AKASHI Takahiro [Wed, 11 Oct 2023 10:07:03 +0000 (19:07 +0900)]
test: dm: simplify SCMI unit test on sandbox
Adding SCMI base protocol makes it inconvenient to hold the agent instance
(udevice) locally since the agent device will be re-created per each test.
Just remove it and simplify the test flows.
The test scenario is not changed at all.
AKASHI Takahiro [Wed, 11 Oct 2023 10:07:01 +0000 (19:07 +0900)]
firmware: scmi: add a version check against base protocol
In SCMI base protocol version 2 (0x20000), new interfaces,
BASE_SET_DEVICE_PERMISSIONS/BASE_SET_PROTOCOL_PERMISSIONS/
BASE_RESET_AGENT_CONFIGURATION, were added. Moreover, the api of
BASE_DISCOVER_AGENT was changed to support self-agent discovery.
So the driver expects SCMI firmware support version 2 of base protocol.
AKASHI Takahiro [Wed, 11 Oct 2023 10:07:00 +0000 (19:07 +0900)]
firmware: scmi: implement SCMI base protocol
SCMI base protocol is mandatory according to the SCMI specification.
With this patch, SCMI base protocol can be accessed via SCMI transport
layers. All the commands, except SCMI_BASE_NOTIFY_ERRORS, are supported.
This is because U-Boot doesn't support interrupts and the current transport
layers are not able to handle asynchronous messages properly.
AKASHI Takahiro [Wed, 11 Oct 2023 10:06:55 +0000 (19:06 +0900)]
firmware: scmi: use a protocol's own channel if assigned
SCMI specification allows any protocol to have its own channel for
the transport. While the current SCMI driver may assign its channel
from a device tree, the core function, devm_scmi_process_msg(), doesn't
use a protocol's channel, but always use an agent's channel.
With this commit, devm_scmi_process_msg() tries to find and use
a protocol's channel. If it doesn't exist, use an agent's.
AKASHI Takahiro [Wed, 11 Oct 2023 10:06:54 +0000 (19:06 +0900)]
scmi: refactor the code to hide a channel from devices
The commit 85dc58289238 ("firmware: scmi: prepare uclass to pass channel
reference") added an explicit parameter, channel, but it seems to make
the code complex.
Hiding this parameter will allow for adding a generic (protocol-agnostic)
helper function, i.e. for PROTOCOL_VERSION, in a later patch.
Tom Rini [Fri, 13 Oct 2023 12:45:55 +0000 (08:45 -0400)]
Merge tag 'xilinx-for-v2024.01-rc1-v3' of https://source.denx.de/u-boot/custodians/u-boot-microblaze
Xilinx changes for v2024.01-rc1 v3
clk:
- remove additional compatible strings for Versal NET
net:
- zynq_gem: Fix clock calculation for MDC for higher frequencies
pinctrl:
- core: Extend pinmux status buffere size
- zynqmp driver: Show also tristate configuration
test:
- add test case for pxe get
Xilinx:
- describe SelectMAP boot mode
Zynq:
- Fix nand description in DT
ZynqMP:
- DTS sync patches with kernel and also W=1 related fixes
- Add support for KD240, zcu670, e-a2197 with x-prc cards, SC revB/C with i2c
description for other SC based boards
- k24 psu_init cleanup
The first 5 patches are from Mikhail Kshevetskiy, aligning the mtd spinand
core with kernel version 5.15.43, fixing a bug on Winbond, and adding
support for Winbond W25NO2KV.
The other 2 patches are from Patrick Delaunay and they fix a bug and mark
bad the MTD block on erase error.
Artur Rojek [Wed, 13 Sep 2023 15:00:58 +0000 (17:00 +0200)]
usb: gadget: atmel: fix transfer of queued requests
In the existing implementation, multiple requests queued up on an
endpoint are subject to getting evicted without transmission.
