platform/kernel/u-boot.git
2 years agovideo: sandbox: Set a maximum frame-buffer size
Simon Glass [Fri, 19 Nov 2021 20:23:48 +0000 (13:23 -0700)]
video: sandbox: Set a maximum frame-buffer size

If U-Boot starts with the frame buffer set to 16bpp but then runs a test
that uses 32bpp, there is not enough space. Update the driver to use the
maximum possible frame-buffer size, to avoid this.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoconsole: Avoid serial output before the console is running
Simon Glass [Fri, 19 Nov 2021 20:23:47 +0000 (13:23 -0700)]
console: Avoid serial output before the console is running

The video driver uses this for debugging, but if used before relocation it
crashes at present. Avoid trying to output debugging before the console is
ready.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agovideo: sandbox: Avoid duplicate display windows
Simon Glass [Fri, 19 Nov 2021 20:23:46 +0000 (13:23 -0700)]
video: sandbox: Avoid duplicate display windows

When unit tests are run they currently create a new window. Update the
code so that the old one is removed first. This avoids the confusion as to
which one is active.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agosandbox: video: Support 8bpp depth
Simon Glass [Fri, 19 Nov 2021 20:23:45 +0000 (13:23 -0700)]
sandbox: video: Support 8bpp depth

At present sandbox only supports 16 and 32bpp depths, since those are the
easy ones with SDL.

We can support other depths by manually converting the pixel formats. Add
support for this, to enable an 8ppp (monochrome) format.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoMerge branch '2021-12-23-make-OF_BOARD-a-boolean' into next
Tom Rini [Fri, 24 Dec 2021 14:31:35 +0000 (09:31 -0500)]
Merge branch '2021-12-23-make-OF_BOARD-a-boolean' into next

Merge v8 of Simon's series to make CONFIG_OF_BOARD a boolean option.
Quoting him:

With Ilias' efforts we have dropped OF_PRIOR_STAGE and OF_HOSTFILE so
there are only three ways to obtain a devicetree:

   - OF_SEPARATE - the normal way, where the devicetree is built and
      appended to U-Boot
   - OF_EMBED - for development purposes, the devicetree is embedded in
      the ELF file (also used for EFI)
   - OF_BOARD - the board figures it out on its own

The last one is currently set up so that no devicetree is needed at all
in the U-Boot tree. Most boards do provide one, but some don't. Some
don't even provide instructions on how to boot on the board.

The problems with this approach were covered in another patch[1], since
removed from this series.

In practice, OF_BOARD is not really distinct from OF_SEPARATE. Any board
can obtain its devicetree at runtime, even it is has a devicetree built
in U-Boot. This is because U-Boot may be a second-stage bootloader and its
caller may have a better idea about the hardware available in the machine.
This is the case with a few QEMU boards, for example.

So it makes no sense to have OF_BOARD as a 'choice'. It should be an
option, available with either OF_SEPARATE or OF_EMBED. This would allow
rpi3, for example, to run with the devicetree provided by the prior
bootloader.

This series makes this change, adding various missing devicetree files
(and placeholders) to make the build work.

To make the 'prior stage' side of things more deterministic, a new
OF_HAS_PRIOR_STAGE is added, which cannot be disabled by updated a board's
defconfig. This should help to prevent mistakes.

It also adds a run-time message showing where the devicetree came from,
as well as warnings if the board's expected flow is not being used. This
comes originally from the 'standard passage' series, which depends on
this series.

It also provides a few qemu clean-ups discovered along the way. The
qemu-riscv64_spl problem is fixed.

Please see [2] for discussion on the v6 series.

I put Heinrich's Tested-by tag[3] for the series onto the three devicetree
patches (ARM and RISC-V) that I think it most affects. It isn't possible
to apply a tag to a whole series at present and in any case there are
changes in v7.

This series is available at u-boot-dm/ofb-working

[1] https://patchwork.ozlabs.org/project/uboot/patch/20211207001209.3467163-2-sjg@chromium.org/
[2] https://lore.kernel.org/u-boot/20211205133207.GW1220664@bill-the-cat/T/#mcd8c0258827fbc1bb3000b7ff9ba0929df1ddcb2
[3] https://lore.kernel.org/u-boot/93913911-4d20-d28f-ee04-739985184c5e@canonical.com/raw

2 years agofdt: Show a runtime warning based on devicetree source
Simon Glass [Fri, 17 Dec 2021 03:59:39 +0000 (20:59 -0700)]
fdt: Show a runtime warning based on devicetree source

When running, if the devicetree failed to come from the expected source,
show a warning, e.g:

   U-Boot ...

   DRAM:  128 MiB
   Core:  42 devices, 11 uclasses, devicetree: separate
   Warning: Unexpected devicetree source (not from a prior stage)
   Warning: U-Boot may not function properly
   Flash: 64 MiB
   ...

These warnings should only appear if the board config has been changed, or
the prior stage is broken.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Avoid emitting an device tree when not needed
Simon Glass [Fri, 17 Dec 2021 03:59:38 +0000 (20:59 -0700)]
fdt: Avoid emitting an device tree when not needed

U-Boot always needs some sort of a device tree in the build. Some boards
never actually use this, at least in production systems, since a prior
firmware stage sets one up and passes it to U-Boot. At present the only
mechanism to do that is with custom function (OF_BOARD), but future work
will include a standard way of doing this ('standard passage').

It can be confusing to see a device tree emitted from the U-Boot build in
this situation. Add an option to drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Makefile: Ensure that OF_BOARD is used when needed
Simon Glass [Fri, 17 Dec 2021 03:59:37 +0000 (20:59 -0700)]
fdt: Makefile: Ensure that OF_BOARD is used when needed

Boards which define OF_HAS_PRIOR_STAGE must define OF_BOARD at present,
since a custom function is the only way to obtain the devicetree at
runtime.

Add a build error when this requirement is not met, to avoid accepting
any patches which break this requirement.

Add an allowlist for boards which use it, currently none. This allowlist
can be updated for local development, if needed.

