pandora-u-boot.git
4 months agotest: Drop unused suite prototypes
Simon Glass [Mon, 20 Jan 2025 21:25:23 +0000 (14:25 -0700)]
test: Drop unused suite prototypes

Drop some the prototypes for functions which were removed in earlier
series.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agoMerge tag 'u-boot-imx-master-20250120' of https://gitlab.denx.de/u-boot/custodians...
Tom Rini [Mon, 20 Jan 2025 18:08:16 +0000 (12:08 -0600)]
Merge tag 'u-boot-imx-master-20250120' of https://gitlab.denx.de/u-boot/custodians/u-boot-imx

CI: https://source.denx.de/u-boot/custodians/u-boot-imx/-/pipelines/24263

- Add i.MX95 EMDIO support
- Guard binman nodes with CONFIG_OPTEE on imx8m
- Enable CAAM in phycore-imx8mp SPL.
- Fix Fix NULL dereference in imx_pinctrl_probe().

4 months agoarch: arm: mach-k3: Delete tifs node in DT fixup
Udit Kumar [Mon, 13 Jan 2025 04:25:24 +0000 (09:55 +0530)]
arch: arm: mach-k3: Delete tifs node in DT fixup

Delete tifs DT node as part of fixup.

TISCI API reported msmc_size, does not include 64KB reserved size for
tifs aka MSMC comms memory, see documentation[0].

As part of fixup, original code uses TISCI API reported msmc_size as
size for SRAM DT node.

tifs node is similar to l3-cache, which should hold address above
msmc_size, and should be deleted before passing control to OS.

[0] https://software-dl.ti.com/tisci/esd/latest/2_tisci_msgs/general/core.html?highlight=msmc#tisci-msg-query-msmc

Reviewed-by: Neha Malcom Francis <n-francis@ti.com>
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
4 months agoboard: ti: j722s: j722s.env: Set remoteproc firmware names
Beleswar Padhi [Wed, 8 Jan 2025 10:25:16 +0000 (15:55 +0530)]
board: ti: j722s: j722s.env: Set remoteproc firmware names

Include k3_rproc.env to access rproc boot commands and specify rproc
firmware names for adding remoteproc support in J722S SoCs.

Signed-off-by: Beleswar Padhi <b-padhi@ti.com>
Reviewed-by: Bryan Brattlof <bb@ti.com>
4 months agoarch: arm: mach-k3: j784s4_init: Set CTRL_MMR for AUDIO_REFCLK1 clk_sel
Jayesh Choudhary [Mon, 6 Jan 2025 10:04:00 +0000 (15:34 +0530)]
arch: arm: mach-k3: j784s4_init: Set CTRL_MMR for AUDIO_REFCLK1 clk_sel

The default value for the mux to select the parent clock,
AUDIO_REFCLK1_CTRL_CLK_SEL is '11111' (31) but the mux input for 31
is marked as 'Reserved' so the ti-sci-clk call for get-parent fails.
Mark it to a valid value, '11100' (28) for MAIN_PLL4_HSDIV2_CLKOUT
to get rid of the linux failures during boot-time like:
"[    1.573193] ti-sci-clk 44083000.system-controller:clock-controller:
 get-parent failed for dev=157, clk=34, ret=-19"

Signed-off-by: Jayesh Choudhary <j-choudhary@ti.com>
4 months agofs: eliminate YAFFS2 implementation
Heinrich Schuchardt [Sat, 4 Jan 2025 00:34:02 +0000 (01:34 +0100)]
fs: eliminate YAFFS2 implementation

Upstream development stopped 2012.
Linux eliminated YAFFS2 in 2010.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
4 months agophycore-imx8mp: Enable CAAM in spl
Leonard Anderweit [Fri, 17 Jan 2025 13:20:13 +0000 (14:20 +0100)]
phycore-imx8mp: Enable CAAM in spl

Enable CAAM in spl.

Signed-off-by: Leonard Anderweit <l.anderweit@phytec.de>
4 months agopinctrl: imx: Fix NULL dereference in imx_pinctrl_probe()
Jesse Taube [Fri, 17 Jan 2025 03:00:48 +0000 (22:00 -0500)]
pinctrl: imx: Fix NULL dereference in imx_pinctrl_probe()

When converting to ofnode `ofnode_read_u32` was accedentally used to
replace `fdtdec_get_int` instead of `ofnode_read_u32_default`.
Use `ofnode_read_u32_default` to fix this.

Fixes: 59382d2 ("pinctrl: imx: Convert to use livetree API for fdt access")
Signed-off-by: Jesse Taube <Mr.Bossman075@gmail.com>
4 months agoimx8m: Guard binman nodes with CONFIG_OPTEE
Fabio Estevam [Fri, 17 Jan 2025 01:56:42 +0000 (22:56 -0300)]
imx8m: Guard binman nodes with CONFIG_OPTEE

Guard binman nodes with CONFIG_OPTEE to fix the following error
when building without optee support:

  BINMAN  .binman_stamp
Image 'image' has faked external blobs and is non-functional: tee.bin

Image 'image' is missing optional external blobs but is still functional: tee-os

/binman/section/fit/images/tee/tee-os (tee.bin):
   See the documentation for your board. You may need to build Open Portable
   Trusted Execution Environment (OP-TEE) and build with TEE=/path/to/tee.bin

Some images are invalid
make: *** [Makefile:1135: .binman_stamp] Error 103

While at it, only pass tee to the loadables lines when CONFIG_OPTEE is
defined.

Reported-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Fabio Estevam <festevam@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Tested-by: Yannic Moog <y.moog@phytec.de>
4 months agonet: fsl_enetc: Update enetc driver to support i.MX95
Alice Guo [Thu, 16 Jan 2025 04:03:30 +0000 (05:03 +0100)]
net: fsl_enetc: Update enetc driver to support i.MX95

i.MX95 uses enetc version 4.1 controller. Update the enetc for i.MX95.
Add ARM-specific cache handling and i.MX95 specific register layout
handling.