For both control and bulk endpoints, their respective logic found in
usba_control_irq()/usba_ep_irq() guarantees that TX FIFO is empty before
data is sent out, and that request_complete() gets called once the
transaction has been finished. At this point however, if any additional
requests are found on the endpoint queue, they will be processed by
submit_next_request(), which makes no checks against the above
conditions, trashing data on a busy FIFO and neglecting completion
handlers.
Fix the above issues by removing the calls to submit_next_request(),
and thus forcing the pending requests to be processed on the next pass
of the respective endpoint logic. While at it, remove a DBG message, as
that branch becomes part of regular flow.
This restores mass storage mode operation on Microchip ATSAMA5D27 SoC.
In the MTD DFU backend, it is needed to mark the NAND block bad when the
erase failed with the -EIO error, as it is done in UBI and JFFS2 code.
This operation is not done in the MTD framework, but the bad block
tag (in BBM or in BBT) is required to avoid to write data on this block
in the next DFU_OP_WRITE loop in mtd_block_op(): the code skip the bad
blocks, tested by mtd_block_isbad().
Without this patch, when the NAND block become bad on DFU write operation
- low probability on new NAND - the DFU write operation will always failed
because the failing block is never marked bad.
This patch also adds a test to avoid to request an erase operation on a
block already marked bad; this test is not performed in MTD framework
in mtd_erase().
Reviewed-by: Michael Trimarchi <michael@amarulasolutions.com> Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
The offset variable = 'off' used in the error trace when limit is reach
on erase operation is incorect as 'erase_op.addr' is used in the loop.
This patch corrects the copy paste issue between the erase loop and
the write loop.
This patch also adds the 'remaining' information to allow to debug of
limit issues.
Fixes: 6015af28ee6d ("dfu: add backend for MTD device") Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com> Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
Winbond uses 3 bytes to identify flash: vendor_id, dev_id_0, dev_id_1,
but current driver uses only first 2 bytes of it for devices
identification. As result Winbond W25N02KV flash (id_bytes: EF, AA, 22)
is identified as W25N01GV (id_bytes: EF, AA, 21).
mtd/spinand: sync supported devices with linux-5.15.43
This adds more supported spinand devices from the Linux kernel
implementation.
This does not include the latest kernel implementation as this would
require a substantial amount of extra work due to the missing
ECC engine abstraction layer in U-Boot.
mtd/spinand: sync core spinand code with linux-5.10.118
This brings us closer to the current Linux kernel implementation of
the spinand core and makes backporting features and fixes easier.
This does not include the latest kernel implementation as this would
require a substantial amount of extra work due to the missing
ECC engine abstraction layer in U-Boot.
mtd/spinand: rework detect procedure for different READ_ID operation
Currently there are 3 different variants of read_id implementation:
1. opcode only. Found in GD5FxGQ4xF.
2. opcode + 1 addr byte. Found in GD5GxGQ4xA/E
3. opcode + 1 dummy byte. Found in other currently supported chips.
Original implementation was for variant 1 and let detect function
of chips with variant 2 and 3 to ignore the first byte. This isn't
robust:
1. For chips of variant 2, if SPI master doesn't keep MOSI low
during read, chip will get a random id offset, and the entire id
buffer will shift by that offset, causing detect failure.
2. For chips of variant 1, if it happens to get a devid that equals
to manufacture id of variant 2 or 3 chips, it'll get incorrectly
detected.
This patch reworks detect procedure to address problems above. New
logic do detection for all variants separatedly, in 1-2-3 order.
Since all current detect methods do exactly the same id matching
procedure, unify them into core.c and remove detect method from
manufacture_ops.