Signed-off-by: Simon Glass <sjg@chromium.org>
[trini: change of_whitelist to of_allowlist]
Signed-off-by: Tom Rini <trini@konsulko.com>
2 years agofdt: Enable OF_HAS_PRIOR_STAGE for most boards with OF_BOARD
Simon Glass [Fri, 17 Dec 2021 03:59:36 +0000 (20:59 -0700)]
fdt: Enable OF_HAS_PRIOR_STAGE for most boards with OF_BOARD

Use this new Kconfig instead of OF_BOARD, so we know for sure which boards
obtain their devicetree from a prior stage. Leave sandbox alone since it
does not. Also don't touch xilinx_versal_virt since it does not have a
specific TARGET Kconfig.

This option implies OF_BOARD for now, but with future work standard
passage may be used instead.

Signed-off-by: Simon Glass <sjg@chromium.org>
[trini: Add rpi_4_32b and rpi_arm64 to the list of boards converted]
Signed-off-by: Tom Rini <trini@konsulko.com>
2 years agofdt: Add a Kconfig for boards with a prior stage
Simon Glass [Fri, 17 Dec 2021 03:59:35 +0000 (20:59 -0700)]
fdt: Add a Kconfig for boards with a prior stage

When U-Boot is started from another firmware program, not just a prior
phase of U-Boot, special behaviour is typically used. In particular, the
device tree may come from that prior stage.

At present this is sort-of indicated by OF_BOARD, although the
correlation is not 1:1, since that option simply means that the board has
a custom mechanism for obtaining the device tree. For example, sandbox
defines OF_BOARD. Also the board_fdt_blob_setup() function can in fact
make use of the devicetree in U-Boot if it wishes, as used by
dragonboard410c until very recently.

Add an explicit Kconfig for this situation. Update the OF_BOARD option to
more-accurately reflect what it is doing, e.g. for sandbox.

Drop the docs in the README as it is out of date.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Report the devicetree source
Simon Glass [Fri, 17 Dec 2021 03:59:34 +0000 (20:59 -0700)]
fdt: Report the devicetree source

It can be confusing to figure out where the devicetree came from. It seems
important enough to warrant a message during boot. Add information about
the number of devices and uclasses too since it is helpful to have some
idea what is going on with driver model.

Report the devicetree source in bdinfo too.

This looks something like this, with > marking the new line.

   U-Boot 2021.10-00190 (Oct 30 2021 - 09:01:29 -0600)

   DRAM:  128 MiB
>  Core:  42 devices, 11 uclasses, devicetree: passage
   Flash: 64 MiB

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Record where the devicetree came from
Simon Glass [Fri, 17 Dec 2021 03:59:33 +0000 (20:59 -0700)]
fdt: Record where the devicetree came from

Keep track of where the devicetree came from, so we can report this later.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agodm: core: Allow getting some basic stats
Simon Glass [Fri, 17 Dec 2021 03:59:32 +0000 (20:59 -0700)]
dm: core: Allow getting some basic stats

Add a function that returns some basic stats about driver model. For now
we only have two.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Don't call board_fdt_blob_setup() without OF_BOARD
Simon Glass [Fri, 17 Dec 2021 03:59:31 +0000 (20:59 -0700)]
fdt: Don't call board_fdt_blob_setup() without OF_BOARD

At present this override function is called even when OF_BOARD is not
enabled. This makes it impossible to disable this feature and in fact
makes the OF_BOARD option useless.

Reinstate its intended purpose, so that it is possible to switch between
the appended devicetree and one provided by the board's custom function.

A follower patch adds warnings for this scenario, but for now we don't
have a Kconfig that definitively tells us that OF_BOARD should be used.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Drop remaining preprocessor macros in fdtdec_setup()
Simon Glass [Fri, 17 Dec 2021 03:59:30 +0000 (20:59 -0700)]
fdt: Drop remaining preprocessor macros in fdtdec_setup()

We only have two choices for obtaining the devicetree. Simplify the code
to make that clear.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Drop OF_CONTROL check in fdtdec_setup()
Simon Glass [Fri, 17 Dec 2021 03:59:29 +0000 (20:59 -0700)]
fdt: Drop OF_CONTROL check in fdtdec_setup()

This function should only be called when OF_CONTROL is enabled. It
fails in fdtdec_prepare_fdt() anyway, since gd->fdt_blob stays as NULL
if OF_CONTROL is not enabled.

Drop this useless check.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Use if() for fdtcontroladdr check
Simon Glass [Fri, 17 Dec 2021 03:59:28 +0000 (20:59 -0700)]
fdt: Use if() for fdtcontroladdr check

Change this to use if() instead of #if

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Drop #ifdef around board_fdt_blob_setup()
Simon Glass [Fri, 17 Dec 2021 03:59:27 +0000 (20:59 -0700)]
fdt: Drop #ifdef around board_fdt_blob_setup()

This serves no purpose. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Drop CONFIG_SPL_BUILD check in fdtdec_setup()
Simon Glass [Fri, 17 Dec 2021 03:59:26 +0000 (20:59 -0700)]
fdt: Drop CONFIG_SPL_BUILD check in fdtdec_setup()

Move this to the header file to clean up the C code.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Drop #ifdefs with MULTI_DTB_FIT
Simon Glass [Fri, 17 Dec 2021 03:59:25 +0000 (20:59 -0700)]
fdt: Drop #ifdefs with MULTI_DTB_FIT

Refactor the code to drop the #ifdefs for this feature.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Move MULTI_DTB_FIT handling out of fdtdec_setup()
Simon Glass [Fri, 17 Dec 2021 03:59:24 +0000 (20:59 -0700)]
fdt: Move MULTI_DTB_FIT handling out of fdtdec_setup()

This logic is a bit convoluted for one function. Move the mulit-FIT part
into its own function.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
2 years agofdt: Make it easier to debug u-boot.dtsi files
Simon Glass [Fri, 17 Dec 2021 03:59:23 +0000 (20:59 -0700)]
fdt: Make it easier to debug u-boot.dtsi files

At present one must hack the Makefile to see what is going on with these
files. Also it doesn't quite work correctly.

Fix this by using an environment variable for debugging. Update the docs
also.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Drop CONFIG_BINMAN_STANDALONE_FDT
Simon Glass [Fri, 17 Dec 2021 03:59:22 +0000 (20:59 -0700)]
fdt: Drop CONFIG_BINMAN_STANDALONE_FDT

This was added as a hack to work around not having an in-tree devicetree.
Now that this is fixed it is not needed.

Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdt: Make OF_BOARD a bool option
Simon Glass [Fri, 17 Dec 2021 03:59:21 +0000 (20:59 -0700)]
fdt: Make OF_BOARD a bool option

This should not be a separate option from OF_SEPARATE. It is a run-time
option to override the devicetree, even if present.

Move the option out of the choice.

Disable BINMAN_FDT for a few boards which don't actually use it.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: highbank: Add devicetree files
Simon Glass [Fri, 17 Dec 2021 03:59:20 +0000 (20:59 -0700)]
arm: highbank: Add devicetree files

Add an empty version of this file, so that we can at least build this
board when devicetrees are required.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: qemu-ppce500: Add a devicetree file
Simon Glass [Fri, 17 Dec 2021 03:59:19 +0000 (20:59 -0700)]
arm: qemu-ppce500: Add a devicetree file

This uses QEMU virt which creates its own devicetree.

Add an empty version of this file, so that we can at least build this
board when devicetrees are required.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: bcm7xxx: Add a devicetree file
Simon Glass [Fri, 17 Dec 2021 03:59:18 +0000 (20:59 -0700)]
arm: bcm7xxx: Add a devicetree file

Add an empty devicetree file for these boards. It seems to be possible to
obtain a real one from another bootloader called 'bolt' but I will leave
this to the maintainer.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: xilinx_versal_virt: Add a devicetree file
Simon Glass [Fri, 17 Dec 2021 03:59:17 +0000 (20:59 -0700)]
arm: xilinx_versal_virt: Add a devicetree file

Add an empty file to prevent build errors when building with
CONFIG_OF_SEPARATE enabled.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: octeontx: Add an empty devicetree file
Simon Glass [Fri, 17 Dec 2021 03:59:16 +0000 (20:59 -0700)]
arm: octeontx: Add an empty devicetree file

Add an empty file to prevent build errors when building with
CONFIG_OF_SEPARATE enabled.

Unfortunately there are no build instructions in the U-Boot tree to enable
a real file to be created.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: xenguest_arm64: Add a empty devicetree file
Simon Glass [Fri, 17 Dec 2021 03:59:15 +0000 (20:59 -0700)]
arm: xenguest_arm64: Add a empty devicetree file

Add an empty version of this file, so that we can at least build this
board when devicetrees are required.

The real devicetree is created by the Xen project on-the-fly.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: vexpress: Add a devicetree files for juno
Simon Glass [Fri, 17 Dec 2021 03:59:14 +0000 (20:59 -0700)]
arm: vexpress: Add a devicetree files for juno

Sync these file, obtained from Linux v5.15.

Add a note for the maintainer, and SPDX lines where they are missing.
The added lines are:

   SPDX-License-Identifier: GPL-2.0 OR BSD-3-Clause

Note, this matches the text in those files, but is not the same as the
GPL-2.0 of some files.

[1] https://releases.linaro.org/android/reference-lcr/juno/7.1-17.05/

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
2 years agoarm: rpi: Sync rpi dts files from Linux
Simon Glass [Fri, 17 Dec 2021 03:59:13 +0000 (20:59 -0700)]
arm: rpi: Sync rpi dts files from Linux

Sync these files, obtained from Linux v5.15.

This adds a devicetree file for rpi_4 which was not there before.

Testing shows no change so far as I can see:
- boots to U-Boot prompt on rpi0, rpi2
- boots to distro on rpi3
- boots to distro on rpi4

I am assuming that syncing with Linux is safe, but the maintainer should
know for sure.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoriscv: qemu: Split devicetree files for qemu_riscv32/64
Simon Glass [Fri, 17 Dec 2021 03:59:12 +0000 (20:59 -0700)]
riscv: qemu: Split devicetree files for qemu_riscv32/64

This uses QEMU virt which creates its own devicetree.

Copy the existing empty version of this file, so splitting the existing
qemu-virt into two, since anyone actually trying to use this will need a
different devicetree for 32- and 64-bit machines.

Tested-by: Heinrich Schuchardt <heinrich.schuchardt@canaonical.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: qemu: Add a devicetree file for qemu_arm64 virt
Simon Glass [Fri, 17 Dec 2021 03:59:11 +0000 (20:59 -0700)]
arm: qemu: Add a devicetree file for qemu_arm64 virt

This uses QEMU virt which creates its own devicetree.

Add an empty version of this file, so that we can at least build this
board when devicetrees are required.

Tested-by: Heinrich Schuchardt <heinrich.schuchardt@canaonical.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: qemu: Add a devicetree file for qemu_arm
Simon Glass [Fri, 17 Dec 2021 03:59:10 +0000 (20:59 -0700)]
arm: qemu: Add a devicetree file for qemu_arm

This uses QEMU virt which creates its own devicetree.

Add an empty version of this file, so that we can at least build this
board when devicetrees are required.

Tested-by: Heinrich Schuchardt <heinrich.schuchardt@canaonical.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: riscv: qemu: Explain how to extract the generated dt
Simon Glass [Fri, 17 Dec 2021 03:59:09 +0000 (20:59 -0700)]
arm: riscv: qemu: Explain how to extract the generated dt

QEMU currently generates a devicetree for use with U-Boot. Explain how to
obtain it.

Also explain how to merge it to produce a devicetree with the U-Boot
features included.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoarm: qemu: Mention -nographic in the docs
Simon Glass [Fri, 17 Dec 2021 03:59:08 +0000 (20:59 -0700)]
arm: qemu: Mention -nographic in the docs

Without this option QEMU appears to hang. Add it to avoid confusion.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
2 years agoMerge branch 'next' of https://source.denx.de/u-boot/custodians/u-boot-marvell into...
Tom Rini [Tue, 21 Dec 2021 13:02:28 +0000 (08:02 -0500)]
Merge branch 'next' of https://source.denx.de/u-boot/custodians/u-boot-marvell into next

- pci_mvebu: Misc improvements and cleanup (Pali)
- turris_mox: Remove extra newline after module topology (Marek)

2 years agodoc: remove duplicate page inclusion
Heinrich Schuchardt [Tue, 21 Dec 2021 08:00:44 +0000 (09:00 +0100)]
doc: remove duplicate page inclusion

doc/usage/index.rst in branch origin/next includes usage/environment twice.
Remove the duplicate entry.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agoarm: mvebu: turris_mox: Remove extra newline after module topology
Marek Behún [Thu, 11 Nov 2021 15:35:49 +0000 (16:35 +0100)]
arm: mvebu: turris_mox: Remove extra newline after module topology