Signed-off-by: Alice Guo <alice.guo@nxp.com>
Signed-off-by: Marek Vasut <marex@denx.de> # Clean up
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Tim Harvey <tharvey@gateworks.com>
4 months agonet: fsl_enetc: add i.MX95 EMDIO support
Alice Guo [Thu, 16 Jan 2025 04:03:29 +0000 (05:03 +0100)]
net: fsl_enetc: add i.MX95 EMDIO support

The verdor ID and device ID of i.MX95 EMDIO are different from LS1028A
EMDIO, so add new vendor ID and device ID to pci_device_id table to
support i.MX95 EMDIO.

Signed-off-by: Alice Guo <alice.guo@nxp.com>
Signed-off-by: Marek Vasut <marex@denx.de> # Clean up
Signed-off-by: Ye Li <ye.li@nxp.com>
4 months agonet: fsl_enetc: Add initial netc-blk-ctrl driver support
Marek Vasut [Thu, 16 Jan 2025 04:03:28 +0000 (05:03 +0100)]
net: fsl_enetc: Add initial netc-blk-ctrl driver support

The netc-blk-ctrl driver is used to configure Integrated Endpoint
Register Block (IERB) and Privileged Register Block (PRB) of NETC.
For i.MX platforms, it is also used to configure the NETCMIX block.

The IERB contains registers that are used for pre-boot initialization,
debug, and non-customer configuration. The PRB controls global reset
and global error handling for NETC. The NETCMIX block is mainly used
to set MII protocol and PCS protocol of the links, it also contains
settings for some other functions.

Note the IERB configuration registers can only be written after being
unlocked by PRB, otherwise, all write operations are inhibited. A warm
reset is performed when the IERB is unlocked, and it results in an FLR
to all NETC devices. Therefore, all NETC device drivers must be probed
or initialized after the warm reset is finished.

Ported from Linux 6.13-rc as of commit
fe5ba6bf91b3 ("net: enetc: add initial netc-blk-ctrl driver support")

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Pass udevice pointer to accessors
Marek Vasut [Thu, 16 Jan 2025 04:03:27 +0000 (05:03 +0100)]
net: fsl_enetc: Pass udevice pointer to accessors

Pass struct udevice * into the register accessors, so the accessors can reach
driver data, which contain device specific register offsets.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Introduce driver data
Marek Vasut [Thu, 16 Jan 2025 04:03:26 +0000 (05:03 +0100)]
net: fsl_enetc: Introduce driver data

Introduce driver data for each PCI device. The driver data carry
offsets of registers which differ between different SoCs.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Split register accessors
Marek Vasut [Thu, 16 Jan 2025 04:03:25 +0000 (05:03 +0100)]
net: fsl_enetc: Split register accessors

Split register accessors to the port base/station interface/port/mac
registers as those are at different offsets on different SoCs. This
is a preparatory patch which will allow addition of adjusted offsets
for new SoCs easily.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Inline register accessors
Marek Vasut [Thu, 16 Jan 2025 04:03:24 +0000 (05:03 +0100)]
net: fsl_enetc: Inline register accessors

Move register accessors from header files and turn them into proper
inline functions, so typechecking can be done on them. Drop no longer
enetc_port_regs() and unused enetc_read() and enetc_bdr_read().

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Introduce enetc_dev_id()
Marek Vasut [Thu, 16 Jan 2025 04:03:23 +0000 (05:03 +0100)]
net: fsl_enetc: Introduce enetc_dev_id()

Introduce mapping function enetc_dev_id(), which converts PCIe BDF of
the ENETC into linear incrementing index usable e.g. as interface index.
This replaces the current ad-hoc calculation used in the code with a
dedicated function. No functional change.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Rename the driver and related structures
Marek Vasut [Thu, 16 Jan 2025 04:03:22 +0000 (05:03 +0100)]
net: fsl_enetc: Rename the driver and related structures

Rename the current driver structure and matching ops and PCI IDs
and add _ls suffix to indicate this content is LS specific. This
is done in preparation for addition of i.MX95 ENETCv4 which will
require slightly different structure content.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Pass udevice pointer into enetc_enable_si_port()
Marek Vasut [Thu, 16 Jan 2025 04:03:21 +0000 (05:03 +0100)]
net: fsl_enetc: Pass udevice pointer into enetc_enable_si_port()

Pass udevice pointer into enetc_enable_si_port() so tests like
enetc_is_ls1028a() an be used in the function. No functional
change.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Fold DT MAC address update into enetc_set_ierb_primary_mac()
Marek Vasut [Thu, 16 Jan 2025 04:03:20 +0000 (05:03 +0100)]
net: fsl_enetc: Fold DT MAC address update into enetc_set_ierb_primary_mac()

The entire content of the loop can be folded into enetc_set_ierb_primary_mac(),
do it. This changes the behavior slightly such that the DT is only updated in
case of a LS1028A, which is the only SoC with ethernet MAC path in DT matching
"/soc/pcie@1f0000000/ethernet@%x,%x" anyway, so this slight change should have
no impact.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Remove ifdef around enetc_set_ierb_primary_mac()
Marek Vasut [Thu, 16 Jan 2025 04:03:19 +0000 (05:03 +0100)]
net: fsl_enetc: Remove ifdef around enetc_set_ierb_primary_mac()

Use enetc_is_ls1028() instead of ifdef around enetc_set_ierb_primary_mac()
and clean up the function. No functional change intended.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agonet: fsl_enetc: Introduce enetc_is_ls1028a()
Marek Vasut [Thu, 16 Jan 2025 04:03:18 +0000 (05:03 +0100)]
net: fsl_enetc: Introduce enetc_is_ls1028a()

Introduce accurate test for LS1028A compatibility based both on
IS_ENABLED(CONFIG_ARCH_LS1028A) and PCI vendor ID. This is done
in preparation for adding ENETCv4 support, which has a different
PCI vendor ID.

Signed-off-by: Marek Vasut <marex@denx.de>
4 months agoMerge patch series "add the support of sha256_hmac and sha256_hkdf"
Tom Rini [Sat, 18 Jan 2025 23:13:01 +0000 (17:13 -0600)]
Merge patch series "add the support of sha256_hmac and sha256_hkdf"

Philippe Reynes <philippe.reynes@softathome.com> says:

This serie adds the support of sha256_hmac and sha256_hkdf.
A first version was sent several months ago just before the
integration of mbedtls. This new version is based on mbedtls.

