Marek Vasut [Mon, 14 Nov 2022 21:49:59 +0000 (22:49 +0100)]
cmd: fdt: Fix iteration over elements above index 1 in fdt get
Always increment both the iterator and pointer into the string
property value by length of the current element + 1 (to cater
for the string delimiter), otherwise the element extracted from
the string property value would be extracted from an offset that
is multiple of the length of the first element, instead of sum
of element lengths until select index.
This fixes 'fdt get value' operation for index above 1 (counting
from index 0).
Acked-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com> Fixes: 13982ced2cc ("cmd: fdt: Add support for reading stringlist property values") Signed-off-by: Marek Vasut <marex@denx.de>
Tom Rini [Thu, 10 Nov 2022 02:14:50 +0000 (19:14 -0700)]
global: Do not default to faking missing binaries for buildman
While it is possible and documented on how to re-run buildman to replace
faked required binary files after the fact, this behavior ends up being
more confusing than helpful in practice. Switch to requiring
BINMAN_ALLOW_MISSING=1 to be passed on the 'make' line to enable this
behavior.
Cc: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Cc: Simon Glass <sjg@chromium.org> Signed-off-by: Tom Rini <trini@konsulko.com> Signed-off-by: Simon Glass <sjg@chromium.org>
Simon Glass [Thu, 10 Nov 2022 02:14:49 +0000 (19:14 -0700)]
binman: Add a separate section about environment variables
These are documented in various several sections. Add a new section that
mentions them all in one place so it is easier to see what environment
variables can be used to control U-Boot's use of binman.
Signed-off-by: Simon Glass <sjg@chromium.org> Suggested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Simon Glass [Thu, 10 Nov 2022 02:14:48 +0000 (19:14 -0700)]
buildman: Detect binman reporting missing blobs
Buildman should consider a build as a success (with warnings) if missing
blobs have been dealt with by binman, even though buildman itself returns
and error code overall. This is how other warnings are dealt with.
We cannot easily access the 103 exit code, so detect the problem in the
output.
With this change, missing blobs result in an exit code of 101, although
they still indicate failure.
Simon Glass [Thu, 10 Nov 2022 02:14:47 +0000 (19:14 -0700)]
buildman: Drop mention of old architectures
Support for some architectures has been removed since buildman was first
written. Also all toolchains are now available at kernel.org so we don't
need the links, except for arc where the kernel.org toolchain fails to
build all boards.
Simon Glass [Thu, 10 Nov 2022 02:14:42 +0000 (19:14 -0700)]
binman: Use an exit code when blobs are missing
At present binman returns success when told to handle missing/faked blobs
or missing bintools. This is confusing since in fact the resulting image
cannot work.
Use exit code 103 to signal this problem, with a -W option to convert
it to a warning.
Rename the flag to --ignore-missing since it controls bintools also.
Add documentation about exit codes while we are here.
Simon Glass [Thu, 10 Nov 2022 02:14:40 +0000 (19:14 -0700)]
Makefile: Correct the binman rule
This currently uses if_changed on a phony target. Use a real file as the
target and add FORCE at the end, as required. Drop the 'inputs' phony
since it is not needed.
Quentin Schulz [Mon, 7 Nov 2022 12:54:54 +0000 (13:54 +0100)]
binman: bintool: remove btool_ prefix from btool names
The binary is looked on the system by the suffix of the packer class.
This means binman was looking for btool_gzip on the system and not gzip.
Since a btool can have its btool_ prefix missing but its module and
binary presence on the system appropriately found, there's no need to
actually keep this prefix after listing all possible btools, so let's
remove it.
This fixes gzip btool by letting Bintool.find_bintool_class handle the
missing prefix and still return the correct class which is then init
with gzip name instead of btool_gzip.
Additionally, there was an issue with the cached module global variable.
The variable only stores the module and not the associated class name
when calling find_bintool_class.
This means that when caching the module on the first call to
find_bintool_class, class_name would be set to Bintoolbtool_gzip but the
module_name gzip only, adding the module in the gzip key in the module
dictionary. When hitting the cache on next calls, the gzip key would be
found, so its value (the module) is used. However the default class_name
(Bintoolgzip) is used, failing the getattr call.
Instead, let's enforce the same class name: Bintool<packer>, whatever
the filename it is contained in.
