Currently, binman always runs the compile tools like cc, objcopy, strip,
etc. using their literal name. Instead, this patch makes it use the
target-specific versions by default, derived from the tool-specific
environment variables (CC, OBJCOPY, STRIP, etc.) or from the
CROSS_COMPILE environment variable.
For example, the u-boot-elf etype directly uses 'strip'. Trying to run
the tests with 'CROSS_COMPILE=i686-linux-gnu- binman test' on an arm64
host results in the '097_elf_strip.dts' test to fail as the arm64
version of 'strip' can't understand the format of the x86 ELF file.
This also adjusts some command.Output() calls that caused test errors or
failures to use the target versions of the tools they call. After this,
patch, an arm64 host can run all tests with no errors or failures using
a correct CROSS_COMPILE value.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
These test files are currently "intended for use on x86 hosts", but most
of the tests using them can still pass when cross-compiled to x86 on an
arm64 host.
This patch enables non-x86 hosts to run the tests by specifying a
cross-compiler via CROSS_COMPILE. The list of variables it sets is taken
from the top-level Makefile. It would be possible to automatically set
an x86 cross-compiler with a few blocks like:
ifneq ($(shell i386-linux-gnu-gcc --version 2> /dev/null),)
CROSS_COMPILE = i386-linux-gnu-
endif
But it wouldn't propagate to the binman process calling this Makefile,
so it's better just raise an error and expect 'binman test' to be run
with a correct CROSS_COMPILE.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
This patch makes buildman create linked working trees instead of clones
of the source repository, but keeps updating the older clones of the
repository that might already exist. These worktrees share "everything
except working directory specific files such as HEAD, index, etc." with
the source repository. See the git-worktree(1) manual page for more
information.
If git-worktree isn't available, silently falls back to cloning the
repository.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
At present 64-bit sunxi boards use the Makefile to create a FIT, using
USE_SPL_FIT_GENERATOR. This is deprecated.
Update sunxi to use binman instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
At present binman warns about missing external blobs only when the
BUILD_ROM is defined. Enable this behaviour always, since many boards
are starting to use these (e.g. ARM Trusted Firmware's BL31).
Signed-off-by: Simon Glass <sjg@chromium.org>
In some cases it is useful to generate a FIT which has a number of DTB
images, selectable by configuration. Add support for this in binman,
using a simple iterator and string substitution.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add an entry for ARM Trusted Firmware's 'BL31' payload, which is the
device's main firmware. Typically this is U-Boot.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present we have an Entry_blob_ext which implement a blob which holds an
external binary. We need to support other entry types that hold external
binaries, e.g. Entry_blob_named_by_arg. Move the support into the base
Entry class to allow this.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the Python sequential-write interface can produce an error when
it calls fdt_finish(), since this needs to add a terminating tag to the
end of the struct section.
Fix this by automatically expanding the buffer if needed.
Signed-off-by: Simon Glass <sjg@chromium.org>
Tidy up a few test functions which lack argument comments. Rename one that
has the same name as a different test.
Also fix up the comment for PrepareImagesAndDtbs().
Signed-off-by: Simon Glass <sjg@chromium.org>
If an entry argument is needed by an entry but the entry argument is not
present, then a strange error can occur when trying to read the file.
Fix this by allowing arguments to be required. Select this option for the
cros-ec-rw entry. If a filename is provided in the node, allow that to be
used.
Also tidy up a few related tests to make the error string easier to find,
and fully ignore unused return values.
Signed-off-by: Simon Glass <sjg@chromium.org>
Currently of_match_ptr is used to avoid referencing compatible strings
when OF_CONTROL is not enabled. This behaviour could be improved by
taking into account also OF_PLATDATA, as when this configuration is
enabled the compatible strings are not used at all.
Signed-off-by: Walter Lozano <walter.lozano@collabora.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present we look for resources based on the path of the Python module
that wants them. Instead we should use Python's pkg_resources feature
which is designed for this purpose.
Update binman to use this.
Signed-off-by: Simon Glass <sjg@chromium.org>
When reading subentries of each image, the FIT entry type directly
concatenates their contents without padding them according to their
offset, size, align, align-size, align-end, pad-before, pad-after
properties.
This patch makes sure these properties are respected by offloading this
image-data building to the section etype, where each subnode of the
"images" node is processed as a section. Alignments and offsets are
respective to the beginning of each image. For example, the following
fragment can end up having "u-boot-spl" start at 0x88 within the final
FIT binary, while "u-boot" would then end up starting at e.g. 0x20088.
fit {
description = "example";
images {
kernel-1 {
description = "U-Boot with SPL";
type = "kernel";
arch = "arm64";
os = "linux";
compression = "none";
u-boot-spl {
};
u-boot {
align = <0x10000>;
};
};
};
}
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reinstate check in testPadInSections(), squash in
"binman: Allow FIT binaries to have missing external blobs"
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Other relevant properties (pad-after, offset, size, align, align-size,
align-end) already work since Pack() sets correct ranges for subentries'
data (.offset, .size variables), but some padding here is necessary to
align the data within this range to match the pad-before property.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Switch to str.startswith for matching like the FIT etype does since the
current version doesn't ignore 'hash-1', 'hash-2', etc.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Setting CONFIG_ENV_ADDR to something other than 0 stops gd->env_addr
from being allocated dynamically. When the environment is in SPI we need
it to be allocated as we can't use a direct memory mapped address.
