]> git.ipfire.org Git - thirdparty/u-boot.git/log
thirdparty/u-boot.git
8 months agodrivers: misc: gsc: add support for prescaled
Tim Harvey [Wed, 18 Dec 2024 19:45:42 +0000 (11:45 -0800)] 
drivers: misc: gsc: add support for prescaled

Add support for Gateworks System Controller pre-scaled ADC input.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
8 months agodrivers: misc: gsc: add support for fan controller
Tim Harvey [Wed, 18 Dec 2024 19:45:41 +0000 (11:45 -0800)] 
drivers: misc: gsc: add support for fan controller

Add support for Gateworks System Controller fan tach input.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
8 months agospi: mxc_spi: use proper clock for SPI bus
Tim Harvey [Wed, 18 Dec 2024 19:42:24 +0000 (11:42 -0800)] 
spi: mxc_spi: use proper clock for SPI bus

The mxc_get_clock function is around for compatibility with older
drivers that are not clock aware. In this case asking for the clk for
MXC_CSPI_CLK does not take into account there are multiple SPI busses on
modern IMX SoC's and it will return the clock for the first bus which
may not be used or configured.

In the case you are not using the first bus you will not get the proper
clock. Fix this by obtaining the clock rate from the bus clock.

This resolves an invalid SPI clock frequency configuration for SPI2 on a
board where SPI1 is not used.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
8 months agonet: rswitch: Do not register disabled ports as ethernet devices
Marek Vasut [Sat, 21 Dec 2024 21:48:19 +0000 (22:48 +0100)] 
net: rswitch: Do not register disabled ports as ethernet devices

In case an rswitch port is described as disabled in DT, do not
register it as ethernet device in U-Boot. This way, such ports
cannot be accessed from U-Boot command line.

Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
8 months agoARM: renesas: Consistently enable ENV_OVERWRITE on Renesas R-Car
Marek Vasut [Sat, 21 Dec 2024 23:58:24 +0000 (00:58 +0100)] 
ARM: renesas: Consistently enable ENV_OVERWRITE on Renesas R-Car

Move CONFIG_ENV_OVERWRITE=y into commont renesas_rcar.config to make sure
this configuration option is consistently enabled on all of Renesas R-Car
Gen2, Gen3, Gen4. Currently this option is not enabled on Gen4, this fix
corrects that omission.

Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
8 months agoMerge branch '2024-12-21-add-arm64-CI-host-support' into next
Tom Rini [Sat, 21 Dec 2024 18:27:55 +0000 (12:27 -0600)] 
Merge branch '2024-12-21-add-arm64-CI-host-support' into next

First, thanks to Simon Glass and also Linaro, we now have access to a
few fast arm64 host machines in our Gitlab instance, to use as CI
runners. This series finishes the work that I pushed earlier and Simon
had started that enables arm64 hosts to be used for most things now.

The first notable change, especially if you use this on your own Gitlab
instance is that "DEFAULT_TAG" is now unused and we instead have:
- DEFAULT_ALL_TAG:
- DEFAULT_ARM64_TAG:
- DEFAULT_AMD64_TAG:
- DEFAULT_FAST_AMD64_TAG:

This lets us say that some jobs can be run on all runners, because they
are small enough that anything we'd connect to CI is fast enough and it
also does not depend on the underlying host architecture. Next we have
tags for any arm64 host, or any amd64 host. Finally, we have a tag for
fast amd64 hosts. What these last three are for is that we have a few
jobs that need to run on amd64 hosts and so we have to restrict them
there, but we also have now reworked the world build jobs to build
(almost) everything in a single job and on the fast amd64 machines this
is still as quick as the old way was, in practice.

To reach this point, we say that the Xtensa jobs can only run on amd64
hosts. Our targets only work with the binary-only toolchain and so this
is a reasonable limit and we exclude them from the world build jobs. We
also need to deal with ensuring the right toolchain is used regardless
what the host architecture is and that we don't use the host toolchain
by accident. Finally, because some of these changes needed to be worked
out in the linter, fix some of the general warnings that notes as well.

8 months agodocker: Update to latest "Jammy" tag
Tom Rini [Sun, 8 Dec 2024 17:07:32 +0000 (11:07 -0600)] 
docker: Update to latest "Jammy" tag

Bring us up to the current Ubuntu "Jammy" tag.

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agodoc: Expand what we say about building the docker container a bit
Tom Rini [Sun, 8 Dec 2024 17:07:31 +0000 (11:07 -0600)] 
doc: Expand what we say about building the docker container a bit

First, try and be slightly clearer about what "buildx" is with respect
to the docker build process.

Second, now that we build the container for both amd64 and arm64, we
should document how to make a docker "builder" that has multiple nodes.
With this one node should be amd64 and one node arm64, and with
reasonably fast arm64 hardware this will be much quicker than using
QEMU.

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agodocker: Prefer the kernelorg toolchains over "root"
Tom Rini [Sun, 8 Dec 2024 17:07:30 +0000 (11:07 -0600)] 
docker: Prefer the kernelorg toolchains over "root"

We should always look in our downloaded toolchains first and then for
host-provided toolchains.

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agodocker: Ensure we use the cross toolchain for aarch64 on arm64
Tom Rini [Sun, 8 Dec 2024 17:07:29 +0000 (11:07 -0600)] 
docker: Ensure we use the cross toolchain for aarch64 on arm64

We do not want to use the host toolchain for building our platforms in
CI (it is both too old, and would be inconsistent with our CI
practices). To do this we need to set the toolchain-prefix so that we
don't end up guessing "/opt/.../aarch64-linux-aarch64-linux-" as the
prefix.

Link: https://source.denx.de/u-boot/custodians/u-boot-dm/-/issues/32
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agoGitlab: Avoid downloading all artifacts in later stages
Tom Rini [Sun, 8 Dec 2024 17:07:28 +0000 (11:07 -0600)] 
Gitlab: Avoid downloading all artifacts in later stages

In the test.py stage of the build we mark the pytest results as
artifacts to save, so that they can be used for reports. This however
leads to all of the artifacts being downloaded (and then not used) in
later stages. Optimize this out by using an empty list of dependencies
here (which is the keyword for which artifacts are needed).

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agoGitLab: Add a default workflow stanza
Tom Rini [Sun, 8 Dec 2024 17:07:27 +0000 (11:07 -0600)] 
GitLab: Add a default workflow stanza

When validating our current pipeline, a warning is produced about a lack
of a default workflow. For how we use it, we can add a simple default of
"always".

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agoGitLab CI: Make wider use of tags
Tom Rini [Sun, 8 Dec 2024 17:07:26 +0000 (11:07 -0600)] 
GitLab CI: Make wider use of tags

First, introduce DEFAULT_ALL_TAG, DEFAULT_ARM64_TAG, DEFAULT_AMD64_TAG
and DEFAULT_FAST_AMD64_TAG and remove the previous DEFAULT_TAG (as
anyone making use of that will need to adjust their jobs). This allows
us to say that some jobs can run on amd64 or arm64 hosts under the all
tag, while some jobs must run on amd64 (the Xtensa jobs due to
binary-only toolchains and sandbox for now) Then we rework the world
build stage to only run on our very fast amd64 hosts, or our arm64 hosts
(which are also very fast). This should result in a similar overall
build time but also a much more consistent overall build time as we
won't have the two big world jobs possibly run on our slower build
nodes.

Signed-off-by: Tom Rini <trini@konsulko.com>
8 months agodocker: Use ${TCVER} more widely
Tom Rini [Sun, 8 Dec 2024 17:07:25 +0000 (11:07 -0600)] 
docker: Use ${TCVER} more widely

Remove the rest of the places where we hard-code the version of the
toolchain we're using.

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agoDocker/CI: Only test Xtensa on amd64 hosts
Tom Rini [Sun, 8 Dec 2024 17:07:24 +0000 (11:07 -0600)] 
Docker/CI: Only test Xtensa on amd64 hosts

The xtensa architecture is interesting in that the platforms we support
are only valid on the binary-only toolchains as the DC233C instruction
set requires those toolchains (and not the FSF instruction set). Only
install the binary toolchain on amd64 hosts and only run the tests on
them as well.