The first patch of this serie add the support of hkdf
using mbedtls.

Link: https://lore.kernel.org/r/20241219130554.49825-1-philippe.reynes@softathome.com
4 months agoconfigs: sandbox: enable mbedtls
Philippe Reynes [Thu, 19 Dec 2024 13:05:54 +0000 (14:05 +0100)]
configs: sandbox: enable mbedtls

Enable mbedtls to have at least one config with
mbedtls enabled.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
4 months agotest: lib: add test for key derivation
Philippe Reynes [Thu, 19 Dec 2024 13:05:53 +0000 (14:05 +0100)]
test: lib: add test for key derivation

Adds a test for the function sha256_hkdf.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Raymond Mao <raymond.mao@linaro.org>
4 months agolib: mbedtls: sha256: add support of key derivation
Philippe Reynes [Thu, 19 Dec 2024 13:05:52 +0000 (14:05 +0100)]
lib: mbedtls: sha256: add support of key derivation

Adds the support of key derivation using the scheme hkdf.
This scheme is defined in rfc5869.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Raymond Mao <raymond.mao@linaro.org>
4 months agotest: lib: add test for sha256_hmac
Philippe Reynes [Thu, 19 Dec 2024 13:05:51 +0000 (14:05 +0100)]
test: lib: add test for sha256_hmac

Adds a test for the function sha256_hmac

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Raymond Mao <raymond.mao@linaro.org>
4 months agolib: sha256: add feature sha256_hmac
Philippe Reynes [Thu, 19 Dec 2024 13:05:50 +0000 (14:05 +0100)]
lib: sha256: add feature sha256_hmac

Adds the support of the hmac based on sha256.
This implementation is based on rfc2104.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Raymond Mao <raymond.mao@linaro.org>
4 months agolib: sha256: move common function to sha256_common.c
Philippe Reynes [Thu, 19 Dec 2024 13:05:49 +0000 (14:05 +0100)]
lib: sha256: move common function to sha256_common.c

The function sha256_csum_wd is defined in lib/sha256.c
and in lib/mbedtls/sha256.c. To avoid duplicating this
function (and future function), we move this function
to the file lib/sha256_common.c

Reviewed-by: Raymond Mao <raymond.mao@linaro.org>
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
4 months agombedtls: enable support of hkdf
Philippe Reynes [Thu, 19 Dec 2024 13:05:48 +0000 (14:05 +0100)]
mbedtls: enable support of hkdf

Adds the support of key derivation using
the scheme hkdf.

Reviewed-by: Raymond Mao <raymond.mao@linaro.org>
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
4 months agotools: renesas_spkgimage.h: use linux/compiler_attributes.h
Philippe Reynes [Thu, 19 Dec 2024 13:05:47 +0000 (14:05 +0100)]
tools: renesas_spkgimage.h: use linux/compiler_attributes.h

The attribute __packed was defined in the file
tools/renesas_spkgimage.h but this attribute is
already defined in linux/compiler_attributes.h.
So we include <linux/compiler_attributes.h.h>
and do not define __packed.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
4 months agotools: kwbimage.h: use linux/compiler_attributes.h
Philippe Reynes [Thu, 19 Dec 2024 13:05:46 +0000 (14:05 +0100)]
tools: kwbimage.h: use linux/compiler_attributes.h

The attribute __packed was defined in the file
tools/kwbimage.h but this attribute is already
defined in linux/compiler_attributes.h. So we
include <linux/compiler_attributes.h.h> and
do not define __packed.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
4 months agoMerge tag 'efi-2025-04-rc1' of https://source.denx.de/u-boot/custodians/u-boot-efi
Tom Rini [Sat, 18 Jan 2025 16:46:23 +0000 (10:46 -0600)]
Merge tag 'efi-2025-04-rc1' of https://source.denx.de/u-boot/custodians/u-boot-efi

Pull request efi-2025-04-rc1

Documentation:

* Correct the defconfig name in the coolpi documentation

UEFI:

* Carve out the biggest part of the RISC-V and ARM linker scripts for EFI
  binary into a common include.
* Correct the values of SizeOfCode and SizeOfInitializedData in generated
  EFI binaries for RISC-V and ARM.
* Avoid gaps between sections in EFI binaries causing a failure in secure
  boot.
* Makefile: let clean remove capsule_in.capsule*.efi-capsule
* Refactor some of the code used for launching EFI binaries.

4 months agoGitlab: Add DEFAULT_FAST_ARM64_TAG as well
Tom Rini [Sat, 21 Dec 2024 16:45:52 +0000 (10:45 -0600)]
Gitlab: Add DEFAULT_FAST_ARM64_TAG as well

Upon further consideration, we should have both DEFAULT_FAST_ARM64_TAG
and DEFAULT_ARM64_TAG values available. This will allow us to later run
a matrix of some jobs, such as sandbox, on any arm64 host and still keep
the world build to only fast arm64 hosts.

Signed-off-by: Tom Rini <trini@konsulko.com>
4 months agosquashfs: Fix sqfs_inode_size() for xattr related SQFS_LSYMLINK_TYPE
Norbert van Bolhuis [Tue, 24 Dec 2024 00:49:04 +0000 (01:49 +0100)]
squashfs: Fix sqfs_inode_size() for xattr related SQFS_LSYMLINK_TYPE

A squashfs filesystem with extended attributes (xattrs) may have
inodes of type SQFS_LSYMLINK_TYPE. This might cause u-boot to fail to
handle the filesystem since it assumes a SYMLINK_TYPE and LSYMLINK_TYPE
inode are the same size. This is wrong, see:
https://github.com/plougher/squashfs-tools/blob/master/squashfs-tools/read_fs.c#L421

Using the mksquashfs '-no-xattrs' argument is probably best, but the
mksquashfs '-xattrs' argument is the default.
This patch fixes squashfs image handling by making sure parsing the
uncompressed inode_table (with sqfs_find_inode) succeeeds. The only change
needed is correctly determining the size of a SQFS_LSYMLINK_TYPE inode.