Tom Rini [Tue, 22 Nov 2022 17:33:48 +0000 (12:33 -0500)]
Merge tag 'xilinx-for-v2023.01-rc3' of https://source.denx.de/u-boot/custodians/u-boot-microblaze
Xilinx changes for v2023.01-rc3
microblaze:
- Enable 32 bit addressing mode for SPIs
zynq:
- Minor DT fixes (PL clock enabling)
zynqmp:
- Disable watchdog by default
- Remove unused xlnx,eeprom chosen support
- Add missing symlink for vck190 SC revB
- Use mdio bus with ethernet-phy-id description
versal:
- Add mini qspi/ospi configuration
versal-net:
- Add soc driver
- Fix Kconfig entry for SOC
- Fix loading address location for MINI configuration
- Disable LMB for mini configuration
net:
- Fix ethernet-phy-id usage in the code
pinctrl:
- Revert high impedance/output enable support
timer:
- Fix timer relocation for Microblaze
- Fix timer wrap in 32bit Xilinx timer driver
arm64: versal: Add octal spi flash mini u-boot configuration
Add configuration file for mini u-boot configuration which runs on a
smaller footprint from on chip memory(OCM). This configuration has
required CONFIG's enabled to support octal spi flash and uses DCC terminal
for console output. Add required dts for octal spi flash mini u-boot
configuration.
spi: cadence-qspi: Fix compilation error in mini u-boot flash reset
When cadence_qspi_versal_flash_reset() function is called in mini
u-boot where there is no firmware support, it is missing defines for
macro's BOOT_MODE_POR_0 & BOOT_MODE_POR_1. Remove them and replace with
already define macro's which have same values as these.
arm64: versal: Add qspi flash mini u-boot configuration
Add configuration file for mini u-boot configuration which runs on a
smaller footprint from on chip memory(OCM). This configuration has
required CONFIG's enabled to support qspi flash and uses DCC terminal
for console output. Add required dts files for qspi mini configuration.
Add soc_xilinx_versal_net driver to identify the family & revision of
versal-net SoC. Add Kconfig option CONFIG_SOC_XILINX_VERSAL_NET to
enable/disable this driver. To enable this driver by default, add this
config to xilinx_versal_net_virt_defconfig file. This driver will be
probed using platdata U_BOOT_DEVICE structure which is specified in
mach-versal-net/cpu.c.
Michal Simek [Wed, 16 Nov 2022 10:59:19 +0000 (11:59 +0100)]
arm64: zynqmp: Describe TI phy as ethernet-phy-id with reset on zcu106
zcu106 also connects ethernet phy reset via tca6416 chip as is done on
other evaluation boards. That's why describe this connection to make sure
that ethernet phy is reset before it's use.
Michal Simek [Sun, 6 Nov 2022 19:48:06 +0000 (11:48 -0800)]
xilinx: versal-net: Fix SYS_LOAD_ADDR to point to OCM
Versal NET mini U-Boot configuration is used for memory testing that's why
load address can't be really placed in memory which doesn't need to work
that's why move it to start of OCM which is the same memory which U-Boot is
running from.
Signed-off-by: Michal Simek <michal.simek@amd.com>
Michal Simek [Tue, 1 Nov 2022 00:08:44 +0000 (17:08 -0700)]
net: phy: Fix ethernet-phy-id <dot> in the code
Use dot instead of comma. The fix doesn't affect anything but it is good to
be aligned with used pattern. The first is used only for string size
calculation and the second change is in the comment.
Fixes: db681d4929ca ("net: phy: Add new read ethernet phy id function") Signed-off-by: Michal Simek <michal.simek@amd.com>
Ovidiu Panait [Wed, 12 Oct 2022 05:36:56 +0000 (08:36 +0300)]
timer: xilinx-timer: use timer_conv_64() to fix timer wrap around
Current xilinx_timer_get_count() implementation does not take into account
the periodic 32-bit wrap arounds, as it directly returns the 32-bit counter
register value. The roll-overs cause problems in the upper timer layers, as
generic timer code expects an incrementing 64-bit value from get_count() to
work correctly.
Add the missing 64-bit up-conversion to fix random hangs/delays in
__udelay().
Fixes: a36d86720f ("microblaze: Convert axi timer to DM driver") Signed-off-by: Ovidiu Panait <ovpanait@gmail.com> Reviewed-by: Michal Simek <michal.simek@amd.com> Link: https://lore.kernel.org/r/20221012053656.1492457-3-ovpanait@gmail.com Signed-off-by: Michal Simek <michal.simek@amd.com>
On systems with older PMUFW using these pinctrl properties can cause system
hang because there is missing feature autodetection.
When it is implemented support for these two properties should go back.
Michal Simek [Wed, 12 Oct 2022 09:35:30 +0000 (11:35 +0200)]
arm64: zynqmp: Disable watchdog by default for virt platform
Disable watchdog based on request in past that not all Linux rootfs have
proper utilities ready to service it. Enable it if your rootfs have proper
watchdog handling.