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agoMerge branch 'master' of https://source.denx.de/u-boot/custodians/u-boot-sh
Tom Rini [Sat, 21 Dec 2024 15:45:30 +0000 (09:45 -0600)] 
Merge branch 'master' of https://source.denx.de/u-boot/custodians/u-boot-sh

- Two rswitch fixes and a clock fix

8 months agonet: rswitch: Add missing cache invalidate of TX descriptor
Marek Vasut [Thu, 19 Dec 2024 10:52:46 +0000 (11:52 +0100)] 
net: rswitch: Add missing cache invalidate of TX descriptor

TFTP transfers of large files, for example 128 MiB, can sporadically
get stuck and the transfer slows down considerably.

This happens because the TX DMA descriptor in DRAM becomes out of sync
with the view of the TX DMA descriptor content from the CPU side, which
is viewed through the CPU caches. In order to guarantee these two views
are consistent, the cache over TX DMA descriptor that has possibly been
written by the rswitch hardware must first be invalidated, only then can
the descriptor be cleared and updated by the CPU, and finally the cache
over that area must be flushed back into DRAM to make sure the rswitch
hardware has consistent view of the updated descriptor content.

The very first invalidation operation was missing, which led to sporadic
corruption of the TX DMA descriptor. Fix it, add the missing invalidation
operation.

Reported-by: Enric Balletbo i Serra <eballetb@redhat.com>
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
Tested-by: Enric Balletbo i Serra <eballetb@redhat.com>
8 months agonet: renesas: Enable TFTP_TSIZE on all Renesas hardware
Marek Vasut [Wed, 18 Dec 2024 19:41:28 +0000 (20:41 +0100)] 
net: renesas: Enable TFTP_TSIZE on all Renesas hardware

TFTP transfer size can be used to re-size the TFTP progress bar on
single line based on the server reported file size. Enable it by
default for Renesas hardware to avoid long scrolling walls of '#'
character during long TFTP transfers.

Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
8 months agoclk: renesas: r8a779h0: Drop CLK_PLL2_DIV2 to clarify ZCn clocks
Geert Uytterhoeven [Fri, 20 Dec 2024 00:04:08 +0000 (01:04 +0100)] 
clk: renesas: r8a779h0: Drop CLK_PLL2_DIV2 to clarify ZCn clocks

Early revisions of the R-Car V4M Series Hardware User’s Manual
contained an incorrect formula for the CPU core clocks:

    ZCnφ = (PLL2VCO x 1/2) x mult/32

Dang-san fixed this by using CLK_PLL2_DIV2 instead of CLK_PLL2 as the
parent clock.

In Rev.0.70 of the documentation, the formula was corrected to:

    ZCnφ = (PLL2VCO x 1/4) x mult/32

As the CPG Block Diagram now shows a separate 1/4 post-divider for PLL2,
the use of CLK_PLL2_DIV2 is a recurring source of confusion.  Hence get
rid of CLK_PLL2_DIV2, and include the proper 1/4 post-divider in the
invocation of the DEF_GEN4_Z() macro, like is done on other R-Car Gen4
(and Gen3) SoCs.

Ported from Linux commit
92850bed9d4d ("clk: renesas: r8a779h0: Drop CLK_PLL2_DIV2 to clarify ZCn clocks")

Reported-by: Vinh Nguyen <vinh.nguyen.xz@renesas.com>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/0d2789cac2bf306145fe0bbf269c2da5942bb68f.1728377724.git.geert+renesas@glider.be
Signed-off-by: Marek Vasut <marek.vasut+renesas@mailbox.org>
8 months agoMerge tag 'xilinx-for-v2025.01-rc5-v2' of https://source.denx.de/u-boot/custodians...
Tom Rini [Fri, 20 Dec 2024 14:26:16 +0000 (08:26 -0600)] 
Merge tag 'xilinx-for-v2025.01-rc5-v2' of https://source.denx.de/u-boot/custodians/u-boot-microblaze

AMD/Xilinx changes for v2025.01-rc5 - second

- Fix find_ram_top() return to support configs > 4GB

8 months agocommon: memtop: Fix the return type for find_ram_top
Venkatesh Yadav Abbarapu [Fri, 20 Dec 2024 03:07:42 +0000 (08:37 +0530)] 
common: memtop: Fix the return type for find_ram_top

As the return type is "int" for find_ram_top() function and
returning the "base" which is of phys_addr_t is breaking when the
"base" address is 64-bit. So fix this by updating the return type as
phys_addr_t.

Signed-off-by: Venkatesh Yadav Abbarapu <venkatesh.abbarapu@amd.com>
Reviewed-by: Michal Simek <michal.simek@amd.com>
Acked-by: Sughosh Ganu <sughosh.ganu@linaro.org>
Link: https://lore.kernel.org/r/20241220030742.1745984-1-venkatesh.abbarapu@amd.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
8 months agoMakefile: Match the full path to ccache for filtering
Manorit Chawdhry [Tue, 10 Dec 2024 09:16:45 +0000 (14:46 +0530)] 
Makefile: Match the full path to ccache for filtering

One can use ccache by keeping ccache in PATH or by providing the full
path to ccache as well. Providing the full path to ccache fails as the
current regex tries to look for ccache being the initial token during
filtering.

Do a greedy search to remove anything before ccache for regex matching.

Fixes: 04b1d84221d5 ("Makefile: fix empty MK_ARCH when using ccache")
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
8 months agoimx: Fix usable memory ranges for imx8m SOCs
Ilias Apalodimas [Wed, 18 Dec 2024 09:00:06 +0000 (11:00 +0200)] 
imx: Fix usable memory ranges for imx8m SOCs

commit e27bddff4b97 ("imx8m: Restrict usable memory to space below 4G boundary")
tried to adjust the usable memory limits on a 4GB boundary.

ram_top is described as 'top address of RAM used by U-Boot' and we want
to preserve that. This is defined as a phys_addr_t and unfortunately
its size differs across architectures. This has lead us to a weird
state where 32bit boards define it 'SZ_4GB - 1' and 64bit boards as
'SZ_4GB' unless it was otherwise defined.

With some recent LMB changes and specifically
commit 1a48b0be93d4 ("lmb: prohibit allocations above ram_top even from same bank")
the board fails to boot properly although the commit above is correct
since it's making sure that no memory above ram_top is usable -- but
added to our memory map so EFI can hand it over to the booted OS.

The reason for that is that during the LMB init we add all usable memory
in lmb_add_memory(). In that function any memory above ram_top gets added
as 'reserved' for LMB. With the current values tha's set to 0xFFFF_FFFF
for this board. Later LMB is trying to protect the memory area U-Boot lives
in with lmb_reserve_common(). The latter fails though since it tries to
add U-Boot top (which is 0xFFFF_FFFF as well) to U-Boot 'bottom'. This call
will fail since 1 byte of that memory range is already marked as 'reserved'.

Since we are close to the release, LMB seems to assume that the address
is rounded up and is the 'next address' and so does parsing and adding
memory ranges from DT files, bump the ram_top of the board by 1byte.

In the long run we should change all of the above and have 32b and 64b
platforms define ram_top identically.

Add a Fixes tag although the commit is correct, so people can figure out
the broken scenarios in the future.

Suggested-by: Sughosh Ganu <sughosh.ganu@linaro.org>
Fixes: commit 1a48b0be93d4 ("lmb: prohibit allocations above ram_top even from same bank")
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Frieder Schrempf <frieder.schrempf@kontron.de>
Reported-by: João Paulo Gonçalves <jpaulo.silvagoncalves@gmail.com>
Closes: https://lore.kernel.org/all/20241216114231.qpfwug3zfqkxn3d5@joaog-nb.corp.toradex.com/
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Fixes: 74f88b72219e ("ARM: imx: imx8m: Fix board_get_usable_ram_top()")
8 months agoRevert "configs: JH7110: enable EFI_LOADER_BOUNCE_BUFFER"
E Shattow [Mon, 16 Dec 2024 06:32:12 +0000 (22:32 -0800)] 
Revert "configs: JH7110: enable EFI_LOADER_BOUNCE_BUFFER"

("Enable EFI_LOADER_BOUNCE_BUFFER") is not the correct fix for the
problem it describes.