Remove extra newline after module topology is printed.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agopci: pci_mvebu: Remove unused DECLARE_GLOBAL_DATA_PTR
Pali Rohár [Thu, 11 Nov 2021 15:35:47 +0000 (16:35 +0100)]
pci: pci_mvebu: Remove unused DECLARE_GLOBAL_DATA_PTR

The global data pointer is not used in this driver, remove it's
declaration.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agopci: pci_mvebu: Replace MBUS_PCI_*_SIZE by resource_size()
Pali Rohár [Thu, 11 Nov 2021 15:35:43 +0000 (16:35 +0100)]
pci: pci_mvebu: Replace MBUS_PCI_*_SIZE by resource_size()

Use more appropriate resource_size() function when working with data in
struct resource.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agopci: pci_mvebu: Move setup for BAR[0] where other BARs are setup
Pali Rohár [Thu, 11 Nov 2021 15:35:42 +0000 (16:35 +0100)]
pci: pci_mvebu: Move setup for BAR[0] where other BARs are setup

Function mvebu_pcie_setup_wins() sets up all other BARs, so move setup of
BAR[0] to this function to have common code at one place.

In the past, commit 193a1e9f196b ("pci: pci_mvebu: set BAR0 after memory
space is set") moved setup of BAR[0] to another location, due to ath10k
not working in kernel, but the reason why was unknown, but it seems to
work now, and we think the issue then was cause by the PCIe Root Port
presenting itself as a Memory Controller and therefore U-Boot's code
have overwritten the BAR. Since the driver now ignores any write
operations to PCIe Root Port BARs, this should not be an issue anymore.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoMerge tag 'v2022.01-rc4' into next
Tom Rini [Mon, 20 Dec 2021 22:12:04 +0000 (17:12 -0500)]
Merge tag 'v2022.01-rc4' into next

Prepare v2022.01-rc4

2 years agoPrepare v2022.01-rc4
Tom Rini [Mon, 20 Dec 2021 16:15:15 +0000 (11:15 -0500)]
Prepare v2022.01-rc4

Signed-off-by: Tom Rini <trini@konsulko.com>
2 years agoRevert "image: Remove #ifdefs from select_ramdisk()"
Tom Rini [Mon, 20 Dec 2021 14:36:32 +0000 (09:36 -0500)]
Revert "image: Remove #ifdefs from select_ramdisk()"

This reverts commit f33a2c1bd0fb371511a485cac1f182ba50db51be.

This causes a crash on some platforms as seen here:
https://lore.kernel.org/r/f153017b-c41a-0d32-67b9-f288e695f900@baylibre.com/

Reported-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
2 years agofw_setenv: Unbreak fw_setenv caused by buggy MEMISLOCKED use
Joakim Tjernlund [Wed, 8 Dec 2021 14:33:11 +0000 (15:33 +0100)]
fw_setenv: Unbreak fw_setenv caused by buggy MEMISLOCKED use

Commit "fw_setenv: lock the flash only if it was locked before"
checks for Locked status with uninitialized erase data.
Address by moving the test for MEMISLOCKED.

Fixes: 8a726b852502 ("fw_setenv: lock the flash only if it was locked before")
Signed-off-by: Joakim Tjernlund <joakim.tjernlund@infinera.com>
2 years agoMerge tag '20211220-fixes-for-2022.01' of https://source.denx.de/u-boot/custodians...
Tom Rini [Mon, 20 Dec 2021 13:51:53 +0000 (08:51 -0500)]
Merge tag '20211220-fixes-for-2022.01' of https://source.denx.de/u-boot/custodians/u-boot-i2c

i2c changes for 20211220-fixes-for-2022.01

- mvtwsi: Swab the register address if its size is > 1

2 years agoi2c: mvtwsi: Swab the register address if its size is > 1
Stefan Roese [Thu, 18 Nov 2021 08:18:41 +0000 (09:18 +0100)]
i2c: mvtwsi: Swab the register address if its size is > 1

Testing on Armada XP with an EEPROM using register address with size
of 2 has shown, that the register address bytes are sent to the I2C
EEPROM in the incorrect order. This patch swabs the address bytes so
that the correct address is transferred to the I2C device.

BTW: This worked without any issues before migrating Armada XP to
DM I2C.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Heiko Schocher <hs@denx.de>
Cc: Samuel Holland <samuel@sholland.org>
Cc: Baruch Siach <baruch@tkos.co.il>
Cc: Pali Rohár <pali@kernel.org>
Cc: Marek Behún <marek.behun@nic.cz>
Tested-by: Marek Behún <marek.behun@nic.cz>
2 years agoMerge branch 'next' of https://source.denx.de/u-boot/custodians/u-boot-marvell into...
Tom Rini [Sun, 19 Dec 2021 13:59:59 +0000 (08:59 -0500)]
Merge branch 'next' of https://source.denx.de/u-boot/custodians/u-boot-marvell into next

- Armada XP etc: Move to DM_I2C (Stefan)
- Some mvebu comphy + mox + fdt_support changes (Marek & Pali)
- mvebu: a38x: improve USB3 serdes configuration (Stefan Eichenberger)
- mvebu: Some maintainer updates (Pali)
- mvebu: Misc minor cleanup (Pali)

2 years agoarm: mvebu: Use printf for printing fatal errors
Pali Rohár [Fri, 17 Dec 2021 17:31:14 +0000 (18:31 +0100)]
arm: mvebu: Use printf for printing fatal errors

There is no point to hide/disable fatal errors via debug() macro.
Print fatal errors loudly.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: Remove commented example code
Pali Rohár [Fri, 17 Dec 2021 17:28:51 +0000 (18:28 +0100)]
arm: mvebu: Remove commented example code

Include file debug_uart.h already contains documentation how to use it.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: serial: Add me as co-maintainer and author of Marvell serial drivers
Pali Rohár [Thu, 16 Dec 2021 11:04:07 +0000 (12:04 +0100)]
arm: mvebu: serial: Add me as co-maintainer and author of Marvell serial drivers

There is no maintainer entry for serial_mvebu_a3700.c. Add entry with Pali
and Stefan as maintainers.