T Karthik Reddy [Wed, 12 Oct 2022 09:03:45 +0000 (11:03 +0200)]
microblaze: Make extended addressing support default
Axi qspi controller supports 32-bit & 24-bit addressing modes
for micron, macronix & spansion flash parts. But for winbond
flashes it only supports 24-bit addressing mode.
Enable CONFIG_SPI_FLASH_BAR to use extended addressing mode
to make 32-bit addressing mode work on all flashes.
This commit adds the menu-driven UEFI Secure Boot Key
enrollment interface. User can enroll PK, KEK, db
and dbx by selecting file.
Only the signed EFI Signature List(s) with an authenticated
header, typically '.auth' file, is accepted.
To clear the PK, KEK, db and dbx, user needs to enroll the null key
signed by PK or KEK.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org> Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Masahisa Kojima [Sun, 20 Nov 2022 00:21:16 +0000 (09:21 +0900)]
eficonfig: expose eficonfig_create_device_path()
Following commits are adding support for UEFI variable management
via the eficonfig menu. Those functions needs to use
eficonfig_create_device_path() to construct the full device path
from device path of the volume and file path, so move it
out of their static declarations.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org>
Masahisa Kojima [Sun, 20 Nov 2022 00:21:15 +0000 (09:21 +0900)]
eficonfig: refactor change boot order implementation
All the eficonfig menus other than "Change Boot Order"
use 'eficonfig_entry' structure for each menu entry.
This commit refactors change boot order implementation
to use 'eficonfig_entry' structure same as other menus
to have consistent menu handling.
This commit also simplifies the data->active handling when
KEY_SPACE is pressed, and sizeof() parameter.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org> Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Masahisa Kojima [Sun, 20 Nov 2022 00:21:14 +0000 (09:21 +0900)]
eficonfig: expose append entry function
Following commits are adding support for UEFI variable management
via the eficonfig menu. Those functions needs to use
append_entry() and append_quit_entry() to construct the
menu, so move them out of their static declarations.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org> Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Masahisa Kojima [Sun, 20 Nov 2022 00:21:13 +0000 (09:21 +0900)]
eficonfig: refactor file selection handling
eficonfig_select_file_handler() is commonly used to select the
file. eficonfig_display_select_file_option() adds an additional
menu to clear the selected file.
eficonfig_display_select_file_option() is not always necessary
for the file selection process, so it must be outside of
eficonfig_select_file_handler().
This commit also renames the following functions to avoid confusion.
eficonfig_select_file_handler() -> eficonfig_process_select_file()
eficonfig_select_file() -> eficonfig_show_file_selection()
eficonfig_display_select_file_option() -> eficonfig_process_show_file_option()
Finally, test_eficonfig.py need to be updated to get aligned with
the above modification.
Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org> Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Suggested-by: Simon Glass <sjg@chromium.org> Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Stefan Roese [Thu, 17 Nov 2022 08:20:34 +0000 (09:20 +0100)]
watchdog: Drop GD_FLG_WDT_READY as it's not used any more
Since commit c2fd0ca1a822
("watchdog: Integrate watchdog triggering into the cyclic framework")
GD_FLG_WDT_READY has become write-only. This patch now removes this
flag completely.
The vacant spot in gd_flags is filled with the newly introduced
GD_FLG_CYCLIC_RUNNING flag.
Suggested-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk> Signed-off-by: Stefan Roese <sr@denx.de>
Quentin Schulz [Tue, 15 Nov 2022 10:20:14 +0000 (11:20 +0100)]
watchdog: designware: make reset really optional
Checking for DM_RESET is not enough since not all watchdog
implementations use a reset lane. Such is the case for Rockchip
implementation for example. Since reset_assert_bulk will only succeed if
the resets property exists in the watchdog DT node, it needs to be
called only if a reset property is present.
This adds a condition on the resets property presence in the watchdog DT
node before assuming a reset lane needs to be fetched with
reset_assert_bulk, by calling ofnode_read_prop.
Samuel Holland [Mon, 31 Oct 2022 04:26:19 +0000 (23:26 -0500)]
spl: Fix SPL_ATF and SPL_OPENSBI dependencies
The code for these two options depends on having the FIT loadables
recorded in the FDT. Thus, these options require the full version of
the SPL_LOAD_FIT code.
Signed-off-by: Samuel Holland <samuel@sholland.org>
DFU implementation does not bound the length field in USB
DFU download setup packets, and it does not verify that
the transfer direction. Fixing the length and transfer
direction.
