently tied to the board but the commands
contained within are more closely tied to the architecture. To
allow usage of those commands when the architecture is ZynqMP but
the board is not, this change moves the cmds into the arch/ tree.

The source file is renamed to zynqmp.c to reflect the command name
as well.

Signed-off-by: Charlie Johnston <charlie.johnston@loftorbital.com>
Link: https://lore.kernel.org/r/20240410195008.405061-2-charlie.johnston@loftorbital.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
6 months agoxilinx: Enable NVMEM framework for all platforms
Michal Simek [Thu, 11 Apr 2024 06:04:16 +0000 (08:04 +0200)]
xilinx: Enable NVMEM framework for all platforms

Boards which have for example MAC address in eeprom but not in Xilinx
format (legacy or FRU) could reference it via nvmem cells.
For example:

&gem0 {
nvmem-cells = <&mac>;
nvmem-cell-names = "mac-address";
};

&eeprom {
#address-cells = <1>;
#size-cells = <1>;
mac: mac-address@f0 {
reg = <0xf0 6>;
};
};

For getting it work above DT changes are required but also CONFIG_NVMEM
should be enabled. That's why enable it by default in generic defconfigs
to be able to use it directly by changing DT only.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/9c8ee7a4c7a16367438a92a4c9581bac9d968f84.1712815454.git.michal.simek@amd.com
6 months agoxilinx: zynqmp: Clean up xilinx_zynqmp.h
Michal Simek [Mon, 15 Apr 2024 07:55:33 +0000 (09:55 +0200)]
xilinx: zynqmp: Clean up xilinx_zynqmp.h

Options are moving to Kconfig by running sed and comments are staying in
that's why do clean up and remove useless comments.

Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/fa117ac482591d3d5957af54fe99e6acc89972e3.1713167731.git.michal.simek@amd.com
6 months agotest/py: net_boot: Add test cases for net boot
Love Kumar [Wed, 5 Jun 2024 09:49:35 +0000 (15:19 +0530)]
test/py: net_boot: Add test cases for net boot

Add tests for booting image using tftpboot/pxe boot commands, tftpboot
boot case loads the FIT image into DDR and boots using bootm command
whereas pxe boot cases downloads the pxe configuration file from the
TFTP server and interprets it to boot the images mentioned in the pxe
configurations file.
This test relies on boardenv_* containing configuration values including
the parameter 'pattern'. tftpboot/pxe boot cases boots the Linux till the
boot log pattern value is matched. For example, if the parameter
'pattern' is defined as 'login:', it will boot till login prompt.

Signed-off-by: Love Kumar <love.kumar@amd.com>
Tested-by: Tom Rini <trini@konsulko.com>
6 months agotest/py: Add support to enable check for bad pattern
Love Kumar [Wed, 22 May 2024 13:15:13 +0000 (18:45 +0530)]
test/py: Add support to enable check for bad pattern

Executing a u-boot command may raise an error or extra bad pattern,
beyond the default bad patterns. Providing a way to enable the console
output error check in test.

For example, description for OS boot test:
import re
check_type = 'kernel_boot_error'
check_pattern = re.compile('ERROR -2: can't get kernel image!')
with u_boot_console.enable_check(check_type, check_pattern):
    u_boot_console.run_command('<boot command>')

Signed-off-by: Love Kumar <love.kumar@amd.com>
6 months agocmd: bcb: Fix bcb compilation when CONFIG_CMD_BCB=n
Mattijs Korpershoek [Mon, 3 Jun 2024 09:04:58 +0000 (11:04 +0200)]
cmd: bcb: Fix bcb compilation when CONFIG_CMD_BCB=n

commit dfeb4f0d7935 ("cmd: bcb: extend BCB C API to allow read/write the fields")
introduced the bcb_get() function.

When CONFIG_CMD_BCB=n, that function is stubbed.
The stubbed function has a wrong prototype: value_size arg is missing.

Add the missing argument to fix build when CONFIG_CMD_BCB=n.

Fixes: dfeb4f0d7935 ("cmd: bcb: extend BCB C API to allow read/write the fields")
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Reviewed-by: Guillaume La Roque <glaroque@baylibre.com>
Reviewed-by: Dmitrii Merkurev <dimorinny@google.com>
Reviewed-by: Julien Masson <jmasson@baylibre.com>
6 months agoimxtract: add support for zstd-compressed images
Dmitry Gerasimov [Sun, 2 Jun 2024 15:25:52 +0000 (17:25 +0200)]
imxtract: add support for zstd-compressed images

Allow extraction of zstd-compressed images from FIT using imxtract
command. This is especially useful when one has to load an image via
some interface (e.g. SPI) rather that just to the memory.

Signed-off-by: Dmitry Gerasimov <di.gerasimov@gmail.com>
6 months agobcmgenet: fix Rx buffer corruption caused by lack of cache flush
Yasuharu Shibata [Sun, 2 Jun 2024 08:24:03 +0000 (17:24 +0900)]
bcmgenet: fix Rx buffer corruption caused by lack of cache flush

When bcmgenet complete to write Rx buffer with the DMA,
some U-Boot commands write data to the buffer directly.
Those write data will become dirty in CPU cache.
After this driver calls free_pkt to the buffer,
the buffer is assigned as the future Rx buffer.
At some point, if bcmgenet writes to a buffer with DMA
and CPU cache flushes dirty data to the buffer,
the buffer is corrupted.

This patch calls flush_dcache_range in free_pkt
to immediately flush the data written by U-Boot command
and prevent data corruption.

This issue can be reproduced using wget on Raspberry Pi4.
If wget receives data larger than
RX_BUF_LENGTH * RX_DESCS = 2048 * 256 bytes,
it will timeout due to data corruption.
In addition, if LOG_DEBUG is enabled in net/tcp.c,
the following error log is output.

TCP RX TCP xSum Error

Signed-off-by: Yasuharu Shibata <yasuharu.shibata@gmail.com>
6 months agoconfigs: j784s4: Enable OSPI NOR boot
Jonathan Humphreys [Fri, 31 May 2024 22:32:41 +0000 (17:32 -0500)]
configs: j784s4: Enable OSPI NOR boot

Set necessary configs to enable the j784s4 device to boot from OSPI NOR
flash.