Signed-off-by: Pali Rohár <pali@kernel.org>
Acked-by: Stefan Roese <sr@denx.de>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: pci: Add me as co-maintainer and author of Marvell PCIe drivers
Pali Rohár [Thu, 16 Dec 2021 11:04:06 +0000 (12:04 +0100)]
arm: mvebu: pci: Add me as co-maintainer and author of Marvell PCIe drivers

There is no maintainer entry for pci-aardvark.c. Add entry for
pci-aardvark.c and pci_mvebu.c with Pali and Stefan as maintainers.

Signed-off-by: Pali Rohár <pali@kernel.org>
Acked-by: Stefan Roese <sr@denx.de>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: a38x: serdes: improve USB3 electrical configuration
Stefan Eichenberger [Thu, 9 Dec 2021 10:19:34 +0000 (11:19 +0100)]
arm: mvebu: a38x: serdes: improve USB3 electrical configuration

This is a backport from Marvell U-Boot:
https://github.com/MarvellEmbeddedProcessors/u-boot-marvell
commit 381d029e7a ("fix: serdes: a38x, a39x: Improve USB3 electrical
configuration")

Improves electrical USB3 receiver jitter tolerance test:

- De-Emphasize force, in functional mode the transmitter should always
have 3.5db de-emphasize, so we are forcing it.

- After forcing De-Emphasize, choose 3.5db (After forcing, default is
6dB so need to change it to 3.5dB).

- Align90 set to 0x58 - this is the sample point in the receiver, after
the clock is recovered this sampler samples at the chosen value, usually
it is supposed to be 0x60(which is the center of the eye), but sometimes
after adding jitter and ISI the center of the eye can move slightly and
the sample point is not necessarily the exact center, and after
optimization (searching the middle of the eye manually) it was seen that
the center of the eye is actually 0x58 and not 0x60.

- FFE Res and FFE Cap set to 0xE & 0xF respectively: improves this
settings is adequate according to how the USB3 spec defines the
interconnect, thus improves USB3 jitter tolerance settings.

- Change the resolution of the DFE to 0x3 which is 6mV(highest
resolution) , this avoids the DFE to saturate and cease to work.

- HPF set to 0x3 which is 5Khz high pass filter, the function of the HPF
is to filter the low frequency patterns(below 5Khz) to make sure that
the signal is not a noise, the setting before was 0x1(205Khz), and the
change came since the USB3 CP0 pattern, that is used in the USB3 jitter
tolerance testing, is similar to PRBS15, which has 2^15=32768bits which
is 32768*200ps (200ps is one Unit interval in USB3(5Gbps)) = 6.5us,
which is in frequency terms: 152Khz. since the PRBS15 is a random
pattern and can theoretically have once in a while a pattern that will
be at frequency of 152Khz, hence the previous setting (205khz HPF) can
possibly filter this pattern which can cause to an error in the
receiver, thus this change to avoid such scenarios.

Signed-off-by: Stefan Eichenberger <eichest@gmail.com>
Signed-off-by: René Straub <rene.straub@netmodule.com>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: a38x: serdes: fix serdes config for USB3
Stefan Eichenberger [Thu, 9 Dec 2021 10:19:33 +0000 (11:19 +0100)]
arm: mvebu: a38x: serdes: fix serdes config for USB3

The electrical serdes configuration for USB3 expects an array as data
argument. For USB3 the second value is used (see data_arr_idx = USB3 =
1). However, because only one value is inside the array mv_seq_exec is
accessing an invalid element and the serdes is configured wrongly.

This wrong initialization is leading to an unreliable detection
mechanism for some USB3 devices. We were able to reproduce the issue
regularly with an LTE modem from Sierra Wireless (SM7455) where it was
not detected as USB3 device in 1/3 of all tests.

This commit fixes the issue by setting data_arr_idx to 0. This is the
same value as the original U-Boot from Marvell is using. There it is
called FIRST_CELL which is a define for 0.
See: https://github.com/MarvellEmbeddedProcessors/u-boot-marvell
commit 56f963ce4c ("fix: serdes: a38x, a39x: Fix USB3 serdes DB
initialization")

Signed-off-by: Stefan Eichenberger <eichest@gmail.com>
Signed-off-by: René Straub <rene.straub@netmodule.com>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agofdt_support: Add fdt_delete_disabled_nodes() and use in Turris MOX
Marek Behún [Fri, 26 Nov 2021 13:57:15 +0000 (14:57 +0100)]
fdt_support: Add fdt_delete_disabled_nodes() and use in Turris MOX

Move Turris MOX specific remove_disabled_nodes() to fdt_support with
name fdt_delete_disabled_nodes(), so that others can potentially use it.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: turris_mox: Fix unstable board topology reading
Marek Behún [Fri, 26 Nov 2021 13:57:14 +0000 (14:57 +0100)]
arm: mvebu: turris_mox: Fix unstable board topology reading

The pre-relocation board topology reading in board_fix_fdt() is
unstable: sometimes wrong data are read from the SPI bus.

This is due to wrong order of SPI bus configuration instructions: we
first need to set the pins to SPI mode, and only after that configure
the bus.

Also add a 1ms delay before enabling chip-select, so that the clock pin
is high for some time before reading the bus.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agophy: marvell: a3700: Convert to official DT bindings in COMPHY driver
Pali Rohár [Fri, 26 Nov 2021 13:57:13 +0000 (14:57 +0100)]
phy: marvell: a3700: Convert to official DT bindings in COMPHY driver

Convert A3720 common PHY driver to official DT bindings.

This puts us closer to be able to synchronize A3720 device-trees with
those from Linux.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Cc: Konstantin Porotchkin <kostap@marvell.com>
Cc: Robert Marko <robert.marko@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Cc: Marcin Wojtas <mw@semihalf.com>
Cc: Grzegorz Jaszczyk <jaz@semihalf.com>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: turris_mox: Enable eth1 in U-Boot if a network module is present
Pali Rohár [Fri, 26 Nov 2021 13:57:12 +0000 (14:57 +0100)]
arm: mvebu: turris_mox: Enable eth1 in U-Boot if a network module is present

Enable eth1 node in U-Boot's device-tree if a network module (SFP, Topaz
or Peridot) is detected.