CVE-2022-2347
Signed-off-by: Venkatesh Yadav Abbarapu <venkatesh.abbarapu@amd.com> Reviewed-by: Marek Vasut <marex@denx.de>
Pali Rohár [Sun, 20 Nov 2022 16:56:28 +0000 (17:56 +0100)]
Nokia RX-51: Do not overwrite standard $loadaddr variable
Instead of overwriting $loadaddr variable, use custom temporary
$fileloadaddr variable. So scripts can access default/original address
stored in $loadaddr at build time.
Pali Rohár [Sun, 20 Nov 2022 16:56:26 +0000 (17:56 +0100)]
arm32: Fix relocation of env_addr if POSITION_INDEPENDENT=y
Apply commit 534f0fbd6520 ("arm64: Fix relocation of env_addr if
POSITION_INDEPENDENT=y") also for 32-bit ARM.
This change fixes crashing of U-Boot on ARMv7 (Omap3 / Cortex-A8) Nokia N900
phone (real HW). Note that qemu emulator of this board with same u-boot.bin
binary has not triggered this crash.
Crash happened after U-Boot printed following debug lines to serial console:
initcall: 0001ea8c (relocated to 8fe0aa8c)
Loading Environment from <NULL>... Using default environment
Destroy Hash Table: 8fe25a98 table = 00000000
Create Hash Table: N=387
Due to usage of PREBOOT in Kconfig, macro CONFIG_PREBOOT is always defined
when CONFIG_USE_PREBOOT is enabled. In case CONFIG_PREBOOT is not
explicitly enabled it is set to empty C string and therefore
'#ifdef CONFIG_PREBOOT' guard does not work. Fix this issue by introducing
a new Kconfig symbol PREBOOT_DEFINED which cause to define new C macro
CONFIG_PREBOOT_DEFINED only when CONFIG_PREBOOT is really defined.
Change usage of '#ifdef CONFIG_PREBOOT' by '#ifdef CONFIG_USE_PREBOOT' for
code which checks if preboot code would be called and by
'#ifdef CONFIG_PREBOOT_DEFINED' for defining preboot code.
- Fix and improve microchip's clock driver to allow sync'ing DTS with linux
- Improve the help message in "SBI_V02" Kconfig
- Improve DTS property "isa-string" parsing rule
Ilias Apalodimas [Fri, 11 Nov 2022 18:04:31 +0000 (20:04 +0200)]
efi_loader: replace a u16_strdup with alloc + memcpy
Heinrich reports that on RISC-V unaligned access is emulated by OpenSBI
which is very slow. Performance wise it's better if we skip the calls
to u16_strdup() -- which in turn calls u16_strsize() and just allocate/copy the
memory directly. The access to dp.length may still be unaligned, but that's
way less than what u16_strsize() would do
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Use malloc() instead of calloc(). Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Ilias Apalodimas [Thu, 10 Nov 2022 13:31:30 +0000 (15:31 +0200)]
efi_loader: add comments on efi_file_from_path() regarding alignment
UEFI specification requires pointers that are passed to protocol member
functions to be aligned. There's a u16_strdup in that function which
doesn't make sense otherwise Add a comment so no one removes it
accidentally
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Ilias Apalodimas [Thu, 10 Nov 2022 08:21:24 +0000 (10:21 +0200)]
efi_loader: initialize return values in efi_uninstall_multiple_protocol_interfaces_int()
If the va_list we got handed over contains no protocols we must return
EFI_SUCCESS. However in that case the current code just returns
an unintialized value.
Fix that by setting the return value in the variable definition
Quentin Schulz [Mon, 7 Nov 2022 13:49:31 +0000 (14:49 +0100)]
doc: fix build with sphinx 5.0 and later.
Sphinx 5.0 and later fails to build when language is set to None:
Warning, treated as error:
Invalid configuration value found: 'language = None'. Update your configuration to a valid langauge code. Falling back to 'en' (English)
Let's set the language to English since it is the language used for the
documentation.
Correct GD_FLG_CYCLIC_RUNNING documentation to match Sphinx style.
Fixes: d7de5ef62935 ("cyclic: use a flag in gd->flags for recursion protection") Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Stefan Roese <sr@denx.de>
riscv: Fix detecting FPU support in standard extension
We should check the string until it hits underscore, in case it
searches multi-letter extensions. For example, "rv64imac_xandes"
will be treated as D extension support since there is a "d" in
"andes", resulting illegal instruction caused by initializing FCSR.