This is a rework of Chuanhong Guo <gch981213@gmail.com> patch
submitted to linux kernel
TRM for J721S2/AM68: https://www.ti.com/lit/pdf/spruj28
Product Page for J721S2: https://www.ti.com/tool/J721S2XSOMXEVM
Product Page for AM68: https://www.ti.com/tool/SK-AM68
Udit Kumar [Fri, 6 Oct 2023 04:45:55 +0000 (10:15 +0530)]
configs: j721s2_evm_r5_defconfig: Increase malloc pool size in DRAM
The malloc capacity in DRAM at R5 SPL is set to 1MB which isn't
sufficient to load the new tispl.bin to
enable loading of tispl.bin the size is increased by 256KB to 1.25MB.
Reid Tonking [Thu, 5 Oct 2023 18:12:58 +0000 (13:12 -0500)]
arm: dts: j7200: dts sync with Linux 6.6-rc1
Sync j7200 dts with Linux 6.6-rc1
- k3-j7200-r5-common-proc-board.dts now inherits from
k3-j7200-common-proc-board.dts instead of k3-j7200-som-p0.dtsi. This
allows us to trim down the r5 file considerably by using existing
properties
- remove pimux nodes from r5 file
- remove duplicate nodes & node properties from r5/u-boot files
- mcu_timer0 now used instead of timer1
mcu_timer0 device id added to dev-data.c file in order to work
- remove cpsw node
This node is no longer required since the compatible is now fixed
- remove dummy_clock_19_2_mhz
This node wasn't being used anyhere, so it was removed
- remove dummy_clock_200mhz
main_sdhci0 & main_sdhci1 no longer need dummy clock for eMMC/SD
- fix secure proxy node
mcu_secproxy changed to used secure_prxy_mcu which is already
defined in k3-j7200-mcu-wakeup.dtsi
- removed &mcu_ringacc property override since they're present in
v6.6-rc1
Roger Quadros [Thu, 5 Oct 2023 13:06:42 +0000 (16:06 +0300)]
board: ti: am64x: Switch to standard boot flow
Switch to using bootstd. Note with this change, we will stop using
distro_bootcmd and instead depend entirely on bootflow method of
starting the system up.
Drop header files that are no longer needed in am64x_evm.h.
k3_dfu.h is available via k3_dfu.env in am64x.env.
Drop unused macro CFG_SYS_SDRAM_BASE1.
Signed-off-by: Roger Quadros <rogerq@kernel.org> Reviewed-by: Nishanth Menon <nm@ti.com>
To quote the author:
So far expo only supports menus. These are quite flexible for various
kinds of settings, but cannot deal with free-form input, such as a
serial number or a machine name.
This series adds support for a textline object, which is a single line
of text. It has a maximum length and its value is stored within the expo
structure.
U-Boot already has a command-line editor which provides most of the
features needed by expo. But the code runs in its own loop and only
returns when the line is finished. This is not suitable for expo, which
must handle a keypress at a time, returning to its caller after each
one.
In order to use the CLI code, some significant refactoring is included
here. This mostly involves moving the internal loop of the CLI to a
separate function and recording its state in a struct, just as was done
for single keypresses some time back. A minor addition is support for
Ctrl-W to delete a word, since strangely this is currently only present
in the simple version.
The video-console system provides most of the features needed by
testline, but a few things are missing. This series provides:
- primitive cursor support so the user can see where he is typing
- saving and restoring of the text-entry context, so that expo can allow
the user to continue where he left off, including deleting previously
entered characters correctly (for Truetype)
- obtaining the nominal width of a string of n characters, so that a
suitable width can be chosen for the textline object
Note that no support is provided for clearing the cursor. This was
addressed in a previous series[1] which could perhaps be rebased. For
this implementation, the cursor is therefore not enabled for the normal
command line, only for expo.
Reading and writing textline objects is supported for FDT and
environment, but not for CMOS RAM, since it would likely use too much
RAM to store a string.
In terms of code size, the overall size increase is 180 bytes for
Thumb02 boards, 160 of whcih is the addition of Ctrl-W to delete a word.
Masahisa Kojima [Thu, 12 Oct 2023 05:31:59 +0000 (14:31 +0900)]
efi_loader: use well-known guid for auto-created boot option
The boot option automatically created by efibootmgr is identified
by the special guid appended in the optional data of boot option.