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Fix build failure, it used to get this implicitly through common.h
until f7ae49fc4f (common: Drop log.h from common header).
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Heiko Schocher <hs@denx.de>
Somewhere between v2020.04 and v2020.07 the mpc8xxx_spi driver broke,
I'm guessing due to this hunk
@@ -559,6 +560,8 @@ int dm_gpio_set_dir_flags(struct gpio_desc *desc, ulong flags)
if (ret)
return ret;
+ /* combine the requested flags (for IN/OUT) and the descriptor flags */
+ flags |= desc->flags;
ret = _dm_gpio_set_dir_flags(desc, flags);
from commit 695e5fd546 ("gpio: update dir_flags management"). But
the blame is mostly on the driver itself which seems rather confused:
The chip select gpios are requested with GPIOD_ACTIVE_LOW, but then in
each activate/deactivate, dm_gpio_set_dir_flags() is called with
merely GPIOD_IS_OUT, and then the driver call set_value(0) for
activate.
That used to work, but with the above hunk, the ACTIVE_LOW setting
from the request becomes persistent, so the gpio driver ends up being
asked to set the value to 1 in mpc8xxx_spi_cs_activate().
So drop the dm_gpio_set_dir_flags() calls in the activate/deactivate
functions, and use a value of 1 to mean "logically enabled".
Ideally, I think we should also drop the GPIOD_ACTIVE_LOW from the
request and make it up to the list of gpio cs in DT to indicate
whether that CS is enabled when driven low (as is of course usually
the case), but that requires changing
arch/powerpc/dts/gdsys/gazerbeam-base.dtsi among others, and I don't
have that hardware to test on. I have, however, tested our
own (mpc8309-based) hardware with this change, and I have also tested
that removing the GPIOD_ACTIVE_LOW from the request and updating our
DT as
- gpios = <&spisel 0 0>;
+ gpios = <&spisel 0 GPIO_ACTIVE_LOW>;
still works.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Linking a U-Boot larger than 1MB fails with PIE enabled:
u-boot/arch/arm/cpu/armv8/start.S:71:(.text+0x3c): relocation
truncated to fit: R_AARCH64_ADR_PREL_LO21 against symbol `__rel_dyn_end'
defined in .bss_start section in u-boot.
This extends the supported range by using adrp & add to load symbols
early while starting up.
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
PIE requires a 4K aligned load address. If this is not met, trap
the startup sequence in a WFI loop rather than running into obscure
failures.
Tested-by: Michal Simek <michal.simek@xilinx.com>
Suggested-by: André Przywara <andre.przywara@arm.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Mention the requirement of 4K aligned load addresses in the
help section for the POSITION_INDEPENDENT option.
Suggested-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Atmel PIT timer is not available for next products that
have another timer hardware block.
To be able to use the common at91 code, guard the code that uses PIT
by ifdefs.
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
We already have a function to retrieve the mac address from one EEPROM.
For boards with a second Ethernet interface, however, we would
require another EEPROM with a second unique MAC address.
Introduce at91_set_eth1addr which will look for a second EEPROM
and set the 'eth1addr' variable with the obtained MAC address.
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Add UTMI support for SAMA7G5. SAMA7G5's UTMI control is done via
XTALF register. Values written at bits 2..0 in this register
correspond to the on board crystal oscillator frequency.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Add master clock (MCK1..MCK4) support for SAMA7G5. SAMA7G5's PMC has
multiple master clocks feeding different subsystems. One of them
feeds image subsystem and is changeable based on image subsystem
needs.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Add sckc driver compatible with common clock framework. Driver
implements slow clock support for SAM9X60 compatible IPs (in this
list it is also present SAMA7G5's slow clock IP).
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Move clock code to compat.c to allow switching to CCF
without mixing CCF code with non CCF code. This prepares the
field for next commits.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Add helper for clock drivers. These will be used by following
commits in the process of switching AT91 clock drivers to CCF.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Add pre-requisite headers for AT91 clock architecture. These
are based on already present files on Linux and will be used
by following commits for AT91 CCF clock drivers.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
clk_get_by_indexed_prop() retrieves a clock with dev member being set
with the pointer to the udevice for the clock controller driver. But
in case of CCF each clock driver has set in dev member the reference
to its parent (the root of the clock tree is a fixed clock, every
node in clock tree is a clock registered with clk_register()). In this
case the subsequent operations like dev_get_clk_ptr() on clocks
retrieved by clk_get_by_indexed_prop() will fail. For this, get the
pointer to the proper clock registered (with clk_register()) using
clk_get_by_id() before proceeding.
Fixes: 1d7993d1d0 ("clk: Port Linux common clock framework [CCF] for imx6q to U-boot (tag: v5.1.12)")
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Clock re-parenting is not binding the clock's device to its new
parent device, it only calls the clock's ops->set_parent() API. The
changes in this commit re-parent the clock device to its new parent
so that subsequent operations like clk_get_parent() to point to the
proper parent.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
In common clock framework the relation b/w parent and child clocks is
determined based on the udevice parent/child information. A clock
parent could be changed based on devices needs. In case this is happen
the functionalities for clock who's parent is changed are broken. Add
a function that reparent a device. This will be used in clk-uclass.c
to reparent a clock device.
Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