Signed-off-by: Yu Chien Peter Lin <peterlin@andestech.com> Reviewed-by: Rick Chen <rick@andestech.com> Reviewed-by: Padmarao Begari <padmarao.begari@microchip.com> Reviewed-by: Samuel Holland <samuel@sholland.org>
Conor Dooley [Tue, 25 Oct 2022 07:58:49 +0000 (08:58 +0100)]
riscv: dts: fix the mpfs's reference clock frequency
The initial devicetree for PolarFire SoC incorrectly created a fixed
frequency clock in the devicetree to represent the msspll, but the
msspll is not a fixed frequency clock. The actual reference clock on a
board is either 125 or 100 MHz, 125 MHz in the case of the icicle kit.
Swap the incorrect representation of the msspll out for the actual
reference clock.
Fixes: dd4ee416a6 ("riscv: dts: Add device tree for Microchip Icicle Kit") Signed-off-by: Conor Dooley <conor.dooley@microchip.com> Reviewed-by: Leo Yu-Chi Liang <ycliang@andestech.com> Reviewed-by: Padmarao Begari <padmarao.begari@microchip.com>
Conor Dooley [Tue, 25 Oct 2022 07:58:48 +0000 (08:58 +0100)]
clk: microchip: mpfs: fix criticality of peripheral clocks
Sync the critical clocks in the U-Boot driver with those marked as
critical in Linux. The Linux driver has an explanation of why each clock
is considered to be critical, so import that too.
Conor Dooley [Tue, 25 Oct 2022 07:58:47 +0000 (08:58 +0100)]
clk: microchip: mpfs: fix periph clk parentage
Not all "periph" clocks are children of the AHB clock, some have the AXI
clock as their parent & the mtimer clock is derived from the external
reference clock directly. Stop assuming the AHB clock to be the parent
of all "periph" clocks and define their correct parents instead.
The original devicetrees for PolarFire SoC messed up & defined the
msspll's output as a fixed-frequency, 600 MHz clock & used that as the
input for the clock controller node. The msspll is not a fixed
frequency clock and later devicetrees handled this properly. Check the
devicetree & if it is one of the fixed ones, register the msspll.
Otherwise, skip registering it & pass the reference clock directly to
the cfg clock registration function so that existing devicetrees are
not broken by this change.
As the MSS PLL is not a "cfg" or a "periph" clock, add a new driver for
it, based on the one in Linux.
Conor Dooley [Tue, 25 Oct 2022 07:58:45 +0000 (08:58 +0100)]
clk: microchip: mpfs: convert parent rate acquistion to get_get_rate()
Currently the clock driver for PolarFire SoC takes a very naive approach
to the relationship between clocks. It reads the dt to get an input
clock, assumes that that is fixed frequency, reads the "clock-frequency"
property & uses that to set up both the "cfg" and "periph" clocks.
Simplifying for the sake of incremental fixes, the "correct" parentage for
the clocks currently supported in U-Boot is that the "cfg" clocks should
be children of the fixed frequency clock in the dt. The AHB clock is one
of these "cfg" clocks and is the parent of the "periph" clocks.
Instead of passing the clock rate of the fixed-frequency clock to the
"cfg" and "periph" registration functions and the name of the parents,
pass their actual parents & use clk_get_rate() to determine their parents
rates.
The "periph" clocks are purely gate clocks and should not be reading the
AHB clocks registers to determine their rates, as they can simply report
the output of clk_get_rate() on their parent.
Conor Dooley [Tue, 25 Oct 2022 07:58:44 +0000 (08:58 +0100)]
dt-bindings: clk: add missing clk ids for microchip mpfs
When this binding header was initally upstreamed, the PLL clocking the
microprocessor subsystem (MSS) and the RTC reference clocks were
omitted. Add them now, matching the IDs used in Linux.
Conor Dooley [Wed, 26 Oct 2022 07:49:19 +0000 (08:49 +0100)]
i2c: microchip: fix erroneous late ack send
A late ack is currently being sent at the end of a transfer due to
incorrect logic in mchp_corei2c_empty_rx(). Currently the Assert Ack
bit is being written to the controller's control reg after the last
byte has been received, causing it to sent another byte with the ack.
Instead, the AA flag should be written to the control register when
the penultimate byte is read so it is sent out for the last byte.
Conor Dooley [Wed, 26 Oct 2022 07:49:18 +0000 (08:49 +0100)]
i2c: microchip: fix ack sending logic
"Master receive mode" was not correctly sending ACKs/NACKs in the
interrupt handler. Bring the handling of M_SLAR_ACK, M_RX_DATA_ACKED &
M_RX_DATA_NACKED in line with the Linux driver.
projects / u-boot.git / log