The same mechanism is implemented in the EDK II reference
implementation, it uses the different guid from the one currently
U-Boot uses.
The guid indicating auto-created boot option is not defined in the
UEFI specification, but some userspace tools such as 'efivar' package
are aware of the guid used in EDK II as auto-created boot option.
So let's use the same guid as EDK II reference implementation.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org> Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
The BitBlt test leaves the serial console output in disarray.
* Call ClearScreen() where needed.
* Test CheckEvent() for EFI_SIMPLE_TEXT_INPUT_PROTOCOL.WaitForKey
via adding navigation keys
* Correct timer comment
Paul Barker [Mon, 9 Oct 2023 11:11:36 +0000 (12:11 +0100)]
doc: Migrate Renesas board docs to rst
Some of the information in README.rmobile is obsolete, references
defconfigs which no longer exist in u-boot or has broken links. The
information which is still relevant is moved into the reStructuredText
documentation under `doc/board/renesas`, and `doc/README.rmobile` is
dropped.
The list of boards in `doc/board/renesas` is converted into a table so
it's easier to see which defconfig to use. The list is expanded based on
reviewing the current u-boot code and the contents of the eLinux wiki
[1] [2].
Simon Glass [Sun, 8 Oct 2023 21:35:56 +0000 (15:35 -0600)]
doc: Add a short intro video
This video covers the basics in a short time, so add a link to it.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Paul Barker <paul.barker.ct@bp.renesas.com> Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Weizhao Ouyang [Sat, 7 Oct 2023 10:52:36 +0000 (10:52 +0000)]
cyclic: doc: Update documentation for CONFIG_CYCLIC_MAX_CPU_TIME_US
Cyclic now just print a warning once instead of disabling the cyclic
function when the cyclic function upon exceeding CPU time usage.
Fixes: ddc8d36a7455 ("cyclic: Don't disable cylic function upon exceeding CPU time") Signed-off-by: Weizhao Ouyang <o451686892@gmail.com>
Rephrase the paragraph. Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Tom Rini [Thu, 5 Oct 2023 16:27:38 +0000 (12:27 -0400)]
sphinx: Bump urllib3 version
While not a direct issue for us, urllib3 before 1.26.17 is vulnerable to
CVE-2023-43804 to bump our version up.
Reported-by: GitHub dependabot Signed-off-by: Tom Rini <trini@konsulko.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Use urllib3 2.0.6 Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Simon Glass [Mon, 2 Oct 2023 01:13:34 +0000 (19:13 -0600)]
expo: Plumb in textlines to a scene
Provide an implementation for textlines in the scene code, so that they
are displayed correctly. Provide a way to have a border around the
textline, with the internal part being the same colour as the
background. This looks more natural.
Simon Glass [Mon, 2 Oct 2023 01:13:31 +0000 (19:13 -0600)]
expo: Add some scene fields needed for text entry
Add the CLI state, a buffer to hold the old value of the text being
edited and a place to save vidconsole entry context. These will be use
by the textline object.
Set an upper limit on the maximum number of characters in a textline
object supported by expo, at least for now.
Simon Glass [Mon, 2 Oct 2023 01:13:30 +0000 (19:13 -0600)]
expo: Allow rendering the background of any object
So far only menus have a background. When other object types are
rendered, they may have a background too. Make this code more generic
so it will be usable by new object types.
Simon Glass [Mon, 2 Oct 2023 01:13:27 +0000 (19:13 -0600)]
expo: Allow highlighting other scene-object types
So far only menus can be highlighted. With the coming addition of
text lines we need to be able to highlight other objects. Add a function
to determine whether an object can be highlighted.
Simon Glass [Mon, 2 Oct 2023 01:13:24 +0000 (19:13 -0600)]
expo: Use switch statements more for object types
In a lot of cases menus are the only objects which are have their own
behaviour in the cedit, e.g. to move between menus. With expo expanding
to support text, this is no-longer true.