Signed-off-by: Norbert van Bolhuis <nvbolhuis@gmail.com>
4 months agoqconfig.py: Update resync message
Tom Rini [Tue, 31 Dec 2024 23:54:12 +0000 (17:54 -0600)]
qconfig.py: Update resync message

When using qconfig.py to resync defconfig files and commit the result,
update it to say:
Resync all defconfig files using qconfig.py

Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
4 months agoMakefile: Make sure all linker input objects exist
Marek Vasut [Sun, 12 Jan 2025 22:32:38 +0000 (23:32 +0100)]
Makefile: Make sure all linker input objects exist

In case the build system builds a directory with empty Makefile,
one which does not contain any obj-y entries, the build fails to
link due to missing built-in.o .

This happens because of this part of scripts/Makefile.build
 81 ifneq ($(strip $(obj-y) $(obj-m) $(obj-) $(subdir-m) $(lib-target)),)
 82 builtin-target := $(obj)/built-in.o
 83 endif
which does not assign builtin-target in case obj-y is empty. The
built-in target is then not built at all, and built-in.o is not
generated by this part of scripts/Makefile.build
 325 ifdef builtin-target
 326 quiet_cmd_link_o_target = AR      $@
 327 # If the list of objects to link is empty, just create an empty built-in.o
 ...
 335 targets += $(builtin-target)
 336 endif # builtin-target
This is the correct behavior.

The final link however expects the built-in.o to exist in every directory
included in the build, even in those where the aforementioned code skipped
generation of built-in.o . Make sure the built-in.o does exist for every
directory used in final link simply by doing touch on every built-in.o used
for the link, which will create empty built-in.o in case any built-in.o is
missing.

A possible alternative fix is the always define the builtin-target
and always generate built-in.o .

Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Tested-by: Quentin Schulz <quentin.schulz@cherry.de>
4 months agoMerge https://source.denx.de/u-boot/custodians/u-boot-sh
Tom Rini [Sat, 18 Jan 2025 14:27:15 +0000 (08:27 -0600)]
Merge https://source.denx.de/u-boot/custodians/u-boot-sh

4 months agoarm64: dts: renesas: Deduplicate extalr_clk bootph-all
Marek Vasut [Sat, 18 Jan 2025 07:15:52 +0000 (08:15 +0100)]
arm64: dts: renesas: Deduplicate extalr_clk bootph-all

Most R-Car Gen3 and Gen4 systems require extalr_clk very early
in the boot process. Currently the extalr_clk { bootph-all } DT
property to indicate this clock should be available early is set
in each SoC U-Boot DT extras file. Deduplicate this assignment to
new common r8a779x-rcar64-u-boot.dtsi file.

Note that R-Car Gen3 R8A77990 E3 and R8A77995 D3 and RZ R8A774C0
do not require extalr_clk early, therefore these SoCs still include
plain r8a779x-u-boot.dtsi in their U-Boot DT extras .

Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
4 months agoarm64: renesas: Rewrite u-boot-elf.srec base address if PIE for R-Car Gen3
Marek Vasut [Sun, 12 Jan 2025 22:47:45 +0000 (23:47 +0100)]
arm64: renesas: Rewrite u-boot-elf.srec base address if PIE for R-Car Gen3

The flash_writer tool and previous recovery tools require the SREC load
address to be 0x5000_0000 . The PIE U-Boot build sets the address to 0x0
instead, so override the address back to make u-boot-elf.srec compatible
with the recovery tools.

Tested-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
4 months agoefi_loader: Make efi_run_image() static
Simon Glass [Thu, 9 Jan 2025 15:02:39 +0000 (08:02 -0700)]
efi_loader: Make efi_run_image() static

This function is not called from outside this file and has no entry in
the header file, so mark it static.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
4 months agoefi_loader: Refactor device and image paths into a function
Simon Glass [Thu, 9 Jan 2025 15:02:38 +0000 (08:02 -0700)]
efi_loader: Refactor device and image paths into a function

Move this code into a function so it can be called from elsewhere.

Note that the recently added network code uses the same
'global variable' approach. It could use a separate clean-up.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
4 months agoMakefile: let clean remove capsule_in.capsule*.efi-capsule
Heinrich Schuchardt [Thu, 16 Jan 2025 14:06:17 +0000 (15:06 +0100)]
Makefile: let clean remove capsule_in.capsule*.efi-capsule

Update the CLEAN_FILES list to remove capsule*.*.efi-capsule.

Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
4 months agoefi_loader: use include in ARM EFI linker script
Heinrich Schuchardt [Thu, 16 Jan 2025 11:39:09 +0000 (12:39 +0100)]
efi_loader: use include in ARM EFI linker script

Use the same include as arm64 for the linker script.

Adjust the 32-bit ARM PE-COFF header accordingly and harmonize it with the
64-bit ARM header.

Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
4 months agoefi_loader: correct SizeOfCode, SizeOfInitializedData
Heinrich Schuchardt [Thu, 16 Jan 2025 11:39:08 +0000 (12:39 +0100)]
efi_loader: correct SizeOfCode, SizeOfInitializedData

The fields SizeOfCode, SizeOfInitializedData, and SizeOfUninitializedData
are define in the PE-COFF specification [1].

* SizeOfCode must match the size of all .text sections.
* SizeOfInitializedData must match the size of all .data sections.
* SizeOfUninitializedData must match the size of all .bss sections.

We only have one .text and one .data section. SizeOfCode and
SizeOfInitializedData have to be calculated as the difference between
the end and the start of the respective section.

As we don't have any .bss sections in the generated EFI binaries.
SizeOfUninitializedData must remain 0.

[1] https://learn.microsoft.com/en-us/windows/win32/debug/pe-format

Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
4 months agoefi_loader: use INCLUDE in EFI linker scripts
Heinrich Schuchardt [Thu, 16 Jan 2025 11:39:07 +0000 (12:39 +0100)]
efi_loader: use INCLUDE in EFI linker scripts

Except for the architecture specific lines ARM and RISC-V can use the same
linker script. Move the common lines to an include.

Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
4 months agoscripts/Makefile.lib: add -L option to LD command for EFI binaries
Heinrich Schuchardt [Thu, 16 Jan 2025 11:39:06 +0000 (12:39 +0100)]
scripts/Makefile.lib: add -L option to LD command for EFI binaries

The linker uses the path specified with -L to search for linker scripts
and for linker script includes.