This is required for proper detection of eth1 comphy in a3700 comphy
driver by the following patches.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: turris_mox: Find DT nodes by compatible or alias instead of path
Marek Behún [Fri, 26 Nov 2021 13:57:11 +0000 (14:57 +0100)]
arm: mvebu: turris_mox: Find DT nodes by compatible or alias instead of path

It is better to find DT nodes by compatible strings or aliases instead
of path.

There were issues with Linux some DTBs having different names of some
nodes, e.g.
  internal-regs
instead of
  internal-regs@d0000000

This should be a generic fix for such issues.

Also since fdt_support now contains needed functions, we can drop our
own implementations.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agofdt_support: Add some useful functions
Marek Behún [Fri, 26 Nov 2021 13:57:10 +0000 (14:57 +0100)]
fdt_support: Add some useful functions

Add functions
  fdt_node_offset_by_pathf(),
  fdt_create_phandle_by_pathf(),
  fdt_set_status_by_pathf()
to get node offset, get/create node phandle and set status for node
given by path/alias formatted with sprintf.

Add functions
  fdt_create_phandle_by_compatible(),
  fdt_set_status_by_compatible()
to get/create node phandle and set status for first node given by
compatible.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agofdt_support: Fix comment for fdt_create_phandle()
Marek Behún [Fri, 26 Nov 2021 13:57:09 +0000 (14:57 +0100)]
fdt_support: Fix comment for fdt_create_phandle()

This function does not necessarily create a new phandle. If a phandle
exists, no new phandle is created.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agofdt_support: Remove FDT_STATUS_FAIL_ERROR_CODE
Marek Behún [Fri, 26 Nov 2021 13:57:08 +0000 (14:57 +0100)]
fdt_support: Remove FDT_STATUS_FAIL_ERROR_CODE

Since no one uses this feature and I am not aware of any parsers of this
in Linux, remove it.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Pratyush Yadav <p.yadav@ti.com>
Cc: Tim Harvey <tharvey@gateworks.com>
Cc: Michael Walle <michael@walle.cc>
Cc: Priyanka Jain <priyanka.jain@nxp.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agofdt_support: Remove fdt_alloc_phandle() in favor of fdt_generate_phandle()
Marek Behún [Fri, 26 Nov 2021 13:57:07 +0000 (14:57 +0100)]
fdt_support: Remove fdt_alloc_phandle() in favor of fdt_generate_phandle()

Commit f0921f5098d ("fdt: Sync up to the latest libfdt") introduced
fdt_generate_phandle() in libfdt, making fdt_alloc_phandle() obsolete in
fdt_support.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Cc: "hui.song" <hui.song_1@nxp.com>
Cc: Meenakshi Aggarwal <meenakshi.aggarwal@nxp.com>
Cc: Priyanka Jain <priyanka.jain@nxp.com>
Cc: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agotreewide: Use fdt_create_phandle() where appropriate
Marek Behún [Fri, 26 Nov 2021 13:57:06 +0000 (14:57 +0100)]
treewide: Use fdt_create_phandle() where appropriate

Replace fdt_alloc_phandle() with subsequent fdt_set_phandle() by
fdt_create_phandle().

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Cc: Aaron Williams <awilliams@marvell.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Vladimir Oltean <vladimir.oltean@nxp.com>
2 years agoinclude/linux/byteorder: Fix compilation of __constant_cpu_to_be32()
Pali Rohár [Fri, 26 Nov 2021 13:57:05 +0000 (14:57 +0100)]
include/linux/byteorder: Fix compilation of __constant_cpu_to_be32()

The macro __constant_cpu_to_be32() uses ___constant_swab32(), which for
some reason is not defined and causes the following error during
compilation:

  include/linux/byteorder/little_endian.h:28:52: warning:
    implicit declaration of function ‘___constant_swab32’;
    did you mean ‘__builtin_bswap32’? [-Wimplicit-function-declaration]
   #define __constant_cpu_to_be32(x) ((__force __be32)___constant_swab32((x)))

Declare all ___constant_swabXX() macros.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: armada-xp-theadorable.dts: Add I2C DT nodes
Stefan Roese [Thu, 18 Nov 2021 08:19:39 +0000 (09:19 +0100)]
arm: mvebu: armada-xp-theadorable.dts: Add I2C DT nodes

Now with DM I2C support enabled we need to describe the I2C busses and
devices in the DT.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: axp: ddr: Switch to using DM I2C API
Stefan Roese [Thu, 18 Nov 2021 08:19:38 +0000 (09:19 +0100)]
arm: mvebu: axp: ddr: Switch to using DM I2C API

No functional change intended. This patch switches from the legacy I2C
API to the DM I2C API, so that this code can be used with DM I2C
enabled.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: axp: Remove unreferenced ddr3_get_eprom_fabric() function
Stefan Roese [Thu, 18 Nov 2021 08:19:37 +0000 (09:19 +0100)]
arm: mvebu: axp: Remove unreferenced ddr3_get_eprom_fabric() function

This function is not referenced in mainline U-Boot. Let's remove now.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: axp/high_speed_env_lib: Switch to DM_I2C API
Stefan Roese [Thu, 18 Nov 2021 08:19:36 +0000 (09:19 +0100)]
arm: mvebu: axp/high_speed_env_lib: Switch to DM_I2C API

After all Armada XP boards have been switched over from legacy I2C
support to DM I2C, let's now also convert this serdes code to use
the DM I2C API.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: theadorable: Switch to using DM I2C API
Stefan Roese [Thu, 18 Nov 2021 08:19:35 +0000 (09:19 +0100)]
arm: mvebu: theadorable: Switch to using DM I2C API

No functional change intended. This patch switches from the legacy I2C
API to the DM I2C API, so that this code can be used with DM I2C
enabled.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: maxbcm_defconfig: Enable DM_I2C
Stefan Roese [Thu, 18 Nov 2021 08:19:34 +0000 (09:19 +0100)]
arm: mvebu: maxbcm_defconfig: Enable DM_I2C

Move to the DM I2C version, so that this board will not get dropped from
mainline.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: ds414_defconfig: Enable DM_I2C
Stefan Roese [Thu, 18 Nov 2021 08:19:33 +0000 (09:19 +0100)]
arm: mvebu: ds414_defconfig: Enable DM_I2C

Move to the DM I2C version, so that this board will not get dropped from
mainline.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Phil Sutter <phil@nwl.cc>
2 years agoarm: mvebu: db-mv784mp-gp_defconfig: Enable DM_I2C
Stefan Roese [Thu, 18 Nov 2021 08:19:32 +0000 (09:19 +0100)]
arm: mvebu: db-mv784mp-gp_defconfig: Enable DM_I2C

Move to the DM I2C version, so that this board will not get dropped from
mainline.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: theadorable_debug_defconfig: Enable DM_I2C
Stefan Roese [Thu, 18 Nov 2021 08:19:31 +0000 (09:19 +0100)]
arm: mvebu: theadorable_debug_defconfig: Enable DM_I2C

Move to the DM I2C version, so that this board will not get dropped from
mainline.