Use a switch() statement so that we can simply insert a new 'case' for
the new object types.
Simon Glass [Mon, 2 Oct 2023 01:13:22 +0000 (19:13 -0600)]
expo: Add better error reporting
When building an expo fails, show some information about which node
caused the problem. Use -ENOENT consistently when the ID is missing.
This makes it easier for the user to debug things.
Simon Glass [Mon, 2 Oct 2023 01:13:21 +0000 (19:13 -0600)]
video: Support showing a cursor
Add rudimentary support for displaying a cursor on a vidconsole. This
helps the user to see where text is being entered.
The implementation so far is very simple: the cursor is just a vertical
bar of fixed width and cannot be erased. To erase the cursor, the text
must be redrawn over it.
This is good enough for expo but will need enhancement to be useful for
the command-line console. For example, it could save and restore the
area behind the cursor.
For now, enable this only for expo, to reduce code size.
Simon Glass [Mon, 2 Oct 2023 01:13:19 +0000 (19:13 -0600)]
video: Allow saving and restoring text-entry state
Text entry operates within a context which includes quite a bit of
information. For example, with Truetype fonts, each character in the
text string has a position stored, so that it is possible to
backspace to that character. This information is built up as strings
are drawn on the display.
For the command line, there is just a single context. It is created
when command-line entry starts and it is destroyed (or at least not
needed anymore) when the user presses <enter> to enter the command.
By contrast, expo needs to be able to switch in and out of a text-entry
context, since it is also displaying other objects in the scene.
Add a way to save and restore the entry context for a vidconsole. This
is only needed for the truetype vidconsole, so add a method for that,
storing the information in an abuf struct.
Simon Glass [Mon, 2 Oct 2023 01:13:18 +0000 (19:13 -0600)]
video: Allow obtaining the nominal size of a string size
At present there is a method for measuring text, but if the actual text
string is not known, it cannot be used.
For text editor we want to set the size of the entry box to cover the
expected text size. Add the concept of a 'norminal' size with a method
to calculate that for the vidconsole.
If the method is not implemented, fall back to using the font size,
which is sufficient for fixed-width fonts.
Simon Glass [Mon, 2 Oct 2023 01:13:14 +0000 (19:13 -0600)]
cli: Terminate the string in cread_line_process_ch()
Rather than relying on the caller, terminate the string inside this
function. Do this each time we return, whether input is finished or
not. It is not needed when the input is aborted, since the string will
be discarded in that case.
Simon Glass [Mon, 2 Oct 2023 01:13:11 +0000 (19:13 -0600)]
cli: Convert cread_line() to use a struct for the main vars
We want to reuse the editing code elsewhere. As a first step, move the
common variables into a struct. This will allow us to eventually put the
contents of the inner loop in a function, so it can be called from
elsewhere.
To quote the author:
I'm submitting a patch series that ports the gen_compile_commands.py
script from the Linux kernel's sources to U-Boot. This script,
originally located in scripts/clang-tools/gen_compile_commands.py,
enables the generation of compile_commands.json file for improved code
navigation and analysis. The series consists of the initial script
import, the necessary modifications for U-Boot compatibility, and
finally some documentation.
For now, this chapter is mostly a reference to the documentation of
gen_compile_commands, in doc/build, but it can be futurely used as
a guide for other IDE-friendly features.
Signed-off-by: Joao Marcos Costa <jmcosta944@gmail.com> Tested-by: Joao Paulo Goncalves <joao.goncalves@toradex.com>
Add acknowledgments for porting and modifying the script. Of course, the
license, author, and copyright notice remain the same as in the original
script.
Signed-off-by: Joao Marcos Costa <jmcosta944@gmail.com> Tested-by: Joao Paulo Goncalves <joao.goncalves@toradex.com>
scripts: Port Linux's gen_compile_commands.py to U-Boot
This script generates a database of compiler flags, namely
compile_commands.json. It is quite useful for text editors that use
clangd LSP (e.g. Vim, Neovim).