Signed-off-by: Jonathan Humphreys <j-humphreys@ti.com>
6 months agospl: Kconfig: ARCH_K3: Set default SPL_STACK_R_MALLOC_SIMPLE_LEN for R5 build
Vignesh Raghavendra [Fri, 31 May 2024 14:46:36 +0000 (20:16 +0530)]
spl: Kconfig: ARCH_K3: Set default SPL_STACK_R_MALLOC_SIMPLE_LEN for R5 build

All ARCH_K3 platforms need about of 2MB of malloc space post
reallocation. Since, this space is allocated from SDRAM, provide a
generous 2MB space by default.

Platforms requiring more than 2MB can override in defconfig as needed.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
6 months agoarm: dts: k3-*-binman: Make default DM file optional
Neha Malcom Francis [Wed, 29 May 2024 07:48:49 +0000 (13:18 +0530)]
arm: dts: k3-*-binman: Make default DM file optional

The default DM firmware path is non-optional as of now. Make it
optional so that users that choose to provide DM via TI_DM argument
instead of BINMAN_INDIRS can do so without build errors.

Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
Reviewed-by: Kamlesh Gurudasani <kamlesh@ti.com>
6 months agoenv: ti: k3_dfu: Drup mmcpart for rootfs
Neha Malcom Francis [Wed, 29 May 2024 06:05:59 +0000 (11:35 +0530)]
env: ti: k3_dfu: Drup mmcpart for rootfs

According to [0], raw access to mmc should not have mmcpart in the
entry. This was fixed in k3_dfu_combined.env but k3_dfu.env had been
overlooked.

[0] doc/usage/dfu.rst

Fixes: 53b406369e9d ("DFU: Check the number of arguments and argument string strictly")
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
6 months agoarm: dts: k3-j721s2-r5: Change GTC clock parent
Neha Malcom Francis [Tue, 28 May 2024 09:49:54 +0000 (15:19 +0530)]
arm: dts: k3-j721s2-r5: Change GTC clock parent

MAIN_PLL0 has a flag set in DM (Device Manager) that removes its
capability to re-initialise clock frequencies. A72 CPU clock (GTC) and
RGMII has MAIN_PLL3 as their parent which does not have this flag. While
RGMII needs re-initialization to default frequency to be able to get
250MHz with its divider, GTC can not get its required 200MHz with its
dividers. Thus move GTC clock parent on J721S2 from MAIN_PLL3_HSDIV1 to
MAIN_PLL0_HSDIV6. This was already done on CPTS node in kernel which was
similarly affected (linked).

Link: https://lore.kernel.org/all/20230605110443.84568-1-n-francis@ti.com/
Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
6 months agonet: phy: Replace PHY_ANEG_TIMEOUT with Kconfig symbol
Marek Vasut [Fri, 31 May 2024 16:47:17 +0000 (18:47 +0200)]
net: phy: Replace PHY_ANEG_TIMEOUT with Kconfig symbol

Switch PHY_ANEG_TIMEOUT to CONFIG_PHY_ANEG_TIMEOUT Kconfig symbol.
This removes one more configuration headers option finalizes its
Kconfig symbol conversion. No functional change expected.

Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
6 months agonet: phy: Turn default auto-negotiation timeout into Kconfig symbol
Marek Vasut [Fri, 31 May 2024 16:47:16 +0000 (18:47 +0200)]
net: phy: Turn default auto-negotiation timeout into Kconfig symbol

Let users configure default auto-negotiation timeout via Kconfig
instead of specifying it in board configuration headers. This is
the first step toward converting this to Kconfig fully, so far the
legacy PHY_ANEG_TIMEOUT in configuration headers takes precedence.

Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
6 months agoMerge patch series "Enable OSPI boot for j721s2"
Tom Rini [Thu, 13 Jun 2024 00:40:46 +0000 (18:40 -0600)]
Merge patch series "Enable OSPI boot for j721s2"

Manorit Chawdhry <m-chawdhry@ti.com> says:

The series enables ospi boot for j721s2.

Test logs: https://gist.github.com/manorit2001/6bb91885c608e3a8cb0267ab2c614781

6 months agoconfigs: j721s2_evm_*_defconfig: Enable OSPI configs
Manorit Chawdhry [Tue, 4 Jun 2024 06:09:13 +0000 (11:39 +0530)]
configs: j721s2_evm_*_defconfig: Enable OSPI configs

Enable OSPI related configs to boot using OSPI

Reviewed-by: Apurva Nandan <a-nandan@ti.com>
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
6 months agoarch: arm: dts: k3-j721s2-*-u-boot.dtsi: Enable the ospi0 node
Manorit Chawdhry [Tue, 4 Jun 2024 06:09:12 +0000 (11:39 +0530)]
arch: arm: dts: k3-j721s2-*-u-boot.dtsi: Enable the ospi0 node

Enable ospi0 node for all boot stages

Reviewed-by: Apurva Nandan <a-nandan@ti.com>
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
6 months agoarch: arm: dts: k3-j721s2-r5: Override ospi and fss for 32-bit mode
Manorit Chawdhry [Tue, 4 Jun 2024 06:09:11 +0000 (11:39 +0530)]
arch: arm: dts: k3-j721s2-r5: Override ospi and fss for 32-bit mode

R5 being a 32-bit processor can't understand the 64-bit mapping being
done in ospi node. Override the ospi node for 32-bit register ranges and
the fss node ( the parent node of ospi ) to map the ranges for the
updated child node correctly.

Reviewed-by: Apurva Nandan <a-nandan@ti.com>
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
6 months agomtd: spi-nor-core: Do not start or end writes at odd address in DTR mode
Pratyush Yadav [Tue, 4 Jun 2024 06:09:10 +0000 (11:39 +0530)]
mtd: spi-nor-core: Do not start or end writes at odd address in DTR mode

On DTR capable flashes like Micron Xcella the writes cannot start or end
at an odd address in DTR mode. Extra 0xff bytes need to be prepended or
appended respectively to make sure both the start and end addresses are
even.