The change of memory addressing leading to side-effects in commit
22f2c9ed9f53 ("efi: memory: use the lmb API's for allocating and
freeing memory") is remedied by commit 1a48b0be93d4 ("lmb: prohibit
allocations above ram_top even from same bank").

This reverts commit 9c792ab336f7146e8d49bca6d9d093d2392ded5d.

8 months agonet: lwip: check if network device is available in do_dhcp
Heinrich Schuchardt [Thu, 5 Dec 2024 19:17:36 +0000 (20:17 +0100)] 
net: lwip: check if network device is available in do_dhcp

eth_get_dev() returns NULL if no network device is available.
Not checking the return value leads to a crash when the device
pointer is dereferenced.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
8 months agonet: lwip: do not return CMD_RET_USAGE if no interface
Heinrich Schuchardt [Thu, 5 Dec 2024 19:32:53 +0000 (20:32 +0100)] 
net: lwip: do not return CMD_RET_USAGE if no interface

If the dns command cannot find a network interface, we should return
CMD_RETFAIURE and not -1 (CMD_RET_USAGE).

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
8 months agobinman: Regenerate tools/binman/entries.rst
Tom Rini [Thu, 19 Dec 2024 16:19:57 +0000 (10:19 -0600)] 
binman: Regenerate tools/binman/entries.rst

There have been a few changes to the areas that this file documents
without having regenerated the file. Do so now.

Signed-off-by: Tom Rini <trini@konsulko.com>
8 months agoMerge tag 'u-boot-dfu-20241219' of https://source.denx.de/u-boot/custodians/u-boot-dfu
Tom Rini [Thu, 19 Dec 2024 16:19:29 +0000 (10:19 -0600)] 
Merge tag 'u-boot-dfu-20241219' of https://source.denx.de/u-boot/custodians/u-boot-dfu

CI: https://source.denx.de/u-boot/custodians/u-boot-dfu/-/pipelines/23951

Android:
- Fix kcmdline_extra support when parsing boot image
- Fix memory leak when after bootargs concatenation
- Fix length calculation when merging bootargs, cmdline and kcmdline

8 months agoMerge branch '2024-12-19-assorted-tooling-updates' into next
Tom Rini [Thu, 19 Dec 2024 15:10:48 +0000 (09:10 -0600)] 
Merge branch '2024-12-19-assorted-tooling-updates' into next

This brings in assortment of updates to our python tooling, from Paul
HENRYS <paul.henrys_ext@softathome.com>

8 months agotools: binman: etype: fit: Append DTB directory to the list of input directories
Paul HENRYS [Mon, 25 Nov 2024 18:16:54 +0000 (19:16 +0100)] 
tools: binman: etype: fit: Append DTB directory to the list of input directories

When specifying a directory containing DTBs with 'fit,fdt-list-dir', it can be
handy not to have to also specify this directory to the input directories of
binman with '-I' option and use the method tools.append_input_dirs() append it.
This avoids to have to specify the DTB directory in both the device tree
provided to binman and through '-I' option to binman.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agotools: u_boot_pylib: Allow to append input directories to indir
Paul HENRYS [Mon, 25 Nov 2024 18:16:53 +0000 (19:16 +0100)] 
tools: u_boot_pylib: Allow to append input directories to indir

append_input_dirs() can be used to append a list of input directories to indir
global list.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agotools: binman: etype: Allow to replace 'NAME' in node names
Paul HENRYS [Mon, 25 Nov 2024 17:54:21 +0000 (18:54 +0100)] 
tools: binman: etype: Allow to replace 'NAME' in node names

This change allows to replace both 'SEQ' and 'NAME' keywords by respectively a
sequence number and the name of the FDT to provide more flexibility in the node
name for the device trees included in the FIT.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agotools: binman: Pass a list of input directory to EnsureCompiled()
Paul HENRYS [Mon, 25 Nov 2024 17:51:11 +0000 (18:51 +0100)] 
tools: binman: Pass a list of input directory to EnsureCompiled()

Input directories can be passed to binman using the '-I' option and those
input directories are now also passed to 'dtc' when run by binman.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agotools: dtoc: Allow passing optional input directories
Paul HENRYS [Mon, 25 Nov 2024 17:51:10 +0000 (18:51 +0100)] 
tools: dtoc: Allow passing optional input directories

An optional list of input directories can be passed to EnsureCompiled() which
allows to pass a list of directories where input files can be searched.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agotools: binman: Add tests for FIT with data encrypted by mkimage
Paul HENRYS [Mon, 25 Nov 2024 17:47:17 +0000 (18:47 +0100)] 
tools: binman: Add tests for FIT with data encrypted by mkimage

Test the property 'fit,encrypt' to encrypt FIT data.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
8 months agotools: binman: add 'fit, encrypt' property to pass keys directory to mkimage
Paul HENRYS [Mon, 25 Nov 2024 17:47:16 +0000 (18:47 +0100)] 
tools: binman: add 'fit, encrypt' property to pass keys directory to mkimage

mkimage can be used for both signing the FIT or encrypt its content and the
option '-k' can be used to pass a directory where both signing and encryption
keys can be retrieved. Adding 'fit,encrypt' property to the 'fit' node, leads to
try to find keys directory among binman include directories.
_get_priv_keys_dir() is renamed as _get_keys_dir() and adapted to support both
signing and encryption nodes in the FIT.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agoaes: Allow to store randomly generated IV in the FIT
Paul HENRYS [Mon, 25 Nov 2024 17:47:15 +0000 (18:47 +0100)] 
aes: Allow to store randomly generated IV in the FIT

When the initialisation vector is randomly generated, its value shall be
stored in the FIT together with the encrypted data. The changes allow to
store the IV in the FIT also in the case where the key is not stored in
the DTB but retrieved somewhere else at runtime.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agopowerpc: Support using upstream devicetrees
J. Neuschäfer [Thu, 5 Dec 2024 17:22:00 +0000 (18:22 +0100)] 
powerpc: Support using upstream devicetrees

For new PowerPC developments, it will be useful to borrow devicetrees
from Linux. This patch makes it possible.

Signed-off-by: J. Neuschäfer <j.ne@posteo.net>
Reviewed-by: Sumit Garg <sumit.garg@linaro.org>
8 months agophy: don't spam console if phys property is absent in device node
Roger Quadros [Tue, 3 Dec 2024 20:57:36 +0000 (22:57 +0200)] 
phy: don't spam console if phys property is absent in device node

In generic_phy_get_bulk(), if 'phys' property is absent in
the device node then it is not an error condition.

Change print message verbosity to debug to avoid spamming
console in such cases.

Signed-off-by: Roger Quadros <rogerq@kernel.org>
Reviewed-by: Siddharth Vadapalli <s-vadapalli@ti.com>
8 months agoverdin-am62: fix FASTBOOT_BUF_ADDR and FASTBOOT_BUF_SIZE
Hiago De Franco [Tue, 3 Dec 2024 12:33:02 +0000 (09:33 -0300)] 
verdin-am62: fix FASTBOOT_BUF_ADDR and FASTBOOT_BUF_SIZE

The current value of CONFIG_FASTBOOT_BUF_ADDR is incorrect for hardware
with 512MB and 1GB of RAM. The RAM address ranges for these devices are:

- 512MB:
  - Start: 0x80000000
  - End:   0xA0000000
- 1GB:
  - Start: 0x80000000
  - End:   0xC0000000

The current buffer address makes fastboot downloads work only on
hardware with 2GB of RAM.