For out-of-tree builds specify the build directory with -L instead of
the absolute path of the linker script. This allows using an INCLUDE
statement.

Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
4 months agoefi_loader: Fix section alignment on EFI binaries
Ilias Apalodimas [Thu, 16 Jan 2025 11:39:05 +0000 (12:39 +0100)]
efi_loader: Fix section alignment on EFI binaries

When creating EFI binaries, the alignment of the text section isn't
correctly factored in. As a result trying to load signed EFI binaries
throws an error with:

efi_image_region_add() efi_image_region_add: new region already part of another
Image not authenticated

Running the binary through sbverify has a similar warning
sbverify ./lib/efi_loader/helloworld.efi
warning: gap in section table:
    .text   : 0x00001000 - 0x00001c00,
    .data   : 0x00002000 - 0x00002200,
gaps in the section table may result in different checksums
warning: data remaining[7680 vs 12720]: gaps between PE/COFF sections?
.....

If we include the alignment in the text section, the signed binary boots
fine, and the relevant sbverify warning goes away
sbverify ./lib/efi_loader/helloworld.efi
warning: data remaining[8704 vs 12720]: gaps between PE/COFF sections?
.....

We should look into the remaining warning at some point as well
regarding the gaps between PE/COFF sections.

Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
4 months agodoc: coolpi: Fix the defconfig name
Andy Yan [Mon, 13 Jan 2025 10:56:13 +0000 (18:56 +0800)]
doc: coolpi: Fix the defconfig name

The defconfig name should be: coolpi-cm5-genbook-rk3588_defconfig

Signed-off-by: Andy Yan <andyshrk@163.com>
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
4 months agoMerge https://source.denx.de/u-boot/custodians/u-boot-riscv
Tom Rini [Fri, 17 Jan 2025 14:27:40 +0000 (08:27 -0600)]
Merge https://source.denx.de/u-boot/custodians/u-boot-riscv

CI: https://source.denx.de/u-boot/custodians/u-boot-riscv/-/pipelines/24215

- RISC-V: Add "riscv,isa-extensions" and multi-letter extension parsing
  support
- RISC-V: Add default cache line size
- Board: Canaan: Add K230-CanMV support
- Board: VisionFive2: Split out target specific configuration

4 months agoMerge tag 'u-boot-imx-master-20250116' of https://gitlab.denx.de/u-boot/custodians...
Tom Rini [Thu, 16 Jan 2025 15:59:54 +0000 (09:59 -0600)]
Merge tag 'u-boot-imx-master-20250116' of https://gitlab.denx.de/u-boot/custodians/u-boot-imx

CI: https://source.denx.de/u-boot/custodians/u-boot-imx/-/pipelines/24221

- Enable watchdog command on Capricorn board.
- Convert imx8m evk boards to BOOTSTD.
- Enable dynamic mmcdev setting on imx8m evk boards.
- Fix i.MX FDT cooling-device property corruption.
- Add OP-TEE support on Phycore i.MX8M boards.

4 months agoimx: fdt: fix cooling-device property corruption
Sébastien Szymanski [Wed, 15 Jan 2025 15:41:13 +0000 (16:41 +0100)]
imx: fdt: fix cooling-device property corruption

The function disable_thermal_cpu_nodes() corrupts the cooling-device
property.
For example, booting an i.MX93 devices with only one A55 core (IMX93x1)
with the cooling-device property set to:

$ dtc -I dtb foo.dtb | grep cooling-device
cooling-device = <0x08 0xffffffff 0xffffffff 0x09 0xffffffff 0xffffffff>;

Linux shows the following error at boot:

[    1.715189] OF: /thermal-zones/cpu-thermal/cooling-maps/map0: could not find phandle 1083699869
[    1.723977] thermal_sys: Add a cooling_device property with at least one device
[    1.731285] thermal thermal_zone0: binding zone cpu-thermal with cdev thermal-devfreq-0 failed:-2

because the cooling-device property in the device tree passed to the
kernel is

$ dtc -I dtb /sys/firmware/fdt | grep cooling-device
cooling-device = <0x4097f29d 0x00 0xb05aef9d>;

The issue is because the wrong variable type is passed to the function
fdt_setprop() called in the function disable_thermal_cpu_nodes(). With
the variable type fixed, the error at boot is gone and the property is
properly set:

$ dtc -I dtb /sys/firmware/fdt | grep cooling-device
cooling-device = <0x08 0xffffffff 0xffffffff>;

Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
4 months agoconfigs: capricorn_cxg3: Enable CMD_WDT
Alexander Sverdlin [Mon, 13 Jan 2025 21:15:15 +0000 (22:15 +0100)]
configs: capricorn_cxg3: Enable CMD_WDT

Capricorn boards have WDT (WDT_SIEMENS_PMIC) and U-Boot is not configured
to trigger the WDT. Enable the "wdt" command to be able to stop the WDT
from the command line.

Signed-off-by: Alexander Sverdlin <alexander.sverdlin@siemens.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
4 months agodoc: phytec: imx8mm: add OP-TEE documentation
Yannic Moog [Fri, 10 Jan 2025 12:35:38 +0000 (13:35 +0100)]
doc: phytec: imx8mm: add OP-TEE documentation

Add instructions on how to build and package OP-TEE for the
phycore-imx8mm based boards. The build instructions are identical for
phyGATE-Tauri-L and phyBOARD-Polis.
Also fix missig '-' for TF-A build instructions.

Signed-off-by: Yannic Moog <y.moog@phytec.de>
4 months agodoc: phytec: imx8mp: add OP-TEE documentation
Yannic Moog [Fri, 10 Jan 2025 12:35:37 +0000 (13:35 +0100)]
doc: phytec: imx8mp: add OP-TEE documentation

Add documentation for the phyBOARD-Pollux i.MX 8M Plus on OP-TEE
integration.
Also add missing '-' to TF-A build instruction while at it.