- Assorted TI K3 updates, use ".dtso" for device tree overlays to match
general usage, mkimage fixes/improvements, assorted platform
updates/fixes, other assorted driver/platform fixes.
Andrew Davis [Thu, 3 Aug 2023 14:54:41 +0000 (09:54 -0500)]
configs: Make TI_SECURE_DEVICE default for K3
All K3 boards now are secure by default, instead of setting this in each
defconfig, make it implied by the ARCH config.
The only exception is IOT2050, which I do not believe will have any
problems with being a TI_SECURE_DEVICE, but for now turn it off to keep
its config the same.
Signed-off-by: Andrew Davis <afd@ti.com> Tested-by: Tom Rini <trini@konsulko.com>
Andre Przywara [Sun, 1 Oct 2023 22:52:12 +0000 (23:52 +0100)]
ARM: psci: move GIC address override to Kconfig
As the code to switch an ARM core from secure to the non-secure state
needs to know the base address of the Generic Interrupt Controller
(GIC), we read an Arm Cortex defined system register that is supposed to
hold that base address. However there are SoCs out there that get this
wrong, and this CBAR register either reads as 0 or points to the wrong
address. To accommodate those systems, so far we use a macro defined in
some platform specific header files, for affected boards.
To simplify future extensions, replace that macro with a Kconfig variable
that holds this override address, and define a default value for SoCs
that need it.
Signed-off-by: Andre Przywara <andre.przywara@arm.com> Reviewed-by: Sam Edwards <CFSworks@gmail.com>
Sean Anderson [Sat, 30 Sep 2023 20:45:46 +0000 (16:45 -0400)]
misc: fs_loader: Fix alignment of fs_loader driver
DM_DRIVER_GET will redeclare the fs_loader driver without the correct
alignment. This causes GCC to use the default section alignment of 32
bytes. This in turn creates a gap in the linker list due to the padding
required to achieve the correct alignment, corrupting all further entries.
Use DM_DRIVER_REF instead, which doesn't redeclare anything.
Fixes: 0998a20cfc6 ("misc: fs_loader: Add function to get the chosen loader") Signed-off-by: Sean Anderson <seanga2@gmail.com>
The man page correctly said that -B was ignored without -E, while the
`mkimage -h` output suggested otherwise. Now that -B can actually be
used by itself, update the man page.
While at it, also amend the `mkimage -h` line to mention the
connection with -E.
The FDT header is a fixed 40 bytes, so its size cannot (and is not)
modified, while its alignment is a property of the address in RAM one
loads the FIT to, so not something mkimage can affect in any way. (In
the file itself, the header is of course at offset 0, which has all
possible alignments already.)
Reported-by: Sean Anderson <seanga2@gmail.com> Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Reviewed-by: Simon Glass <sjg@chromium.org>
Distinguish more clearly between source files meant for producing .dtb
from those meant for producing .dtbo. No functional change, as we
currently have rules for producing a foo.dtbo from either foo.dts or
foo.dtso.
Note that in the linux tree, all device tree overlay sources have been
renamed to .dtso, and the .dts->.dtbo rule is gone since v6.5 (commit 81d362732bac). So this is also a step towards staying closer to linux
with respect to both Kbuild and device tree sources.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Reviewed-by: Simon Glass <sjg@chromium.org>
Distinguish more clearly between source files meant for producing .dtb
from those meant for producing .dtbo. No functional change, as we
currently have rules for producing a foo.dtbo from either foo.dts or
foo.dtso.
Note that in the linux tree, all device tree overlay sources have been
renamed to .dtso, and the .dts->.dtbo rule is gone since v6.5 (commit 81d362732bac). So this is also a step towards staying closer to linux
with respect to both Kbuild and device tree sources.
projects / u-boot.git / log