To fix this issue:
1. Lower CONFIG_FASTBOOT_BUF_ADDR to 0x88200000. This makes the buffer
   to be equal ot CONFIG_SYS_LOAD_ADDR, which is a safe address to
   download files using fastboot.

2. Reduce the size of the download buffer to 128MB, avoiding overlaps
   with RAMDISK_ADDR_R. This is a good enough value for individual
   downloads.

These changes make fastboot downloads work with 512MB and 1GB devices.

Fixes: defe30a78b76 ("verdin-am62: add DFU, USB and UUU fastboot support")
Signed-off-by: Hiago De Franco <hiago.franco@toradex.com>
Acked-by: Francesco Dolcini <francesco.dolcini@toradex.com>
8 months agoarch: x86: lib: Add support of legacy 64-bit entry point at 0x200
Paul HENRYS [Wed, 20 Nov 2024 10:39:43 +0000 (11:39 +0100)] 
arch: x86: lib: Add support of legacy 64-bit entry point at 0x200

Support of legacy 64-bit entry point was already present when booting a bzimage
with 'zboot' but not supported with 'bootm' when the x86_64 Linux kernel is
embedded in a FIT image.

Signed-off-by: Paul HENRYS <paul.henrys_ext@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agoMerge patch series "sandbox: repair compile and run-time for OF_EMBED case"
Tom Rini [Wed, 18 Dec 2024 18:36:38 +0000 (12:36 -0600)] 
Merge patch series "sandbox: repair compile and run-time for OF_EMBED case"

Evgeny Bachinin <EABachinin@salutedevices.com> says:

This patch-set repairs ability to use sandbox with CONFIG_OF_EMBED=y.

For now, to use OF_EMBED, the following must be done
1) sandbox64_defconfig should have:
  ```
  -CONFIG_OF_LIVE=y
  +CONFIG_OF_EMBED=y
  ```

2) On sandbox when CONFIG_OF_EMBED=y, the u-boot process can't start
due to:
  ```
  Bloblist at b000 not found (err=-2)
  initcall failed at call 000000000011829c (err=-2: No such file or \
      directory)
  ### ERROR ### Please RESET the board ###
  ```
  So, it is natural desire to disable CONFIG_BLOBLIST just to test
sandbox with OF_EMBED=y (disable it one way or another):
  ```
    config SANDBOX
  - select BLOBLIST
  +     select BLOBLIST if SOME_NON_EXISTING_OPTION
  ```

3) As a result, having such changes (CONFIG_OF_EMBED=y &&
CONFIG_BLOBLIST=n) leads to the link & run-time errors, being fixed
by this patch series.

Link: https://lore.kernel.org/r/20241202-sandbox_repair_of_embed-v1-0-05aff4b0ccf7@salutedevices.com
[trini: The final patch is dropped as no longer relevant with Simon's
        rework to that function which is now applied.]

8 months agoMerge patch series "fdt: Correct condition for receiving bloblist"
Tom Rini [Wed, 18 Dec 2024 18:34:58 +0000 (12:34 -0600)] 
Merge patch series "fdt: Correct condition for receiving bloblist"

This series is from Simon Glass <sjg@chromium.org>.

It first corrects the test for checking if a bloblist could have come
from TPL and so we check for a device tree, and then changes the
argument order for board_fdt_blob_setup() to be more inline with our
usual argument ordering.

Link: https://lore.kernel.org/r/20241102174944.412088-1-sjg@chromium.org
8 months agotest: sandbox: fix link error with do_ut_bootm if BLOBLIST=n
Evgeny Bachinin [Mon, 2 Dec 2024 13:45:25 +0000 (16:45 +0300)] 
test: sandbox: fix link error with do_ut_bootm if BLOBLIST=n

Having CONFIG_OF_EMBED=y && CONFIG_BLOBLIST=n leads to the link
error:
```
ld: /tmp/ccwtRVty.ltrans28.ltrans.o:(.data.rel+0x4e8): undefined \
                                            reference to `do_ut_bootm'
```

Fixes: fe158657a5b ("test: inconsistent bootm tests")
Signed-off-by: Evgeny Bachinin <EABachinin@salutedevices.com>
8 months agotest: sandbox: fix link error with do_ut_bloblist if BLOBLIST=n
Evgeny Bachinin [Mon, 2 Dec 2024 13:45:24 +0000 (16:45 +0300)] 
test: sandbox: fix link error with do_ut_bloblist if BLOBLIST=n

Having CONFIG_OF_EMBED=y && CONFIG_BLOBLIST=n leads to the link
error:
```
ld: /tmp/ccwtRVty.ltrans28.ltrans.o:(.data.rel+0x4b0): undefined \
                                         reference to `do_ut_bloblist'
```

Fixes: 6ea5df39e8d ("test: Only enable bloblist test when supported")
Signed-off-by: Evgeny Bachinin <EABachinin@salutedevices.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agotest: sandbox: fix invalid_use_of_IF_ENABLED_INT if BLOBLIST=n
Evgeny Bachinin [Mon, 2 Dec 2024 13:45:23 +0000 (16:45 +0300)] 
test: sandbox: fix invalid_use_of_IF_ENABLED_INT if BLOBLIST=n

Having CONFIG_OF_EMBED=y && CONFIG_BLOBLIST=n leads to the link
error:
```
ld: /tmp/ccRVty.ltrans40.ltrans.o: in function `lib_test_is_enabled':
test/lib/kconfig.c:24: undefined reference to \
                                       `invalid_use_of_IF_ENABLED_INT'
ld: test/lib/kconfig.c:26: undefined reference to \
                                `invalid_use_of_CONFIG_IF_ENABLED_INT'
```

Fixes: 29784d62ede ("test: Add some tests for kconfig.h")
Signed-off-by: Evgeny Bachinin <EABachinin@salutedevices.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agosandbox: fix bloblist_finish() linker error if BLOBLIST=n
Evgeny Bachinin [Mon, 2 Dec 2024 13:45:22 +0000 (16:45 +0300)] 
sandbox: fix bloblist_finish() linker error if BLOBLIST=n

Having CONFIG_OF_EMBED=y && CONFIG_BLOBLIST=n leads to the link
error:
```
ld: /tmp/ccwtRVty.ltrans0.ltrans.o: in function `state_uninit':
arch/sandbox/cpu/state.c:508: undefined reference to `bloblist_finish'
```

Fixes: 1c52fcca72b ("sandbox: Write out bloblist when exiting")
Signed-off-by: Evgeny Bachinin <EABachinin@salutedevices.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 months agofdt: Swap the signature for board_fdt_blob_setup()
Simon Glass [Sat, 2 Nov 2024 17:49:42 +0000 (11:49 -0600)] 
fdt: Swap the signature for board_fdt_blob_setup()

This returns a devicetree and updates a parameter with an error code.
Swap it, since this fits better with the way U-Boot normally works. It
also (more easily) allows leaving the existing pointer unchanged.

No yaks were harmed in this change, but there is a very small code-size
reduction.