Signed-off-by: Pratyush Yadav <p.yadav@ti.com>
Signed-off-by: Apurva Nandan <a-nandan@ti.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Tested-by: Jonathan Humphreys <j-humphreys@ti.com>
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
6 months agoMerge tag 'efi-next-20240611' of https://source.denx.de/u-boot/custodians/u-boot...
Tom Rini [Tue, 11 Jun 2024 13:42:55 +0000 (07:42 -0600)]
Merge tag 'efi-next-20240611' of https://source.denx.de/u-boot/custodians/u-boot-efi into next

Pull request efi-next-20240611

UEFI:

* Allow specifying a device-tree in an EFI load option
  using the efidebug or eficonfig command.
* Let the EFI boot manager fall back to an OS provided device-tree
  if no device-tree is specified.

6 months agoefi_loader: rename struct efi_initrd_dp to efi_lo_dp_prefix
Heinrich Schuchardt [Mon, 10 Jun 2024 08:00:01 +0000 (10:00 +0200)]
efi_loader: rename struct efi_initrd_dp to efi_lo_dp_prefix

As we now also store device-tree device-paths in load options rename
struct efi_initrd_dp to efi_lo_dp_prefix.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
6 months agoefi_loader: load distro dtb in bootmgr
Heinrich Schuchardt [Fri, 26 Apr 2024 14:13:21 +0000 (16:13 +0200)]
efi_loader: load distro dtb in bootmgr

If no device-tree is specified, try to load a device-tree from the boot
device use the $fdtfile concatenated to either of the paths '/dtb/', '/',
'/dtb/current/'.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agoefi_loader: export efi_load_image_from_path
Heinrich Schuchardt [Fri, 26 Apr 2024 14:13:20 +0000 (16:13 +0200)]
efi_loader: export efi_load_image_from_path

We can reuse this function to load the device-tree.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agoefi_loader: return binary from efi_dp_from_lo()
Heinrich Schuchardt [Fri, 26 Apr 2024 14:13:19 +0000 (16:13 +0200)]
efi_loader: return binary from efi_dp_from_lo()

For finding distro supplied device-trees we need to know from which device
we are booting. This can be identified via the device-path of the binary.

Up to now efi_dp_from_lo() only could return the initrd or fdt device-path.
Allow returning the binary device-path, too.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agoefi_loader: move distro_efi_get_fdt_name()
Heinrich Schuchardt [Fri, 26 Apr 2024 14:13:18 +0000 (16:13 +0200)]
efi_loader: move distro_efi_get_fdt_name()

Move distro_efi_get_fdt_name() to a separate C module
and rename it to efi_get_distro_fdt_name().

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agoefi_loader: load device-tree specified in boot option
Heinrich Schuchardt [Fri, 26 Apr 2024 14:13:17 +0000 (16:13 +0200)]
efi_loader: load device-tree specified in boot option

We allow to specify the triple of binary, initrd, and device-tree in boot
options.

Add the code to actually load the specified device-tree.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agocmd: efidebug: add support for setting fdt
Heinrich Schuchardt [Fri, 26 Apr 2024 14:13:12 +0000 (16:13 +0200)]
cmd: efidebug: add support for setting fdt

We already support creating a load option where the device-path
field contains the concatenation of the binary device-path and
optionally the device path of the initrd which we expose via the
EFI_LOAD_FILE2_PROTOCOL.

Allow to append another device-path pointing to the device-tree
identified by the device-tree GUID.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agocmd: eficonfig: add support for setting fdt
Heinrich Schuchardt [Fri, 26 Apr 2024 14:13:11 +0000 (16:13 +0200)]
cmd: eficonfig: add support for setting fdt

We already support creating a load option where the device-path
field contains the concatenation of the binary device-path and
optionally the device path of the initrd which we expose via the
EFI_LOAD_FILE2_PROTOCOL.

Allow to append another device-path pointing to the device-tree
identified by the device-tree GUID.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agoefi_loader: allow concatenation with contained end node
Heinrich Schuchardt [Fri, 24 May 2024 12:54:26 +0000 (14:54 +0200)]
efi_loader: allow concatenation with contained end node

Allow appending a device-path to a device-path that contains an end node
as separator. We need this feature for creating boot options specifying
kernel, initrd, and dtb.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agoAdded arm64 assembly for examples/api crt0
Brunham, Kalen [Fri, 17 May 2024 19:13:48 +0000 (19:13 +0000)]
Added arm64 assembly for examples/api crt0

I've encountered a problem when compiling the 'examples/api' directory for ARM64 in U-boot. The problem lies in the assembly code in 'examples/api/crt0.S' where the current CONFIG_ARM code is only 32-bit. When targeting ARM64, a 64-bit version is necessary.

I have proposed a fix by including a 'CONFIG_ARM64' section in the assembly code as shown below.  These changes have been check via https://github.com/u-boot/u-boot/pull/538.

Feedback is welcome.

Signed-off-by: Kalen Brunham <kalen.brunham@intel.com>
6 months agoabootimg: Add init_boot image support
Roman Stratiienko [Thu, 23 May 2024 07:06:07 +0000 (07:06 +0000)]
abootimg: Add init_boot image support

Quote from [1]:

"For devices launching with Android 13, the generic ramdisk is removed
from the boot image and placed in a separate init_boot image.
This change leaves the boot image with only the GKI kernel."

While at it, update wrong error handling message when vendor_boot
cannot be loaded.

[1]: https://source.android.com/docs/core/architecture/partitions/generic-boot

Signed-off-by: Roman Stratiienko <r.stratiienko@gmail.com>
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
6 months agoandroid: Fix ramdisk loading for bootimage v3+
Roman Stratiienko [Sun, 19 May 2024 13:09:51 +0000 (13:09 +0000)]
android: Fix ramdisk loading for bootimage v3+

The boot_ramdisk and vendor_ramdisk must be both concatenated together.
Without this change, Android root is missing some of the necessary tools
to complete virtual AB OTA.