Signed-off-by: Yannic Moog <y.moog@phytec.de>
4 months agophycore-imx8mm_defconfig: add optee load address
Yannic Moog [Fri, 10 Jan 2025 12:35:36 +0000 (13:35 +0100)]
phycore-imx8mm_defconfig: add optee load address

The phyCORE i.MX 8M Mini expects 0x56000000 address to load optee.

Signed-off-by: Yannic Moog <y.moog@phytec.de>
4 months agoimx8mm-phygate-tauri-l_defconfig: add optee load address
Yannic Moog [Fri, 10 Jan 2025 12:35:35 +0000 (13:35 +0100)]
imx8mm-phygate-tauri-l_defconfig: add optee load address

The phyGATE-Tauri-L expects 0x56000000 address to load optee.

Signed-off-by: Yannic Moog <y.moog@phytec.de>
4 months agophycore-imx8mp_defconfig: add optee load address
Yannic Moog [Fri, 10 Jan 2025 12:35:34 +0000 (13:35 +0100)]
phycore-imx8mp_defconfig: add optee load address

The phyBOARD-Pollux expects 0x56000000 address to load optee.

Signed-off-by: Yannic Moog <y.moog@phytec.de>
4 months agoarm: imx8m: add OP-TEE node
Yannic Moog [Fri, 10 Jan 2025 12:35:33 +0000 (13:35 +0100)]
arm: imx8m: add OP-TEE node

Add tee node in SoC u-boot device trees. Use a kconfig entry to specify
load and entry addresses for the op-tee image in the respective
defconfig.
Default IMX8M_OPTEE_LOAD_ADDR is supplied. To keep it simple, the same
addresses are used for each SoC as are defined in upstream tf-a
(BL32_BASE) [1].

[1] https://github.com/ARM-software/arm-trusted-firmware/tree/master/plat/imx/imx8m

Signed-off-by: Yannic Moog <y.moog@phytec.de>
4 months agogpio: adp5585: Update compatible string
Peng Fan [Thu, 9 Jan 2025 03:29:18 +0000 (11:29 +0800)]
gpio: adp5585: Update compatible string

i.MX93-11x11-EVK has switched to use upstream device tree, and use
"adi,adp5585". Since i.MX93-11x11-EVK is the only user of this driver,
so it is safe to drop "adp5585".

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx93-11x11-evk: drop duplicated nodes
Peng Fan [Thu, 9 Jan 2025 03:29:17 +0000 (11:29 +0800)]
imx: imx93-11x11-evk: drop duplicated nodes

The upstream dts already has the lpi2c2 subnodes, so drop the duplicated
ones in u-boot dtsi.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mq_evk: Switch to BOOTSTD
Peng Fan [Thu, 9 Jan 2025 03:29:16 +0000 (11:29 +0800)]
imx: imx8mq_evk: Switch to BOOTSTD

Move env to imx8mq_evk.env
Switch to support BOOTSTD with a bsp bootcmd as fallback.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mq_evk: Enable dynamic settings to mmcdev and mmcroot
Peng Fan [Thu, 9 Jan 2025 03:29:15 +0000 (11:29 +0800)]
imx: imx8mq_evk: Enable dynamic settings to mmcdev and mmcroot

Enable dynamic settings to mmcdev and mmcroot for i.MX8MQ-EVK
Since we are here, clean up the including headers

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mq_evk: Drop DECLARE_GLOBAL_DATA_PTR
Peng Fan [Thu, 9 Jan 2025 03:29:14 +0000 (11:29 +0800)]
imx: imx8mq_evk: Drop DECLARE_GLOBAL_DATA_PTR

There is no gd used in the board file, so drop DECLARE_GLOBAL_DATA_PTR.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mq_evk: Cleanup headers
Peng Fan [Thu, 9 Jan 2025 03:29:13 +0000 (11:29 +0800)]
imx: imx8mq_evk: Cleanup headers

Drop unused headers and sort the order.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mm_evk: Switch to BOOTSTD
Peng Fan [Thu, 9 Jan 2025 03:29:12 +0000 (11:29 +0800)]
imx: imx8mm_evk: Switch to BOOTSTD

Move env to imx8mm_evk.env
Switch to support BOOTSTD with a bsp bootcmd as fallback.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mm_evk: Enable dynamic settings to mmcdev and mmcroot
Peng Fan [Thu, 9 Jan 2025 03:29:11 +0000 (11:29 +0800)]
imx: imx8mm_evk: Enable dynamic settings to mmcdev and mmcroot

Enable dynamic settings to mmcdev and mmcroot for i.MX8MM-EVK

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mn_evk: Switch to BOOTSTD
Peng Fan [Thu, 9 Jan 2025 03:29:10 +0000 (11:29 +0800)]
imx: imx8mn_evk: Switch to BOOTSTD

Move env to imx8mn_evk.env
Switch to support BOOTSTD with a bsp bootcmd as fallback.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mn_evk: Enable dynamic settings to mmcdev and mmcroot
Peng Fan [Thu, 9 Jan 2025 03:29:09 +0000 (11:29 +0800)]
imx: imx8mn_evk: Enable dynamic settings to mmcdev and mmcroot

Enable dynamic settings to mmcdev and mmcroot for i.MX8MN-EVK

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mn_evk: Drop DECLARE_GLOBAL_DATA_PTR
Peng Fan [Thu, 9 Jan 2025 03:29:08 +0000 (11:29 +0800)]
imx: imx8mn_evk: Drop DECLARE_GLOBAL_DATA_PTR

There is no users of global data in imx8mn_evk.c, drop it.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mn_evk: Cleanup headers
Peng Fan [Thu, 9 Jan 2025 03:29:07 +0000 (11:29 +0800)]
imx: imx8mn_evk: Cleanup headers

Drop unused headers and sort them

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mp_evk: Switch to BOOTSTD
Peng Fan [Thu, 9 Jan 2025 03:29:06 +0000 (11:29 +0800)]
imx: imx8mp_evk: Switch to BOOTSTD

Move env to imx8mp_evk.env.
Switch to support BOOTSTD with a bsp bootcmd as fallback.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx8mp_evk: Enable dynamic settings to mmcdev and mmcroot
Peng Fan [Thu, 9 Jan 2025 03:29:05 +0000 (11:29 +0800)]
imx: imx8mp_evk: Enable dynamic settings to mmcdev and mmcroot