For sifive, the OF_BOARD option must be set for the function to be
called, so there is no point in checking it again. Also OF_SEPARATE is
defined always.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Matthias Brugger <mbrugger@suse.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
[trini: Update total_compute]
Signed-off-by: Tom Rini <trini@konsulko.com>
8 months agofdt: Correct condition for receiving bloblist
Simon Glass [Sat, 2 Nov 2024 17:49:41 +0000 (11:49 -0600)] 
fdt: Correct condition for receiving bloblist

The condition for receiving a bloblist from TPL is reversed. This was
only noticed are the other fixes landed. Fix it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Matthias Brugger <mbrugger@suse.com>
8 months agoMerge branch 'next' of https://source.denx.de/u-boot/custodians/u-boot-riscv into...
Tom Rini [Wed, 18 Dec 2024 14:01:48 +0000 (08:01 -0600)] 
Merge branch 'next' of https://source.denx.de/u-boot/custodians/u-boot-riscv into next

CI: https://source.denx.de/u-boot/custodians/u-boot-riscv/-/pipelines/23926

- Board: Support LicheeRV Nano
- Board: Support bananapi-f3
- Board: Switch to OF_UPSTREAM for StarFive JH7110
- Board: Add sdhci driver for TH1520 SoC

8 months agoMerge tag 'u-boot-stm32-20241218' of https://source.denx.de/u-boot/custodians/u-boot-stm
Tom Rini [Wed, 18 Dec 2024 14:01:17 +0000 (08:01 -0600)] 
Merge tag 'u-boot-stm32-20241218' of https://source.denx.de/u-boot/custodians/u-boot-stm

CI: https://source.denx.de/u-boot/custodians/u-boot-stm/-/pipelines/23931

- Restore SPL boot from sdcard for STM32MP1 platforms
- Fix STACK_SIZE for STM32 MCU's board

8 months agoboot: android: rework bootargs concatenation
Nicolas Belin [Tue, 17 Dec 2024 13:29:10 +0000 (14:29 +0100)] 
boot: android: rework bootargs concatenation

Rework the bootargs concatenation allocating more accurately
the length that is needed.
Do not forget an extra byte for the null termination byte as,
in some cases, the allocation was 1 byte short.

Fixes: 86f4695b ("image: Fix Android boot image support")
Signed-off-by: Nicolas Belin <nbelin@baylibre.com>
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Link: https://lore.kernel.org/r/20241217-fix-bootargs-concatenation-v2-3-b2fd7cf4e130@baylibre.com
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
8 months agoboot: android: free newbootargs when done
Nicolas Belin [Tue, 17 Dec 2024 13:29:09 +0000 (14:29 +0100)] 
boot: android: free newbootargs when done

Free newbootargs when the concatenation is done and bootargs env
is set.

Fixes: 86f4695b ("image: Fix Android boot image support")
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Signed-off-by: Nicolas Belin <nbelin@baylibre.com>
Link: https://lore.kernel.org/r/20241217-fix-bootargs-concatenation-v2-2-b2fd7cf4e130@baylibre.com
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
8 months agoboot: android: fix extra command line support
Nicolas Belin [Tue, 17 Dec 2024 13:29:08 +0000 (14:29 +0100)] 
boot: android: fix extra command line support

Check that the value at the address kcmdline_extra is not 0
instead of checking the address value itself keeping it
consistent with what is done for kcmdline.

Fixes: b36b227b ("android: boot: support extra command line")
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Signed-off-by: Nicolas Belin <nbelin@baylibre.com>
Link: https://lore.kernel.org/r/20241217-fix-bootargs-concatenation-v2-1-b2fd7cf4e130@baylibre.com
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
8 months agoconfigs: stm32mp1: Restore boot SPL from sdcard for Engicam i.Core STM32MP1 C.TOUCH 2.0
Patrice Chotard [Mon, 16 Dec 2024 10:22:19 +0000 (11:22 +0100)] 
configs: stm32mp1: Restore boot SPL from sdcard for Engicam i.Core STM32MP1 C.TOUCH 2.0

Restore boot SPL from sdcard for Engicam i.Core STM32MP1 C.TOUCH 2.0.

Fixes: 2a00d73d081a ("spl: mmc: Try to clean up raw-mode options")
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
8 months agoconfigs: stm32mp1: Restore boot SPL from sdcard for Engicam MicroGEA STM32MP1 MicroDe...
Patrice Chotard [Mon, 16 Dec 2024 10:22:18 +0000 (11:22 +0100)] 
configs: stm32mp1: Restore boot SPL from sdcard for Engicam MicroGEA STM32MP1 MicroDev 2.0 7" OF

Restore boot SPL from sdcard for Engicam MicroGEA STM32MP1 MicroDev 2.0 7" OF

Fixes: 2a00d73d081a ("spl: mmc: Try to clean up raw-mode options")
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
8 months agoconfigs: stm32mp1: Restore boot SPL from sdcard for Engicam i.Core STM32MP1 EDIMM2.2
Patrice Chotard [Mon, 16 Dec 2024 10:22:17 +0000 (11:22 +0100)] 
configs: stm32mp1: Restore boot SPL from sdcard for Engicam i.Core STM32MP1 EDIMM2.2

Restore boot SPL from sdcard for Engicam i.Core STM32MP1 EDIMM2.2.

Fixes: 2a00d73d081a ("spl: mmc: Try to clean up raw-mode options")
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
8 months agoconfigs: stm32mp1: Restore boot SPL from sdcard for Engicam i.Core STM32MP1 C.TOUCH 2.0
Patrice Chotard [Mon, 16 Dec 2024 10:22:16 +0000 (11:22 +0100)] 
configs: stm32mp1: Restore boot SPL from sdcard for Engicam i.Core STM32MP1 C.TOUCH 2.0

Restore boot SPL from sdcard for Engicam i.Core STM32MP1 C.TOUCH 2.0.

Fixes: 2a00d73d081a ("spl: mmc: Try to clean up raw-mode options")
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
8 months agoconfigs: stm32mp1: Restore boot SPL from sdcard for stm32mp15
Patrice Chotard [Mon, 16 Dec 2024 10:22:15 +0000 (11:22 +0100)] 
configs: stm32mp1: Restore boot SPL from sdcard for stm32mp15

Restore boot SPL from sdcard for STM32MP1 platforms.

Fixes: 2a00d73d081a ("spl: mmc: Try to clean up raw-mode options")
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
8 months agoARM: dts: stm32: Deduplicate CONFIG_OF_SPL_REMOVE_PROPS on DH STM32MP15xx DHSOM
Marek Vasut [Sun, 15 Dec 2024 23:29:15 +0000 (00:29 +0100)] 
ARM: dts: stm32: Deduplicate CONFIG_OF_SPL_REMOVE_PROPS on DH STM32MP15xx DHSOM

The content of CONFIG_OF_SPL_REMOVE_PROPS is the same in both
STM32MP15xx DHCOM and DHCOR defconfigs, deduplicate the content
into stm32mp15_dhsom.config .

Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
8 months agoARM: dts: stm32: Drop access-controllers from SPL DT on DH STM32MP15xx DHSOM
Marek Vasut [Sun, 15 Dec 2024 23:29:14 +0000 (00:29 +0100)] 
ARM: dts: stm32: Drop access-controllers from SPL DT on DH STM32MP15xx DHSOM

The access-controllers DT property is not useful in STM32MP15xx SPL,
remove it to reduce SPL control DT size. No functional change.

Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
8 months agoARM: dts: stm32: Reinstate SPL_SYS_MMCSD_RAW_MODE on DH STM32MP15xx DHSOM
Marek Vasut [Sun, 15 Dec 2024 23:29:13 +0000 (00:29 +0100)] 
ARM: dts: stm32: Reinstate SPL_SYS_MMCSD_RAW_MODE on DH STM32MP15xx DHSOM

Commit 2a00d73d081a ("spl: mmc: Try to clean up raw-mode options")
broke booting from SD card on STM32MP15xx , reinstate raw mode SD
boot configuration options and select the correct raw mode partition
for STM32MP15xx to fix SD boot on STM32MP15xx DHSOM.

Fixes: 2a00d73d081a ("spl: mmc: Try to clean up raw-mode options")
Reported-by: Patrice Chotard <patrice.chotard@foss.st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
8 months agoARM: dts: stm32: Reinstate missing root oscillators on DH STM32MP15xx DHCOR
Marek Vasut [Sun, 15 Dec 2024 23:29:12 +0000 (00:29 +0100)] 
ARM: dts: stm32: Reinstate missing root oscillators on DH STM32MP15xx DHCOR

The root oscillators reference used to be in rcc node since
3d15245502c4 ("ARM: dts: stm32mp1: explicit clock reference needed by RCC clock driver")
however this is not part of upstream stm32mp151.dtsi . The
RCC driver does need this reference, reinstate it locally.

Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
8 months agoARM: dts: stm32: Reinstate missing root oscillators on STM32MP15xx
Marek Vasut [Sun, 15 Dec 2024 23:31:38 +0000 (00:31 +0100)] 
ARM: dts: stm32: Reinstate missing root oscillators on STM32MP15xx

The root oscillators reference used to be in rcc node since
3d15245502c4 ("ARM: dts: stm32mp1: explicit clock reference needed by RCC clock driver")
however this is not part of upstream stm32mp151.dtsi . The
RCC driver does need this reference, reinstate it globally.

Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
Tested-by: Patrice Chotard <patrice.chotard@foss.st.com>
8 months agoKconfig: Set STACK_SIZE to 16KB for STM32 MCUs
Patrice Chotard [Wed, 18 Dec 2024 07:58:33 +0000 (08:58 +0100)] 
Kconfig: Set STACK_SIZE to 16KB for STM32 MCUs

Since commit 6534d26ee9a5 ("lmb: do away with arch_lmb_reserve()"),
STM32F746-disco hangs when loading device tree just before starting
kernel:

Retrieving file: /stm32f746-disco.dtb
Kernel image @ 0xc0008000 [ 0x000000 - 0x19ae00 ]
Flattened Device Tree blob at c0408000
   Booting using the fdt blob at 0xc0408000
Working FDT set to c0408000
   Loading Device Tree to c05f8000, end c05ff71c ...

Adjust STACK_SIZE to 16KB for STM32 MCUs (F4/F7 and H7) boards
to fix kernel boot process as some of these boards embeds a limited
amount of memory.

Fixes: 6534d26ee9a5 ("lmb: do away with arch_lmb_reserve()")
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
8 months agoriscv: spl: add FIT name for RISC-V Falcon mode
Randolph [Fri, 13 Dec 2024 06:51:16 +0000 (14:51 +0800)] 
riscv: spl: add FIT name for RISC-V Falcon mode

The FIT name in RISC-V Falcon mode should be different from that in
normal boot mode; it is called linux.itb. If the setting is missing
in common/spl, the normal boot file name will be used.

Signed-off-by: Randolph <randolph@andestech.com>
8 months agodoc: spacemit: bananapi_f3: document Banana Pi F3 board
Kongyang Liu [Sun, 15 Dec 2024 05:02:42 +0000 (13:02 +0800)] 
doc: spacemit: bananapi_f3: document Banana Pi F3 board

Add document for Banana Pi F3 board which based on SpacemiT's K1 SoC.

Signed-off-by: Kongyang Liu <seashell11234455@gmail.com>
Signed-off-by: Huan Zhou <pericycle.cc@gmail.com>
Tested-by: Marcel Ziswiler <marcel@ziswiler.com>
Reviewed-by: Yixun Lan <dlan@gentoo.org>
8 months agoriscv: spacemit: bananapi_f3: initial support added
Kongyang Liu [Sun, 15 Dec 2024 05:02:41 +0000 (13:02 +0800)] 
riscv: spacemit: bananapi_f3: initial support added

Add basic support for SpacemiT's Banana Pi F3 board.
Update the k1.dtsi align with mainline.
Note that the device tree files follow the mainline Linux source[1].

Links: https://patches.linaro.org/project/linux-serial/patch/20240730-k1-01-basic-dt-v5-8-98263aae83be@gentoo.org/ [1]

Signed-off-by: Kongyang Liu <seashell11234455@gmail.com>
Signed-off-by: Huan Zhou <pericycle.cc@gmail.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
Reviewed-by: Yixun Lan <dlan@gentoo.org>
Tested-by: Marcel Ziswiler <marcel@ziswiler.com>
8 months agoconfigs: th1520_lpi4a: enable mmc controller support
Maksim Kiselev [Wed, 11 Dec 2024 20:11:01 +0000 (23:11 +0300)] 
configs: th1520_lpi4a: enable mmc controller support

Enable driver for DesignWare MSHC. TH1520 supports all
speed modes up to HS400ES and UHS SDR105.

Also enable ADMA and mmc command.

Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Maksim Kiselev <bigunclemax@gmail.com>
8 months agoriscv: dts: t-head: Add sdhci and emmc nodes
Maksim Kiselev [Wed, 11 Dec 2024 20:11:00 +0000 (23:11 +0300)] 
riscv: dts: t-head: Add sdhci and emmc nodes

Add SDHCI and EMMC controlles nodes on TH-1520 SoC. And enable them for
Lichee module 4A.

Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Maksim Kiselev <bigunclemax@gmail.com>
8 months agommc: snps_sdhci: Add sdhci driver support for TH1520 SoC
Maksim Kiselev [Wed, 11 Dec 2024 20:10:59 +0000 (23:10 +0300)] 
mmc: snps_sdhci: Add sdhci driver support for TH1520 SoC

Add support for DesignWare SDHCI host controller on Alibaba TH1520 SoC

Signed-off-by: Maksim Kiselev <bigunclemax@gmail.com>
Tested-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
8 months agoriscv: cpu: jh7110: Sort the list of imply statements
Hal Feng [Sun, 8 Dec 2024 09:19:42 +0000 (17:19 +0800)] 
riscv: cpu: jh7110: Sort the list of imply statements

The imply statements should be sorted in the sequence
of appearance in .config.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Suggested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoboard: starfive: spl: Support multiple DTBs for JH7110 based boards
Hal Feng [Sun, 8 Dec 2024 09:19:41 +0000 (17:19 +0800)] 
board: starfive: spl: Support multiple DTBs for JH7110 based boards

Get product ID and the other information from EEPROM, use them to select
the correct DTB.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoboard: starfive: spl: Fix the wrong use of CONFIG_IS_ENABLED()
Hal Feng [Sun, 8 Dec 2024 09:19:40 +0000 (17:19 +0800)] 
board: starfive: spl: Fix the wrong use of CONFIG_IS_ENABLED()

The prefix "SPL_" is not needed when using CONFIG_IS_ENABLED().

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Fixes: 5ecf9b0b8a75 ("board: starfive: add StarFive VisionFive v2 board support")
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoriscv: dts: jh7110: Support multiple DTBs in a Fit image
Hal Feng [Sun, 8 Dec 2024 09:19:39 +0000 (17:19 +0800)] 
riscv: dts: jh7110: Support multiple DTBs in a Fit image

Support multiple DTBs for JH7110 based boards, so they can
select the correct DT at runtime.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Reviewed-by: E Shattow <lucent@gmail.com>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoconfigs: visionfive2: Enable MULTI_DTB_FIT for JH7110 based board DT
Hal Feng [Sun, 8 Dec 2024 09:19:38 +0000 (17:19 +0800)] 
configs: visionfive2: Enable MULTI_DTB_FIT for JH7110 based board DT

So JH7110 based boards can select their own DT at runtime.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Reviewed-by: E Shattow <lucent@gmail.com>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoboard: starfive: spl: Drop the unneeded DT modification code
Hal Feng [Sun, 8 Dec 2024 09:19:37 +0000 (17:19 +0800)] 
board: starfive: spl: Drop the unneeded DT modification code

As OF_UPSTREAM is implemented, these code are redundant.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoriscv: dts: jh7110: Add u-boot device tree for JH7110 based boards
Hal Feng [Sun, 8 Dec 2024 09:19:36 +0000 (17:19 +0800)] 
riscv: dts: jh7110: Add u-boot device tree for JH7110 based boards

To support the other JH7110 based boards, add u-boot
device tree for them.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Reviewed-by: E Shattow <lucent@gmail.com>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: H Bell <dmoo_dv@protonmail.com>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoriscv: dts: jh7110: Move common code to the new jh7110-common-u-boot.dtsi
Hal Feng [Sun, 8 Dec 2024 09:19:35 +0000 (17:19 +0800)] 
riscv: dts: jh7110: Move common code to the new jh7110-common-u-boot.dtsi

To support JH7110 based boards besides v1.3B,
add a common dtsi and add common code to it.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Reviewed-by: E Shattow <lucent@gmail.com>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agopcie: starfive: Make the driver compatible with upstream DT
Hal Feng [Sun, 8 Dec 2024 09:19:34 +0000 (17:19 +0800)] 
pcie: starfive: Make the driver compatible with upstream DT

There are difference between upstream DT and the old DT
in terms of reg base, reset gpio and syscon. Make the driver
compatible with upstream DT.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agommc: dw_mmc: Add "starfive, jh7110-mmc" compatible to match upstream DT
Hal Feng [Sun, 8 Dec 2024 09:19:33 +0000 (17:19 +0800)] 
mmc: dw_mmc: Add "starfive, jh7110-mmc" compatible to match upstream DT

Make the U-Boot JH7110 MMC driver compatible with upstream DT.

Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoriscv: dts: jh7110: Make u-boot device trees adapting to upstream DT
Hal Feng [Sun, 8 Dec 2024 09:19:32 +0000 (17:19 +0800)] 
riscv: dts: jh7110: Make u-boot device trees adapting to upstream DT

Add u-boot features to the U-Boot device tree.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Reviewed-by: E Shattow <lucent@gmail.com>
Acked-by: Sumit Garg <sumit.garg@linaro.org>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agodts: starfive: Switch to using upstream DT
Hal Feng [Sun, 8 Dec 2024 09:19:31 +0000 (17:19 +0800)] 
dts: starfive: Switch to using upstream DT

Enable OF_UPSTREAM to use upstream DT and add starfive/ prefix to
the DEFAULT_DEVICE_TREE. Rename jh7110-starfive-visionfive-2-u-boot.dtsi
to jh7110-starfive-visionfive-2-v1.3b-u-boot.dtsi and set the v1.3b
device tree as the default device tree.

Drop redundant DT files from arch/riscv/dts/ and redundant clock and
reset definitions from include/dt-bindings/.

Since the old clock definitions is a little different from those in
upstream Linux, update the clock definitions in clock drivers
accordingly.

Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: E Shattow <lucent@gmail.com>
Acked-by: Sumit Garg <sumit.garg@linaro.org>
Signed-off-by: Hal Feng <hal.feng@starfivetech.com>
8 months agoboard: add support for LicheeRV Nano
Thomas Bonnefille [Tue, 12 Nov 2024 14:57:38 +0000 (15:57 +0100)] 
board: add support for LicheeRV Nano

The LicheeRV Nano is a small SBC using the Sophgo SG2002 RISCV SoC.

Signed-off-by: Thomas Bonnefille <thomas.bonnefille@bootlin.com>
8 months agoriscv: dts: sophgo: add device tree for LicheeRV Nano
Thomas Bonnefille [Tue, 12 Nov 2024 14:57:37 +0000 (15:57 +0100)] 
riscv: dts: sophgo: add device tree for LicheeRV Nano

Import a slightly modified version of the LicheeRV Nano and SG2002
device trees from the Linux Kernel. The current supported IPs are UART,
MMC, Timer, PLIC and CLINT.

Signed-off-by: Thomas Bonnefille <thomas.bonnefille@bootlin.com>
8 months agodoc: add LicheeRV Nano and SG2002 SoC
Thomas Bonnefille [Tue, 12 Nov 2024 14:57:36 +0000 (15:57 +0100)] 
doc: add LicheeRV Nano and SG2002 SoC

Provide a page describing the usage of U-Boot on the LicheeRV Nano and a
description of the board.

Signed-off-by: Thomas Bonnefille <thomas.bonnefille@bootlin.com>
Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com>
8 months agoMerge branch 'u-boot-nand-20241212' of https://source.denx.de/u-boot/custodians/u...
Tom Rini [Tue, 17 Dec 2024 19:16:37 +0000 (13:16 -0600)] 
Merge branch 'u-boot-nand-20241212' of https://source.denx.de/u-boot/custodians/u-boot-nand-flash into next

CI: https://source.denx.de/u-boot/custodians/u-boot-nand-flash/-/pipelines/23837

Small addition to uboot-nand. Nothing relevant for now. Anyway
a nice new command coming from Miquel Raynal and small changes.

8 months agonand: Add a watch command
Michael Trimarchi [Thu, 12 Dec 2024 18:28:07 +0000 (19:28 +0100)] 
nand: Add a watch command

This is a debug command to monitor the retention state of the data on
the array. The command needs a duplication of the mtd_read_oob()
function to actually return the maximum number of bitflips encountered
while reading the page. We could write a specific implementation for the
Sunxi driver but this is probably enough.

    nand watch <off> <size> - check an area for bitflips
    nand watch.part <part> - check a partition for bitflips
    nand watch.chip - check the whole device for bitflips

The output may be a bit verbose and could look like:

    => nand watch.chip
    device 0 whole chip
    size adjusted to 0xff60000 (5 bad blocks)

    NAND watch for bitflips in area 0x0-0xff60000:
    Page       0 (0x00000000) -> error -74
    Page       1 (0x00000800) -> error -74
    Page       2 (0x00001000) -> error -74
    Page       3 (0x00001800) -> error -74
    Page       4 (0x00002000) -> error -74
    Page       5 (0x00002800) -> error -74
    Page       6 (0x00003000) -> error -74
    Page       7 (0x00003800) -> error -74
    Page       8 (0x00004000) -> error -74
    Page       9 (0x00004800) -> error -74
    Page      10 (0x00005000) -> error -74
    Page      11 (0x00005800) -> error -74
    Page      12 (0x00006000) -> error -74
    Page      13 (0x00006800) -> error -74
    Page      14 (0x00007000) -> error -74
    Page      15 (0x00007800) -> error -74
    Page      16 (0x00008000) -> error -74
    Page      17 (0x00008800) -> error -74
    Page      18 (0x00009000) -> error -74
    Page      19 (0x00009800) -> error -74
    Page      20 (0x0000a000) -> error -74
    Page      21 (0x0000a800) -> error -74
    Page      22 (0x0000b000) -> error -74
    Page      23 (0x0000b800) -> error -74
    Page    1110 (0x0022b000) -> up to  1 bf/chunk
    Page    1122 (0x00231000) -> up to  1 bf/chunk
    Page    1132 (0x00236000) -> up to  1 bf/chunk
    Page    1362 (0x002a9000) -> up to  1 bf/chunk
    Page    4990 (0x009bf000) -> up to  1 bf/chunk
    Page    5728 (0x00b30000) -> up to  1 bf/chunk
    Page    7116 (0x00de6000) -> up to  1 bf/chunk
    Page    7160 (0x00dfc000) -> up to  1 bf/chunk
    Page    7494 (0x00ea3000) -> up to  1 bf/chunk
    Page   10842 (0x0152d000) -> up to  1 bf/chunk
    Page   11614 (0x016af000) -> up to  1 bf/chunk
    Page   11970 (0x01761000) -> up to  1 bf/chunk
    Page   12536 (0x0187c000) -> up to  1 bf/chunk
    Page   12687 (0x018c7800) -> up to  1 bf/chunk
    Page   14298 (0x01bed000) -> up to  1 bf/chunk
    Page   18268 (0x023ae000) -> up to  1 bf/chunk
    Page   18760 (0x024a4000) -> up to  1 bf/chunk
    Page   21440 (0x029e0000) -> up to  1 bf/chunk
    Page   22336 (0x02ba0000) -> up to  1 bf/chunk
    Page   22592 (0x02c20000) -> up to  1 bf/chunk
    Page   23872 (0x02ea0000) -> up to  1 bf/chunk
    Page   27584 (0x035e0000) -> up to  1 bf/chunk
    Page   35008 (0x04460000) -> up to  1 bf/chunk
    Page   37184 (0x048a0000) -> up to  1 bf/chunk
    Page   41728 (0x05180000) -> up to  1 bf/chunk
    Page   42176 (0x05260000) -> up to  1 bf/chunk
    Page   43200 (0x05460000) -> up to  1 bf/chunk
    Page   43328 (0x054a0000) -> up to  1 bf/chunk
    Page   45376 (0x058a0000) -> up to  1 bf/chunk
    Page   47040 (0x05be0000) -> up to  1 bf/chunk
    Page   47552 (0x05ce0000) -> up to  1 bf/chunk
    Page   49344 (0x06060000) -> up to  1 bf/chunk
    Page   49856 (0x06160000) -> up to  1 bf/chunk
    Page   62784 (0x07aa0000) -> up to  1 bf/chunk
    Page   65153 (0x07f40800) -> up to  1 bf/chunk
    Page   65228 (0x07f66000) -> up to  1 bf/chunk
    Page   65382 (0x07fb3000) -> up to  1 bf/chunk
    Page   98624 (0x0c0a0000) -> up to  1 bf/chunk
    Page  101952 (0x0c720000) -> up to  1 bf/chunk
    Page  107584 (0x0d220000) -> up to  1 bf/chunk
    Page  118208 (0x0e6e0000) -> up to  1 bf/chunk
    Page  126656 (0x0f760000) -> up to  1 bf/chunk
    Page  127680 (0x0f960000) -> up to  1 bf/chunk
    Page  129920 (0x0fdc0000) -> up to  1 bf/chunk
    Maximum number of bitflips: 1
    Pages with bitflips: 44/130752

It is also possible to reduce the output with the .quiet suffix in order
to just show the summary.