Signed-off-by: Roman Stratiienko <r.stratiienko@gmail.com>
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
6 months agoarm: move _end to linker symbols
Ilias Apalodimas [Tue, 28 May 2024 06:18:27 +0000 (09:18 +0300)]
arm: move _end to linker symbols

commit 6e2228fb052b ("Merge patch series "Clean up arm linker scripts")
was cleaning up linker scripts for armv7 and v8 but was leaving
_end and __secure_stack_start/end.

commit d0b5d9da5de2 ("arm: make _end compiler-generated")
was moving _end to be compiler generated. _end is defined as c variable
in its own section to force the compiler emit relative a reference.
However, defining those in the linker script will do the same thing
since [0].

So let's remove the special sections from the linker scripts, the
variable definitions from sections.c and define them as a symbols.
It's worth noting that _image_binary_end symbol is now redundant and
can be removed in the future.

- SPL

The .end section has been removed from the new binary
[ 5] .end
     PROGBITS         00000000fffdf488  000000000002f488  0
     0000000000000000 0000000000000000  0                 1
     [0000000000000003]: WRITE, ALLOC

$~ bloat-o-meter kria_old/spl/u-boot-spl krina_new/spl/u-boot-spl
add/remove: 0/0 grow/shrink: 0/0 up/down: 0/0 (0)
Function                                     old     new   delta
Total: Before=115980, After=115980, chg +0.00%

$~ readelf -sW kria_old/u-boot kria_new/u-boot | grep -w _end
 12047: 000000000813a0f0     0 OBJECT  GLOBAL DEFAULT   11 _end
 12047: 000000000813a118     0 NOTYPE  GLOBAL DEFAULT   11 _end

$~ readelf -sW kria_old/spl/u-boot-spl kria_new/spl/u-boot-spl | grep -w _end
  1605: 00000000fffdf488     0 OBJECT  GLOBAL DEFAULT    5 _end
  1603: 00000000fffdf498     0 NOTYPE  GLOBAL DEFAULT    4 _end

$~ readelf -sW old/u-boot new/u-boot | grep -w _end
  8847: 0000000000103710     0 OBJECT  GLOBAL DEFAULT   11 _end
  8847: 0000000000103738     0 NOTYPE  GLOBAL DEFAULT   11 _end

$~ readelf -sW old_v7/u-boot new_v7/u-boot | grep -w _end
 10638: 000da824     0 OBJECT  GLOBAL DEFAULT   10 _end
 10637: 000da84c     0 NOTYPE  GLOBAL DEFAULT   10 _end

- For both QEMU instances
$~ bloat-o-meter old/u-boot new/u-boot
add/remove: 0/0 grow/shrink: 1/0 up/down: 20/0 (20)
Function                                     old     new   delta
version_string                                50      70     +20
Total: Before=656915, After=656935, chg +0.00%

[0] binutils commit 6b3b0ab89663 ("Make linker assigned symbol dynamic only for shared object")

Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
6 months agoimage: Set load_end on partial loads
Mattijs Korpershoek [Thu, 23 May 2024 09:27:09 +0000 (11:27 +0200)]
image: Set load_end on partial loads

When decompressing, it's possible that the algorithm only performs
a partial decompression.
This usually happens when CONFIG_SYS_BOOTM_LEN is too small for
the uncompressed image.

When that happens, image_decomp() returns an error and *load_end == load.
The error is then handled by handle_decomp_error().

handle_decomp_error() expects the number of uncompressed bytes in
uncomp_size but receives *load_end - load == load - load == 0.

Because of this, handle_decomp_error does not report the expected
"Image too large: increase CONFIG_SYS_BOOTM_LEN" error message.

Modify the image_decomp() logic to always report the decompressed size,
even when a partial decompression happened.

Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
6 months agofs: relax ext4_write_file() dependency
Baruch Siach [Wed, 15 May 2024 04:37:16 +0000 (07:37 +0300)]
fs: relax ext4_write_file() dependency

ext4_write_file() depends on CONFIG_EXT4_WRITE. Allow build without
CONFIG_CMD_EXT4_WRITE.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
6 months agodm: use newly added linux/compat alloc functions
Maxim Moskalets [Wed, 15 May 2024 20:13:30 +0000 (23:13 +0300)]
dm: use newly added linux/compat alloc functions

Signed-off-by: Maxim Moskalets <maximmosk4@gmail.com>
6 months agomemory: ti-gpmc: use printf to dump settings/timings
Roger Quadros [Wed, 15 May 2024 12:20:08 +0000 (15:20 +0300)]
memory: ti-gpmc: use printf to dump settings/timings

pr_info() depends on CONFIG_LOGLEVEL > 6. If user has
enabled CONFIG_TI_GPMC_DEBUG then we should print the
GPMC settings/timings regardless of CONFIG_LOGLEVEL.

So use printf() instead of pr_info().

Signed-off-by: Roger Quadros <rogerq@kernel.org>
6 months agoarm: dts: k3-j7200: Move to OF_UPSTREAM
Aniket Limaye [Tue, 21 May 2024 05:48:42 +0000 (11:18 +0530)]
arm: dts: k3-j7200: Move to OF_UPSTREAM

Move to using OF_UPSTREAM config and thus using the devicetree-rebasing
subtree.

Signed-off-by: Aniket Limaye <a-limaye@ti.com>
Acked-by: Sumit Garg <sumit.garg@linaro.org>
6 months agoinclude: extension_board: Increase overlay file name size
Daniel Schultz [Wed, 22 May 2024 07:21:00 +0000 (00:21 -0700)]
include: extension_board: Increase overlay file name size

Upstream overlays like the ARM64 TI
k3-am625-beagleplay-csi2-tevi-ov5640.dtso can easily have more then
32 characters. Increase the overlay length to 64 characters to
apply overlays with longer names.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Acked-by: Kory Maincent <kory.maincent@bootlin.com>
6 months agoMerge patch series "*** phyCORE-AM62x: DDR detection / Inject DDR timing deltas ***"
Tom Rini [Fri, 7 Jun 2024 20:03:36 +0000 (14:03 -0600)]
Merge patch series "*** phyCORE-AM62x: DDR detection / Inject DDR timing deltas ***"

Wadim Egorov <w.egorov@phytec.de> says:

Changes in v2:
  - Reabse to current next
  - Add Tested-by: John Ma <jma@phytec.com>
  - Add Kconfig option to select RAM size statically
  - Make board/phytec/common/k3 always compile for CONFIG_ARCH_K3

v1: https://lists.denx.de/pipermail/u-boot/2024-May/553057.html

6 months agoboard: phytec: am62x: Add support for 1 & 4 GB RAM variants
Wadim Egorov [Wed, 22 May 2024 07:55:05 +0000 (09:55 +0200)]
board: phytec: am62x: Add support for 1 & 4 GB RAM variants

Use content of EEPROM to detect the actual RAM size and adjust
DDR timings, size and banks accordingly.
Also enable the SoM detection per default in the defconfigs.

Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: John Ma <jma@phytec.com>
6 months agoboard: phytec: common: Introduce a method to inject DDR timings deltas
Wadim Egorov [Wed, 22 May 2024 07:55:04 +0000 (09:55 +0200)]
board: phytec: common: Introduce a method to inject DDR timings deltas

Introduce fdt_apply_ddrss_timings_patch() to allow board code to
override DDR settings in the device tree prior to DDRSS driver probing.

Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: John Ma <jma@phytec.com>
6 months agoarm: mach-k3: am625: Call do_board_detect() before DDR probing
Wadim Egorov [Wed, 22 May 2024 07:55:03 +0000 (09:55 +0200)]
arm: mach-k3: am625: Call do_board_detect() before DDR probing

Call do_board_detect() hook before the K3 DDRSS driver gets probed.
It will allow boards to adjust DDR timings in do_board_detect().

Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: John Ma <jma@phytec.com>
6 months agoboard: phytec: Fix function definitions in AM6x SOM detection
Wadim Egorov [Wed, 22 May 2024 07:55:02 +0000 (09:55 +0200)]
board: phytec: Fix function definitions in AM6x SOM detection

Functions are declared as phytec_am6* and not phytec_am62*.
Update the definitions to match the declarations.

Fixes: 9d152c23279c ("board: phytec: Add SOM detection for AM6x")
Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: John Ma <jma@phytec.com>
Reviewed-by: Dhruva Gole <d-gole@ti.com>
6 months agoboard: phytec: Make AM6 SoM detection depend on I2C
Wadim Egorov [Wed, 22 May 2024 07:55:01 +0000 (09:55 +0200)]
board: phytec: Make AM6 SoM detection depend on I2C

SoM detection is using I2C driver model functions.
Let's depend on I2C.

Signed-off-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: John Ma <jma@phytec.com>
Reviewed-by: Dhruva Gole <d-gole@ti.com>
6 months agoMerge patch series "PHYTEC SOM Detection API v3"
Tom Rini [Fri, 7 Jun 2024 20:02:02 +0000 (14:02 -0600)]
Merge patch series "PHYTEC SOM Detection API v3"

Daniel Schultz <d.schultz@phytec.de> says:

This patch series adds support for the EEPROM v3 API.

V3 is backwards compatible to V2 and therefore, the V2 image still
exists at the beginning. Only the API version changed from 2 to 3.

V3 is a block-based memory layout organized as singled-linked list
with different types of blocks. This is a more flexible approach and
allows us to extend it by more block types in the future.

The V3 data starts with a 8-byte large header which defines the
block count (u8), V3 subversion (u8) and data payload length (u16).
Additionally the header contains a CRC8 checksum a 3 reserved bytes.

Each block starts with a 4-byte large header which defined the
block type (u8), the absolute address of the next block (u16) and a
CRC8 checksum. The content itself is defined via the block type and
we currently have 2 different types:

1) MAC: Contains the Ethernet interface number (u8), MAC address
        (6 x u8) and a CRC8 checksum.

6 months agoconfigs: phycore_am62x_a53_defconfig: Enable CONFIG_ENV_OVERWRITE
Daniel Schultz [Wed, 22 May 2024 06:18:27 +0000 (23:18 -0700)]
configs: phycore_am62x_a53_defconfig: Enable CONFIG_ENV_OVERWRITE

Enable CONFIG_ENV_OVERWRITE to overwrite ethaddr in the environment.
This is required because our environment is not located in the
boot partition.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Tested-by: Wadim Egorov <w.egorov@phytec.de>
6 months agoboard: phytec: common: k3: Set MAC
Daniel Schultz [Wed, 22 May 2024 06:18:26 +0000 (23:18 -0700)]
board: phytec: common: k3: Set MAC

Read the EEPROM API v3 content and set all available
MAC-Addresses to the environment.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Reviewed-by: Wadim Egorov <w.egorov@phytec.de>
Tested-by: Wadim Egorov <w.egorov@phytec.de>
6 months agoboard: phytec: common: Add API v3
Daniel Schultz [Wed, 22 May 2024 06:18:25 +0000 (23:18 -0700)]
board: phytec: common: Add API v3

This API is based on a block structure with a 8 Byte large
API v3 header and various of different blocks following. It extends
our current API v2, which is always 32 Byte large, and is located
directly after v2.

Add the MAC block as first block type. It contains the physical
Ehternet interface number, a MAC address and a CRC checksum over the
MAC payload.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Tested-by: Wadim Egorov <w.egorov@phytec.de>
6 months agoboard: phytec: common: Move API v2 init to new function
Daniel Schultz [Wed, 22 May 2024 06:18:24 +0000 (23:18 -0700)]
board: phytec: common: Move API v2 init to new function

Move the entire initialization code for API v2 into a dedicated
function. This rework will allow to easily integrate the API v3
as next step during init.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Tested-by: Wadim Egorov <w.egorov@phytec.de>
6 months agoboard: phytec: common: Define PHYTEC_API2_DATA_LEN
Daniel Schultz [Wed, 22 May 2024 06:18:23 +0000 (23:18 -0700)]
board: phytec: common: Define PHYTEC_API2_DATA_LEN