Enable dynamic settings to mmcdev and mmcroot for i.MX8MP-EVK

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx91_evk: switch to BOOTSTD
Peng Fan [Thu, 9 Jan 2025 03:29:04 +0000 (11:29 +0800)]
imx: imx91_evk: switch to BOOTSTD

Switch to support BOOTSTD with a bsp bootcmd as fallback.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx93_qsb: switch to BOOTSTD
Peng Fan [Thu, 9 Jan 2025 03:29:03 +0000 (11:29 +0800)]
imx: imx93_qsb: switch to BOOTSTD

Switch to support BOOTSTD with a bsp bootcmd as fallback.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agoimx: imx93_evk: switch to BOOTSTD
Peng Fan [Thu, 9 Jan 2025 03:29:02 +0000 (11:29 +0800)]
imx: imx93_evk: switch to BOOTSTD

Switch to support BOOTSTD with a bsp bootcmd as fallback.
Move the env to imx93_evk.env

Signed-off-by: Peng Fan <peng.fan@nxp.com>
4 months agodoc: canaan: Add K230 CanMV board
Junhui Liu [Tue, 14 Jan 2025 16:46:41 +0000 (00:46 +0800)]
doc: canaan: Add K230 CanMV board

Add description of compiling u-boot for K230 CanMV.

Since the vendor's u-boot-spl verifies u-boot header [1], it is
necessary to use the Python script from vendor to add the header to the
u-boot image.

[1] https://github.com/kendryte/k230_sdk/blob/v1.8/src/little/uboot/board/canaan/common/k230_board_common.h#L52

Signed-off-by: Junhui Liu <junhui.liu@pigmoral.tech>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoriscv: canaan: k230_canmv: Add initial support
Junhui Liu [Tue, 14 Jan 2025 16:46:40 +0000 (00:46 +0800)]
riscv: canaan: k230_canmv: Add initial support

Add support for K230 CanMV board with serial console and usb otg
support. It can boot via vendor's u-boot-spl and boot into Linux
via tftp through the onboard RTL8152.

Signed-off-by: Junhui Liu <junhui.liu@pigmoral.tech>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoriscv: cpu: k230: Add support for Canaan Kendryte K230 SoC
Junhui Liu [Tue, 14 Jan 2025 16:46:39 +0000 (00:46 +0800)]
riscv: cpu: k230: Add support for Canaan Kendryte K230 SoC

Add Canaan K230 SoC with sysreset support, running without cache
enabled.

Signed-off-by: Junhui Liu <junhui.liu@pigmoral.tech>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoriscv: dts: canaan: Add basic device tree for K230 CanMV board
Junhui Liu [Tue, 14 Jan 2025 16:46:38 +0000 (00:46 +0800)]
riscv: dts: canaan: Add basic device tree for K230 CanMV board

Add initial dts for K230-CanMV powered by Canaan Kendryte K230 SoC,
which has two RISC-V C908 cores, a big core with vector 1.0 extension
and a small core without vector extension.

This patch is basically comes from Linux Kernel [1] and it assumes
u-boot is running on the big core. Additionally, bootctl and reboot nodes
are added to support sysreset [2] and an clk_dummy node is added to
satisfy dependencies for usb [3].

Currently, u-boot is booted by the vendor's u-boot-spl. To meet the
requirements [4][5] of vendor's u-boot-spl for u-boot, a binman node with
mkimage child node is added here, which will compress u-boot.bin with
gzip and generate an image named "uboot" in the file u-boot-gz.img.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/conor/linux.git/log/?h=k230-basic
[2] https://github.com/kendryte/k230_sdk/blob/v1.8/src/big/rt-smart/kernel/bsp/maix3/board/interdrv/sysctl/sysctl_boot/sysctl_boot.c#L67
[3] https://lore.kernel.org/linux-riscv/tencent_AD84B436C2F31108B66B4739D6E306C5E80A@qq.com/
[4] https://github.com/kendryte/k230_sdk/blob/v1.8/src/little/uboot/board/canaan/common/k230_img.c#L306
[5] https://github.com/kendryte/k230_sdk/blob/v1.8/src/little/uboot/board/canaan/common/k230_img.c#L125

Signed-off-by: Junhui Liu <junhui.liu@pigmoral.tech>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agousb: dwc2: Add support for Canaan K230
Junhui Liu [Tue, 14 Jan 2025 16:46:37 +0000 (00:46 +0800)]
usb: dwc2: Add support for Canaan K230

Canaan Kendryte K230 SoC instantiates a dwc2 v4.30a core. This patch
adds the compatible for it.

Signed-off-by: Junhui Liu <junhui.liu@pigmoral.tech>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoKconfig: Add a default cache line size for RISC-V
Yu-Chien Peter Lin [Fri, 10 Jan 2025 08:53:08 +0000 (16:53 +0800)]
Kconfig: Add a default cache line size for RISC-V

The RISC-V ISA profile RVA23U64 requires extension Zic64b (Cache blocks
must be 64 bytes in size, naturally aligned in the address space).

Some RISC-V platforms do not define the d-cache line size through SYS_CACHE_SHIFT_n.
Set a default value of 64 bytes for such cases.

Signed-off-by: Yu-Chien Peter Lin <peter.lin@sifive.com>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoriscv: Fallback to riscv,isa
Mayuresh Chitale [Mon, 6 Jan 2025 13:04:05 +0000 (13:04 +0000)]
riscv: Fallback to riscv,isa

Update the cpu probing to fallback to "riscv,isa" property if
"riscv,isa-extensions" is not available and modify the riscv CMO code
to use the block size that was probed during cpu setup.

Signed-off-by: Mayuresh Chitale <mchitale@ventanamicro.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoriscv: Enhance extension probing
Mayuresh Chitale [Mon, 6 Jan 2025 13:04:04 +0000 (13:04 +0000)]
riscv: Enhance extension probing

Enhance the existing extension probing mechanism by adding support for
more extensions and probing using the "riscv,isa" property. This patch
is ported from the latest upstream linux.