Signed-off-by: Stefan Roese <sr@denx.de>
2 years agoMerge tag 'efi-2022-01-rc4-3' of https://source.denx.de/u-boot/custodians/u-boot-efi
Tom Rini [Sat, 18 Dec 2021 19:39:21 +0000 (14:39 -0500)]
Merge tag 'efi-2022-01-rc4-3' of https://source.denx.de/u-boot/custodians/u-boot-efi

Pull request for efi-2022-01-rc4-3

Documentation:

* add Calxeda Highbank/Midway board documentation

Bug fixes:

* call part_init() in blk_get_device_by_str() only for MMC
* fix an 'undefined' error in some driver model macros

2 years agodm: fix an 'undefined' error in some macros
AKASHI Takahiro [Fri, 10 Dec 2021 06:49:36 +0000 (15:49 +0900)]
dm: fix an 'undefined' error in some macros

Due to a non-existing parameter name in macro's, use of those macro's will
cause a compiler error of "undefined reference".
Unfortunately, dm test doesn't fail because a wrong name ("&dev", hence it
is accidentally a valid name in the context of a caller site) is passed on.

Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Fixes: f262d4ca4b2b ("dm: core: Add a way to read platdata for all
child devices")
Fixes: 903e83ee8464 ("dm: core: Add a way to iterate through children,
probing each")
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agoblk: add a helper function, blk_probe_or_unbind()
AKASHI Takahiro [Fri, 10 Dec 2021 06:49:29 +0000 (15:49 +0900)]
blk: add a helper function, blk_probe_or_unbind()

This function will be commonly used in block device drivers
in the succeeding patches.

Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agopart: call part_init() in blk_get_device_by_str() only for MMC
AKASHI Takahiro [Fri, 10 Dec 2021 06:49:28 +0000 (15:49 +0900)]
part: call part_init() in blk_get_device_by_str() only for MMC

In blk_get_device_by_str(), the comment says: "Updates the partition table
for the specified hw partition."
Since hw partition is supported only on MMC, it makes no sense to do so
for other devices.

Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agodoc: board: Add Calxeda Highbank/Midway documentation
Andre Przywara [Tue, 14 Dec 2021 17:47:00 +0000 (17:47 +0000)]
doc: board: Add Calxeda Highbank/Midway documentation

The Calxeda servers are using U-Boot as the primary bootloader, which
was shipped as part of a firmware upgrade package.
Even though the machines are considered legacy at this point, the port
still works, so deserves some documentation.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agoMerge tag 'dm-pull-17dec21' of https://source.denx.de/u-boot/custodians/u-boot-dm...
Tom Rini [Fri, 17 Dec 2021 23:18:15 +0000 (18:18 -0500)]
Merge tag 'dm-pull-17dec21' of https://source.denx.de/u-boot/custodians/u-boot-dm into next

binman support for ATF FIP
fdtgrep fixes for empty devicetree

2 years agofdtgrep: Handle an empty output tree
Simon Glass [Wed, 8 Dec 2021 16:55:35 +0000 (09:55 -0700)]
fdtgrep: Handle an empty output tree

In strange cases it is possible for fdtgrep to find nothing to output.
Typically this means that the resulting SPL device tree is not going to
allow anything to boot, but at present the tree is actually invalid,
since it only has an END tag in the struct region.

The FDT spec requires at least a root node. So add a special case to
include at least this, if the FDT_REG_SUPERNODES flag is set.

This ensures that grepping an empty tree still produces a valid tree.

Also add comments to the enum since it is not completely obvious from
the names now.

The typical symptom of this problem is a message from binman:

   pylibfdt error -11: FDT_ERR_BADSTRUCTURE

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agofdtgrep: Correct alignment of struct section
Simon Glass [Wed, 8 Dec 2021 16:55:34 +0000 (09:55 -0700)]
fdtgrep: Correct alignment of struct section

When outputting a devicetree we should not align the struct section to a
16-byte boundary. The normal position is fine, which is 8-byte aligned.

This avoids leaving adding 8 extra zero bytes in the output tree in the
case where the reserved section is empty (i.e has 16 zero bytes).

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoiot2050: binman: add missing-msg for blobs
Ivan Mikhaylov [Thu, 9 Dec 2021 16:10:53 +0000 (16:10 +0000)]
iot2050: binman: add missing-msg for blobs

Add the 'missing-msg' for blobs for more detailed output on missing system
firmware and SEBoot blobs.

Signed-off-by: Ivan Mikhaylov <ivan.mikhaylov@siemens.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Fix minor typos:
Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Use less hard coded magic when inserting new PATH
Andy Shevchenko [Mon, 6 Dec 2021 11:44:13 +0000 (14:44 +0300)]
binman: Use less hard coded magic when inserting new PATH

Instead of joining hard coded '..' to the run-time path of the executable,
take just a dirname out of it. Besides that, use $(srctree) where it makes
sense.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
2 years agobinman: Do not pollute source tree when build with `make O=...`
Andy Shevchenko [Mon, 6 Dec 2021 11:44:12 +0000 (14:44 +0300)]
binman: Do not pollute source tree when build with `make O=...`

Importing libraries in Python caches the bytecode by default.
Since we run scripts in source tree it ignores the current directory
settings, which is $(srctree), and creates cache just in the middle
of the source tree. Move cache to the current directory.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
2 years agoConvert CONFIG_PHYSMEM to Kconfig
Simon Glass [Wed, 24 Nov 2021 16:26:42 +0000 (09:26 -0700)]
Convert CONFIG_PHYSMEM to Kconfig

This converts the following to Kconfig:
   CONFIG_PHYSMEM

Signed-off-by: Simon Glass <sjg@chromium.org>
Suggested-by: Sean Anderson <seanga2@gmail.com>
2 years agobinman: Add support for ATF FIP
Simon Glass [Wed, 24 Nov 2021 04:08:59 +0000 (21:08 -0700)]
binman: Add support for ATF FIP

This format is used in firmware binaries so we may as well supported it.