The EEPROM image length for API v2 is fixed to 32 bytes. No need
to use sizeof while this value won't change. This value is
also be required for API v3 to know where the API v3 header starts.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Tested-by: Wadim Egorov <w.egorov@phytec.de>
6 months agoboard: phytec: common: Move eeprom read to new function
Daniel Schultz [Wed, 22 May 2024 06:18:22 +0000 (23:18 -0700)]
board: phytec: common: Move eeprom read to new function

We need to read multiple times from different offsets in API v3.
Move the EEPROM read logic into a dedicated function to make it
usable multiple times.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Tested-by: Wadim Egorov <w.egorov@phytec.de>
6 months agoMerge patch series "Add remoteproc support for J784S4 SoCs"
Tom Rini [Fri, 7 Jun 2024 20:01:38 +0000 (14:01 -0600)]
Merge patch series "Add remoteproc support for J784S4 SoCs"

Beleswar Padhi <b-padhi@ti.com> says:

Hello All,

This series adds remoteproc support for TI's J784S4 SoCs. The K3 J784S4
SoCs have four dual-core R5F subsystems and four C71x DSP subsystems.
Enable the remoteproc drivers in defconfig and set the rproc firmware
names to add remoteproc support.

Note: No driver changes are required as J784S4 SoCs have the same data
as J721S2 SoCs. Thus, utilize the existing compatible string for driver
probe.

6 months agoboard: ti: j784s4: j784s4.env: Set remoteproc firmware names
Beleswar Padhi [Wed, 22 May 2024 04:53:31 +0000 (10:23 +0530)]
board: ti: j784s4: j784s4.env: Set remoteproc firmware names

Include k3_rproc.env to access rproc boot commands and specify
rproc firmware names for adding remoteproc support in J784S4 SoCs.

Signed-off-by: Beleswar Padhi <b-padhi@ti.com>
6 months agoconfigs: j784s4_evm_a72_defconfig: Enable remoteproc drivers
Beleswar Padhi [Wed, 22 May 2024 04:53:30 +0000 (10:23 +0530)]
configs: j784s4_evm_a72_defconfig: Enable remoteproc drivers

The K3 J784S4 SoC has four dual-core R5F subsystems and four C71x DSP
subsystems. Set config values to enable the remoteproc functionality
with these R5F and DSP subsystems.

Signed-off-by: Beleswar Padhi <b-padhi@ti.com>
6 months agoMerge patch series "Adding support to load secure firmware for HS devices"
Tom Rini [Fri, 7 Jun 2024 20:01:15 +0000 (14:01 -0600)]
Merge patch series "Adding support to load secure firmware for HS devices"

Udit Kumar <u-kumar1@ti.com> says:

Some use case needs rproc firmware to be loaded at u-boot stage,
using following commands at u-boot shell, firmware could be loaded

=> setenv dorprocboot 1
=> run boot_rprocs

For Secure devices, secure version of rproc firmware should be loaded,
which is appended by sec keyword[0].
but currently non-secure firmware is loaded even for secure devices.
So adding support for loading secure firmware on Secured devices.

[0]: https://gist.github.com/uditkumarti/cd8bf6a448079b59145d17a0e8bf13b7

Bootlogs:
GP : https://gist.github.com/uditkumarti/23a00c313e1c28b62537aab733a585df#file-gp_device line 65 onwards
HS : https://gist.github.com/uditkumarti/23a00c313e1c28b62537aab733a585df#file-hs-device line 60 onwards

6 months agomach-k3: common.c: add a flag for booting authenticated rproc binaries
Manorit Chawdhry [Tue, 21 May 2024 10:56:48 +0000 (16:26 +0530)]
mach-k3: common.c: add a flag for booting authenticated rproc binaries

The flag will be used for booting authenticated remote procs from hs-se
devices which can optionally be used in hs-fs devices also.

Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> # Intel Edison
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
6 months agoinclude: env: ti: Add support for secure firmwares
Udit Kumar [Tue, 21 May 2024 10:56:47 +0000 (16:26 +0530)]
include: env: ti: Add support for secure firmwares

Secure firmwares must be loaded if SOC is secure,
currently rproc framework chooses non-secure firmware always.

So adding support to load secure firmware, when SOC is secure

Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> # Intel Edison
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
6 months agodrivers: remoteproc: ti_k3 : enable secure booting with firmware images
Manorit Chawdhry [Tue, 21 May 2024 10:56:46 +0000 (16:26 +0530)]
drivers: remoteproc: ti_k3 : enable secure booting with firmware images

Remoteproc firmware images aren't authenticated in the current boot flow.
Authenticates remoteproc firmware images to complete the root of trust
in secure booting.

Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> # Intel Edison
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
6 months agoinclude: mach-k3: move k3 security functions to security.h
Manorit Chawdhry [Tue, 21 May 2024 10:56:45 +0000 (16:26 +0530)]
include: mach-k3: move k3 security functions to security.h

ti_secure_image_post_process and ti_secure_image_check_binary is used
for the authentication purposes in the current boot flow. Authentication
of remoteproc firmware images require ti_secure_image_post_process to be
available outside mach-k3.

Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> # Intel Edison
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
6 months agoMerge patch series "k3-j721e: Enable OF_UPSTREAM for J721E"
Tom Rini [Fri, 7 Jun 2024 20:00:26 +0000 (14:00 -0600)]
Merge patch series "k3-j721e: Enable OF_UPSTREAM for J721E"

Neha Malcom Francis <n-francis@ti.com> says:

This series was plucked out from the larger series [1] that did some
templating reformatting and also enabled OF_UPSTREAM for J721E. The
former has been kept aside till all the platforms affected have moved to
using OF_UPSTREAM to have less conflicts while merging.

Patches split J721E EVM and J721E SK to using separate builds, as well
as enable OF_UPSTREAM for both the platforms.

Boot logs:
https://gist.github.com/nehamalcom/8f326376b6c6b1196084721405159bb9

[1] https://lore.kernel.org/all/20240322131011.1029620-1-n-francis@ti.com/

6 months agoarm: dts: k3-j721e: Move to OF_UPSTREAM
Neha Malcom Francis [Mon, 20 May 2024 09:59:16 +0000 (15:29 +0530)]
arm: dts: k3-j721e: Move to OF_UPSTREAM

Move to using OF_UPSTREAM config and thus using the devicetree-rebasing
subtree.

Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
Acked-by: Sumit Garg <sumit.garg@linaro.org>
6 months agoconfigs: j721e_sk: Move to separate defconfig for J721E SK board
Neha Malcom Francis [Mon, 20 May 2024 09:59:15 +0000 (15:29 +0530)]
configs: j721e_sk: Move to separate defconfig for J721E SK board

Add defconfig for J721E SK R5 and A72 configuration.

This includes and modifies the J721E EVM defconfigs:
j721e_evm_r5_defconfig -> j721e_sk_r5_defconfig
j721e_evm_a72_defconfig -> j721e_sk_a72_defconfig

Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
6 months agoarm: dts: k3-j721e-r5*: Introduce k3-j721e-r5.dtsi
Neha Malcom Francis [Mon, 20 May 2024 09:59:14 +0000 (15:29 +0530)]
arm: dts: k3-j721e-r5*: Introduce k3-j721e-r5.dtsi

Introduce k3-j721e-r5.dtsi to be used by board R5 DTS files. This
helps sync SoC changes across boards.

Signed-off-by: Neha Malcom Francis <n-francis@ti.com>
Reviewed-by: Manorit Chawdhry <m-chawdhry@ti.com>
6 months agoMerge patch series "arm: dts: k3-am625-verdin: Enable LPDDR4 WDQS control"
Tom Rini [Fri, 7 Jun 2024 19:58:25 +0000 (13:58 -0600)]
Merge patch series "arm: dts: k3-am625-verdin: Enable LPDDR4 WDQS control"

Emanuele Ghidoli <emanuele.ghidoli@toradex.com> says:

Manually, since SysConfig tool do not have the relevant option,
set PHY_LP4_WDQS_OE_EXTEND to 1.
Since WDQS control mode is required on our modules LPDDR4,
this enables WDQS control mode 1.

6 months agoarm: dts: k3-am625-verdin: Enable LPDDR4 WDQS control
Emanuele Ghidoli [Wed, 15 May 2024 08:00:58 +0000 (10:00 +0200)]
arm: dts: k3-am625-verdin: Enable LPDDR4 WDQS control

Manually, since SysConfig tool do not have the relevant option,
set PHY_LP4_WDQS_OE_EXTEND to 1.
Since WDQS control mode is required on our modules LPDDR4,
this enables WDQS control mode 1.

Signed-off-by: Emanuele Ghidoli <emanuele.ghidoli@toradex.com>
6 months agoarm: dts: k3-am625-verdin: Update autogenerated LPDDR4 configuration
Emanuele Ghidoli [Wed, 15 May 2024 08:00:57 +0000 (10:00 +0200)]
arm: dts: k3-am625-verdin: Update autogenerated LPDDR4 configuration

Update the autogenerated LPDDR4 configuration using the latest available
SysConfig tool.
This changes are cosmetic and are made to track the last used tool version.

Signed-off-by: Emanuele Ghidoli <emanuele.ghidoli@toradex.com>
6 months agoMerge patch series "'eeprom' command improvements"
Tom Rini [Fri, 7 Jun 2024 16:48:41 +0000 (10:48 -0600)]
Merge patch series "'eeprom' command improvements"

Marek Behún <kabel@kernel.org> says:

This series contains improvements for the 'eeprom' command:
- refactors
- fixes
- improvements
- ability to use driver model EEPROMs (uclass UCLASS_I2C_EEPROM)
- more flexible EEPROM layout support

It should not cause any behavior change for any existing board.

This series is a dependency for some DDR issue fixes for Turris Omnia.
I will be sending that one separately.

github PR link (with CI):
  https://github.com/u-boot/u-boot/pull/540
- there is a failure for
    test.py for sandbox sandbox_clang
  but it seems unrelated to these changes

6 months agocmd: eeprom: Extend to EEPROMs probed via driver model
Marek Behún [Tue, 21 May 2024 07:13:35 +0000 (09:13 +0200)]
cmd: eeprom: Extend to EEPROMs probed via driver model

Extend the 'eeprom' command to allow accessing EEPROMs probed via
driver model, uclass UCLASS_I2C_EEPROM.

When the CONFIG_I2C_EEPROM config option is enabled (and so the
i2c-eeprom driver is built), the 'eeprom' command now accepts driver
model device name as EEPROM specifier for the 'eeprom' command, in
addition to the legacy [[bus] devaddr] specifier.

Moreover if no device specifier is given, then the first
UCLASS_I2C_EEPROM device is used, if found.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocmd: eeprom: Don't read/write whole EEPROM if not necessary
Marek Behún [Tue, 21 May 2024 07:13:34 +0000 (09:13 +0200)]
cmd: eeprom: Don't read/write whole EEPROM if not necessary

Don't read/write whole EEPROM if not necessary when printing / updating
EEPROM layout fields. Only read/write layout.data_size bytes.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocmd: eeprom: Refactor command execution into function by action
Marek Behún [Tue, 21 May 2024 07:13:33 +0000 (09:13 +0200)]
cmd: eeprom: Refactor command execution into function by action

Refactor the eeprom_execute_command() function into separate functions
do_eeprom_rw(), do_eeprom_print() and do_eeprom_update().

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocmd: eeprom: Refactor eeprom device specifier parsing
Marek Behún [Tue, 21 May 2024 07:13:32 +0000 (09:13 +0200)]
cmd: eeprom: Refactor eeprom device specifier parsing

In preparation for allowing to access eeprom by driver-model device
name, refactor the eeprom device specifier parsing. Instead of filling
two parameters (i2c_bus, i2c_addr), the parsing function now fills one
parameter of type struct eeprom_dev_spec.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocmd: eeprom: Deduplicate parse_i2c_bus_addr() calls
Marek Behún [Tue, 21 May 2024 07:13:31 +0000 (09:13 +0200)]
cmd: eeprom: Deduplicate parse_i2c_bus_addr() calls

Deduplicate the calls to parse_i2c_bus_addr().