Signed-off-by: Mayuresh Chitale <mchitale@ventanamicro.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoriscv: dts: starfive: split out visionfive2 target specific configuration
E Shattow [Wed, 1 Jan 2025 06:35:57 +0000 (22:35 -0800)]
riscv: dts: starfive: split out visionfive2 target specific configuration

Split out StarFive VisionFive2 multi-board target specific configuration
into starfive-visionfive2-binman.dtsi in preparation for removal of
jh7110-u-boot and jh7110-common-u-boot in part or whole as sent upstream.

Signed-off-by: E Shattow <e@freeshell.de>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
4 months agoMerge patch series "bootstd: Support recording images"
Tom Rini [Wed, 15 Jan 2025 23:34:26 +0000 (17:34 -0600)]
Merge patch series "bootstd: Support recording images"

Simon Glass <sjg@chromium.org> says:

This series provides a way to keep track of the images used in bootstd,
including the type of each image.

At present this is sort-of handled by struct bootflow but in quite an
ad-hoc way. The structure has become quite large and is hard to query.
Future work will be able to reduce its size.

Ultimately the 'bootflow info' command may change to also show images as
a list, but that is left for later, as this series is already fairly
long. So for now, just introduce the concept and adjust bootstd to use
it, with a simple command to list the images.

This series includes various alist enhancements, to make use of this new
data structure a little easier.

[trini: Drop patch 18 and 19 for now due to size considerations]

Link: https://lore.kernel.org/r/20241115231926.211999-1-sjg@chromium.org
4 months agobootstd: Export bootdev_get_from_blk()
Simon Glass [Fri, 15 Nov 2024 23:19:24 +0000 (16:19 -0700)]
bootstd: Export bootdev_get_from_blk()

Export this function so it can be used from other files.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootstd: Add a simple command to list images
Simon Glass [Fri, 15 Nov 2024 23:19:23 +0000 (16:19 -0700)]
bootstd: Add a simple command to list images

Add a new 'bootstd images' command, which lists the images which have
been loaded.

Update some existing tests to use it. Provide some documentation about
images in general and this command in particular.

Use a more realistic kernel command-line to make the test easier to
follow.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootstd: Update cros bootmeth to record images
Simon Glass [Fri, 15 Nov 2024 23:19:22 +0000 (16:19 -0700)]
bootstd: Update cros bootmeth to record images

Record images loaded by this bootmeth.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agoUpdate bootmeth_alloc_other() to record images
Simon Glass [Fri, 15 Nov 2024 23:19:21 +0000 (16:19 -0700)]
Update bootmeth_alloc_other() to record images

Update this function to add the image to the list.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agoboot: Update pxe bootmeth to record images
Simon Glass [Fri, 15 Nov 2024 23:19:20 +0000 (16:19 -0700)]
boot: Update pxe bootmeth to record images

Record images loaded by this bootmeth.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agoboot: Update extlinux pxe_getfile_func() to include type
Simon Glass [Fri, 15 Nov 2024 23:19:19 +0000 (16:19 -0700)]
boot: Update extlinux pxe_getfile_func() to include type

Add a file-type parameter to this function and update all users. Add a
proper comment to the function which we are here.

This will allow tracking of the file types loaded by the extlinux
bootmeth.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootmeth_efi: Check the filename-allocation in the network path
Simon Glass [Fri, 15 Nov 2024 23:19:18 +0000 (16:19 -0700)]
bootmeth_efi: Check the filename-allocation in the network path

If the filename cannot be set we should give up. Add the missing error
check.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootmeth: Update the read_file() method to include a type
Simon Glass [Fri, 15 Nov 2024 23:19:17 +0000 (16:19 -0700)]
bootmeth: Update the read_file() method to include a type

We want to record the type of each file which is loaded. Add an new
parameter for this, to the read_file() method. Update all users.

Make bootmeth_common_read_file() store information about the image that
is read.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
4 months agobootmeth_efi: Simplify reading files by using the common function
Simon Glass [Fri, 15 Nov 2024 23:19:16 +0000 (16:19 -0700)]
bootmeth_efi: Simplify reading files by using the common function

The efiload_read_file() does similar things to a common function, so
update it to use that instead.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
4 months agoboot: pxe: Drop the duplicate comment on get_pxe_file()
Simon Glass [Fri, 15 Nov 2024 23:19:15 +0000 (16:19 -0700)]
boot: pxe: Drop the duplicate comment on get_pxe_file()

This function is exported, so document it in the header file. Drop the
duplicate comment in the C file.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootstd: Update bootmeth_alloc_file() to record images
Simon Glass [Fri, 15 Nov 2024 23:19:14 +0000 (16:19 -0700)]
bootstd: Update bootmeth_alloc_file() to record images

As a first step to recording images and where they came from, update
this function to do so, since it is used by two bootmeths

Create a helper function in the bootflow system, since recorded
images are always associated with bootflows.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootstd: Maintain a list of images
Simon Glass [Fri, 15 Nov 2024 23:19:13 +0000 (16:19 -0700)]
bootstd: Maintain a list of images

We want to keep track of images which are loaded, or those which could
perhaps be loaded. This will make it easier to manage memory allocation,
as well as permit removal of the EFI set_efi_bootdev() feature.

Add a list of these, attached to the bootflow. For now the list is
empty.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootstd: Move the bootflow list into an alist
Simon Glass [Fri, 15 Nov 2024 23:19:12 +0000 (16:19 -0700)]
bootstd: Move the bootflow list into an alist

Use an alist for this data structure as it is somewhat simpler to
manage. This means that bootstd holds a simple list of bootflow structs
and can drop it at will, without chasing down lists.

Signed-off-by: Simon Glass <sjg@chromium.org>
4 months agobootstd: Drop the bootdev-specific list of bootflows
Simon Glass [Fri, 15 Nov 2024 23:19:11 +0000 (16:19 -0700)]
bootstd: Drop the bootdev-specific list of bootflows

This list is only used by two functions, which can be updated to iterate
through the global list. Take this approach, which allows the bootdev
list to be dropped.

Overall this makes the code slightly more complicated, but will allow
moving the bootflow list into an alist

Signed-off-by: Simon Glass <sjg@chromium.org>