With this patch binman supports creating, listing and updating FIPs, as
well as extracting files from one, provided that an FDTMAP is also present
somewhere in the image.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Add a utility module for ATF FIP
Simon Glass [Wed, 24 Nov 2021 04:08:58 +0000 (21:08 -0700)]
binman: Add a utility module for ATF FIP

Add support for this format which is used by ARM Trusted Firmware to find
firmware binaries to load.

FIP is like a simpler version of FMAP but uses a UUID instead of a name,
for each entry.

It supports reading a FIP, writing a FIP and parsing the ATF source code
to get a list of supported UUIDs.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoMerge tag 'u-boot-at91-2022.04-a' of https://source.denx.de/u-boot/custodians/u-boot...
Tom Rini [Fri, 17 Dec 2021 12:25:34 +0000 (07:25 -0500)]
Merge tag 'u-boot-at91-2022.04-a' of https://source.denx.de/u-boot/custodians/u-boot-at91 into next

First set of u-boot-at91 features for the 2022.04 cycle:

This feature set includes : support for the new QSPI hardware on
sama7g5, small fixes on sam9x60 and sama7g5, some additions of commands
and PIO controller on sam9x60/sam9x60ek.

2 years agoMerge commit '4720b83d2c711062cfb55f03591b8f12c897d7cb' of https://github.com/tienfon...
Tom Rini [Fri, 17 Dec 2021 12:24:56 +0000 (07:24 -0500)]
Merge commit '4720b83d2c711062cfb55f03591b8f12c897d7cb' of https://github.com/tienfong/uboot_mainline

2 years agoarm: socfpga: arria10: Enable double peripheral RBF configuration
Tien Fong Chee [Sun, 7 Nov 2021 15:08:56 +0000 (23:08 +0800)]
arm: socfpga: arria10: Enable double peripheral RBF configuration

Double peripheral RBF configuration are needed on some devices or boards
to stabilize the IO configuration system.

Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
Signed-off-by: Sin Hui Kho <sin.hui.kho@intel.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@intel.com>
2 years agoarm: socfpga: arria10: Reset MPFE NoC after program periph / combined RBF
Tien Fong Chee [Sun, 7 Nov 2021 15:08:55 +0000 (23:08 +0800)]
arm: socfpga: arria10: Reset MPFE NoC after program periph / combined RBF

This patch triggers warm reset to recover the MPFE NoC from corruption
due to high frequency transient clock output from HPS EMIF IOPLL at
VCO startup after peripheral RBF is programmed.

Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
Signed-off-by: Sin Hui Kho <sin.hui.kho@intel.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@intel.com>
2 years agoarm: socfpga: arria10: Setting image magic value to romcode initswstate reg
Tien Fong Chee [Sun, 7 Nov 2021 15:08:54 +0000 (23:08 +0800)]
arm: socfpga: arria10: Setting image magic value to romcode initswstate reg

The romcode_initswstate register need to be set with FSBL_IMAGE_IS_VALID
value if the current FSBL image is found valid, otherwise BootROM will
look for next subsequent valid FSBL image when warm reset is triggered.

Signed-off-by: Tien Fong Chee <tien.fong.chee@intel.com>
Signed-off-by: Sin Hui Kho <sin.hui.kho@intel.com>
Reviewed-by: Tien Fong Chee <tien.fong.chee@intel.com>
2 years agoARM: mach-at91: Add compile time option to choose proper timer
Hari Prasath [Mon, 13 Dec 2021 05:00:03 +0000 (10:30 +0530)]
ARM: mach-at91: Add compile time option to choose proper timer

New SoC's of AT91 family with ARM-9 core includes a regular timer and a 64-bit
timer.This patch adds a compile time option to the Makefile such that the old
timer driver is chosen and compiled as default if none of timer configuration
options are explicitly defined in the board configs.

Signed-off-by: Hari Prasath <Hari.PrasathGE@microchip.com>
Reviewed-by: Eugen Hristev <eugen.hristev@microchip.com>
2 years agoMerge tag 'clk-2022.01-rc3' of https://source.denx.de/u-boot/custodians/u-boot-clk
Tom Rini [Wed, 15 Dec 2021 19:51:44 +0000 (14:51 -0500)]
Merge tag 'clk-2022.01-rc3' of https://source.denx.de/u-boot/custodians/u-boot-clk

Clock patches for v2022.01-rc3

This adds better logging support for many CCF drivers, and clarifies some
documentation regarding clk_get_rate.

2 years agoclk: define LOG_CATEGORY for generic and ccf clocks
Patrick Delaunay [Fri, 19 Nov 2021 14:12:07 +0000 (15:12 +0100)]
clk: define LOG_CATEGORY for generic and ccf clocks

Define LOG_CATEGORY to allow filtering with log command
for generic clock and CCF clocks.

This patch also change existing printf, debug and pr_ macro
to log_ or dev_ macro.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
2 years agoclk: cosmetic: reorder include files
Patrick Delaunay [Fri, 19 Nov 2021 14:12:06 +0000 (15:12 +0100)]
clk: cosmetic: reorder include files

Reorder include files in the U-Boot expected order:

the common.h header should always be first,
followed by other headers in order,
then headers with directories,
then local files.

It is a preliminary step for next patch.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
2 years agoclk: fix clk_get_rate() documentation
Giulio Benetti [Sun, 14 Feb 2021 02:17:18 +0000 (03:17 +0100)]
clk: fix clk_get_rate() documentation

Improve clk_get_rate() @return documentation that otherwise is a bit
ambiguous. At the moment I expect to return 0 as error since the return
type is 'ulong', instead the function really returns negative value in
case the corresponding function pointer is null and returns 0 if the clock
is invalid.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>