    => nand watch.chip
    device 0 whole chip
    size adjusted to 0xff60000 (5 bad blocks)

    NAND watch for bitflips in area 0x0-0xff60000:
    Maximum number of bitflips: 1
    Pages with bitflips: 44/130752

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com>
8 months agomtd: nand: raw: atmel: remove unnecessary return value
Marcus Folkesson [Fri, 9 Aug 2024 12:15:43 +0000 (14:15 +0200)] 
mtd: nand: raw: atmel: remove unnecessary return value

The condition 'ret' is always true as it is never set to other than
-EIO.

Remove 'ret' and the condition for copy.

Signed-off-by: Marcus Folkesson <marcus.folkesson@gmail.com>
Reviewed-by: Michael Trimarchi <micheal@amarulasolutions.com>
8 months agocmd: mtd: simplify mtd_show_device()
Heinrich Schuchardt [Mon, 9 Sep 2024 23:23:34 +0000 (01:23 +0200)] 
cmd: mtd: simplify mtd_show_device()

CONFIG_DM=y is always true.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Michael Trimarchi <michael@amarulasolutions.com>
8 months agoMerge tag 'u-boot-imx-next-20241217' of https://gitlab.denx.de/u-boot/custodians...
Tom Rini [Tue, 17 Dec 2024 19:15:58 +0000 (13:15 -0600)] 
Merge tag 'u-boot-imx-next-20241217' 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/23877

- Add support for Apalis iMX8 1300MHz version.
- Don't advertise Gbit on (R)MII on the FEC controller.
- Fix srktool -c usage by removing spaces.

8 months agoARM: stm32mp: Fix dram_bank_mmu_setup() for LMB located above ram_top
Patrice Chotard [Fri, 13 Dec 2024 13:26:55 +0000 (14:26 +0100)] 
ARM: stm32mp: Fix dram_bank_mmu_setup() for LMB located above ram_top

Previously, all LMB marked with LMB_NOMAP (above and below ram_top)
are considered as invalid entry in TLB.

Since commit 1a48b0be93d4 ("lmb: prohibit allocations above ram_top
 even from same bank") all LMB located above ram_top are now marked
LMB_NOOVERWRITE and no more LMB_MAP.

This area above ram_top is reserved for OPTEE and must not be cacheable,
otherwise this leads to a Panic on some boards (Issue on STM32MP135F-DK).

Restore previous behavior by marking invalid entry all TLB above ram_top.

Fixes: 1a48b0be93d4 ("lmb: prohibit allocations above ram_top even from same bank")
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
cc: Sughosh Ganu <sughosh.ganu@linaro.org>
Acked-by: Sughosh Ganu <sughosh.ganu@linaro.org>
8 months agogitlab: Add an rpi to the sjg lab
Simon Glass [Fri, 13 Dec 2024 00:59:03 +0000 (17:59 -0700)] 
gitlab: Add an rpi to the sjg lab

I have an original rpi installed now, loaded with OS Lite (32-bit)
Add an entry for it so that it can be used for testing.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
8 months agoboot: Use correct PHASE_ variable for expo
Simon Glass [Sat, 7 Dec 2024 14:52:27 +0000 (07:52 -0700)] 
boot: Use correct PHASE_ variable for expo

This patch was written before the XPL change-over. Update the Makefile
condition to the new way.

Signed-off-by: Simon Glass <sjg@chromium.org>
Fixes: ae3b5928d61 ("x86: coreboot: Allow building an expo for...")
Reviewed-by: Quentin Schulz <quentin.schulz@cherry.de>
8 months agotest/py: Always use the current dir as the source tree
Simon Glass [Wed, 11 Dec 2024 13:18:58 +0000 (06:18 -0700)] 
test/py: Always use the current dir as the source tree

The logic in get_details() retrieves the default source directory from
the Labgrid settings. This is convenient for interactive use, since it
allows pytests to be run from any directory and still find the source
tree.

However, it is not actually correct.

Gitlab sets the current directory to the source tree and expects that to
be used. At present it is ignored. The result is that Gitlab builds
whatever happens to be in the default source directory, ignoring the
tree it is supposed to be building.

Fix this by using the directory of the source tree, always. This is
obtained by looking at the grandparent of the conftest.py file.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reported-by: Tom Rini <trini@konsulko.com>
Fixes: bf89a8f1fc2 ("test: Introduce the concept of a role")
Tested-by: Tom Rini <trini@konsulko.com>
8 months agoMerge patch series "Tegra: fix clock init"
Tom Rini [Mon, 16 Dec 2024 23:07:42 +0000 (17:07 -0600)] 
Merge patch series "Tegra: fix clock init"

Svyatoslav Ryhel <clamor95@gmail.com> says:

This should fix:
https://lore.kernel.org/all/20241201164810.GT3600562@bill-the-cat/T/#m2b62b471fd37e4c9725c98547e2a919cf11db5e5

Link: https://lore.kernel.org/r/20241213145319.10956-1-clamor95@gmail.com
8 months agodriver: clk: tegra: init basic clocks on probe
Svyatoslav Ryhel [Fri, 13 Dec 2024 14:53:19 +0000 (16:53 +0200)] 
driver: clk: tegra: init basic clocks on probe

In case DM drivers probe earlier than board clock setup is done
init of basic clocks should be done in CAR driver probe as well.
Add it to avoid possible clock related problems.

Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Svyatoslav Ryhel <clamor95@gmail.com>
8 months agodriver: clk: tegra: partially support PLL clocks
Svyatoslav Ryhel [Fri, 13 Dec 2024 14:53:18 +0000 (16:53 +0200)] 
driver: clk: tegra: partially support PLL clocks

Return PLL id into struct clk if PLL is parsed from device
tree instead of throwing an error. Allow requesting PLL
clock rate via get_rate op.

Signed-off-by: Svyatoslav Ryhel <clamor95@gmail.com>
8 months agoMerge branch 'master' of https://source.denx.de/u-boot/custodians/u-boot-mmc
Tom Rini [Mon, 16 Dec 2024 22:17:13 +0000 (16:17 -0600)] 
Merge branch 'master' of https://source.denx.de/u-boot/custodians/u-boot-mmc

CI: https://source.denx.de/u-boot/custodians/u-boot-mmc/-/pipelines/23871

- Fix potential timer value truncation

8 months agoimx: hab: fix srktool -c usage by removing spaces
Christoph Fritz [Mon, 16 Dec 2024 14:59:00 +0000 (15:59 +0100)] 
imx: hab: fix srktool -c usage by removing spaces

The srktool option -c does not allow spaces between certificate
filenames. Only commas (',') should separate the filenames. If spaces
are incorrectly included, srktool will not display an error or warning
message but will only process the first certificate in the list.

So adapt documentation accordingly.

Signed-off-by: Christoph Fritz <chf.fritz@googlemail.com>