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocmd: eeprom: Hide eeprom layout versioning behind a Kconfig option
Marek Behún [Tue, 21 May 2024 07:13:30 +0000 (09:13 +0200)]
cmd: eeprom: Hide eeprom layout versioning behind a Kconfig option

Add a new Kconfig option EEPROM_LAYOUT_VERSIONS, and hide eeprom
layout versionsing code behind it. Only print the relevant help in
'eeprom' command usage if this option is enabled.

Enable this new option for cm_fx6_defconfig and cm_t43_defconfig.
These are the only boards using EEPROM layout versioning.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocmd: eeprom: Fix usage help for the eeprom command
Marek Behún [Tue, 21 May 2024 07:13:29 +0000 (09:13 +0200)]
cmd: eeprom: Fix usage help for the eeprom command

The bus and devaddr arguments of the eeprom command are optional, and if
only one is given, it is assumed to be devaddr. Change the usage help
from
  <bus> <devaddr>
to
  [[bus] [devaddr]

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocommon: eeprom_field: Drop unnecessary comparison
Marek Behún [Tue, 21 May 2024 07:13:28 +0000 (09:13 +0200)]
common: eeprom_field: Drop unnecessary comparison

The byte variable is of type unsigned char, it is never less than zero.
The error case is handled by *endptr, so drop the comparison altogether.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocommon: eeprom_field: Fix updating binary field
Marek Behún [Tue, 21 May 2024 07:13:27 +0000 (09:13 +0200)]
common: eeprom_field: Fix updating binary field

The __eeprom_field_update_bin() function is expected to parse a hex
string into bytes (potentially in reverse order), but the
simple_strtoul() function is given 0 as base. This does not work since
the string does not contain '0x' prefix. Add explicit base 16.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocommon: eeprom_layout: Split field finding code from the field update function
Marek Behún [Tue, 21 May 2024 07:13:26 +0000 (09:13 +0200)]
common: eeprom_layout: Split field finding code from the field update function

Split the eeprom layout field finding code from the
eeprom_layout_update_field() function in order to make it usable in
alternative implementations of update method.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agocommon: eeprom_layout: Assign default layout methods and parameters before specific...
Marek Behún [Tue, 21 May 2024 07:13:25 +0000 (09:13 +0200)]
common: eeprom_layout: Assign default layout methods and parameters before specific ones

Assign the default eeprom layout parameter .data_size and methods
.print() and .update() before calling eeprom_layout_assign() in
eeprom_layout_setup().

This allows eeprom_layout_assign() to overwrite these if needed.

Signed-off-by: Marek Behún <kabel@kernel.org>
6 months agoMerge tag 'v2024.07-rc4' into next
Tom Rini [Tue, 4 Jun 2024 00:42:11 +0000 (18:42 -0600)]
Merge tag 'v2024.07-rc4' into next

Prepare v2024.070-rc4

6 months agoPrepare v2024.07-rc4
Tom Rini [Tue, 4 Jun 2024 00:34:59 +0000 (18:34 -0600)]
Prepare v2024.07-rc4

Signed-off-by: Tom Rini <trini@konsulko.com>
6 months agoMerge tag 'u-boot-imx-next-20240603' of https://gitlab.denx.de/u-boot/custodians...
Tom Rini [Mon, 3 Jun 2024 17:42:51 +0000 (11:42 -0600)]
Merge tag 'u-boot-imx-next-20240603' of https://gitlab.denx.de/u-boot/custodians/u-boot-imx into next

CI: https://source.denx.de/u-boot/custodians/u-boot-imx/-/pipelines/20956

- Support different RAM sizes on imx8m phycoce boards.
- Support new toradex variants.
- Support Samsung 4GB DDR and Realtek RTL8211E PHY on imx8mm-cl-iot-gate.
- Convert imx8mm-phycore and imx8mp-phycore boards to use OF_UPSTREAM.

6 months agoMerge branch 'next-cleanup' of https://source.denx.de/u-boot/custodians/u-boot-sh...
Tom Rini [Mon, 3 Jun 2024 17:42:38 +0000 (11:42 -0600)]
Merge branch 'next-cleanup' of https://source.denx.de/u-boot/custodians/u-boot-sh into next

6 months agoMerge https://source.denx.de/u-boot/custodians/u-boot-samsung
Tom Rini [Mon, 3 Jun 2024 15:48:03 +0000 (09:48 -0600)]
Merge https://source.denx.de/u-boot/custodians/u-boot-samsung

6 months agoMerge branch 'master-cleanup' of https://source.denx.de/u-boot/custodians/u-boot-sh
Tom Rini [Mon, 3 Jun 2024 15:47:43 +0000 (09:47 -0600)]
Merge branch 'master-cleanup' of https://source.denx.de/u-boot/custodians/u-boot-sh

6 months agoimx8mm-cl-iot-gate: Add support for the Realtek RTL8211E PHY
Fabio Estevam [Tue, 28 May 2024 19:15:10 +0000 (16:15 -0300)]
imx8mm-cl-iot-gate: Add support for the Realtek RTL8211E PHY

Newer imx8mm-cl-iot-gate versions are populated with a Realtek RTL8211E
PHY instead of the Atheros AR8033.

Adapted Compulab's patch from:
https://github.com/compulab-yokneam/meta-bsp-imx8mm/blob/iot-gate-imx8_5.10.72/recipes-bsp/u-boot/compulab/imx8mm/0125-imx8mm-net-enable-phy-Realtek-RTL8211E.patch

to support both PHYs in U-Boot.

Signed-off-by: Fabio Estevam <festevam@denx.de>
6 months agoimx8mm-cl-iot-gate: Add support for Samsung 4GB DDR
Fabio Estevam [Tue, 28 May 2024 19:15:09 +0000 (16:15 -0300)]
imx8mm-cl-iot-gate: Add support for Samsung 4GB DDR

Newer versions of the imx8mm-cl-iot-gate boards may come populated with a
Samsung 4GB DDR model.

Add support for it.

Signed-off-by: Fabio Estevam <festevam@denx.de>