| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Unlike with ARM32, barebox prologue on ARM64 starts with the function
prologue, because of the absence of the naked attribute. The code is
written with that in mind (6 branches instead of 8 to account for the
two instructions inserted by the compiler), but it's still suprising.
Add a hint about that in the code.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20211108075209.2366770-5-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is currently used on ARM and sandbox, but it's relevant for
other platforms like x86 as well if driver code wants to make use
of the functions defined within.
MIPS and PowerPC already defines the symbols, but that's ok because
<bitio.h> doesn't override existing macros. This works for MIPS, but
not PowerPC, where those aren't macros. Fix that up.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20210812121944.4419-3-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Many arm32 board entry points use arm_setup_stack() to set up
the stack from C code. This necessitates using __naked, which
probably has been our most frequent cause of misscompiled C code.
GCC is quite clear that:
Only basic asm statements can safely be included in naked functions
While using extended asm or a mixture of basic asm and C code may
appear to work, they cannot be depended upon to work reliably and
are not supported.
But some boards use it anyway, because it's nice to avoid writing
assembly. Reading generated assembly to spot compiler miscompilation
isn't that nice though, so add some documentation, comments
and compiler diagnostics to hopefully reduce future porting effort.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20210917121152.16033-1-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
|
| |
Use a long instead of an u32. For 32-bit arm there is no difference, for
arm64 this results in being able to use the whole register size of 64
bits.
Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.barebox.org/20210916093356.173149-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
ARM trusted firmware has some common data structures passed to bl31.
This patch imports the code supporting this taken from U-Boot.
The defines and data structures are taken directly from U-Boot,
the support code is heavily modified for the sake of readability.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Link: https://lore.barebox.org/20210615141641.31577-9-s.hauer@pengutronix.de
Link: https://lore.barebox.org/20210621092802.27275-9-s.hauer@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
relocate_to_adr() currently relocates the PBL code to another address,
but doesn't move the piggy data (compressed barebox image). This patch
adds relocate_to_adr_full() which moves the full image including the
piggy data. This is needed for upcoming Rockchip RK3568 support. Here
we are started at address 0x0, but must move away from that address as
this is where the ATF/OP-Tee binaries are expected to be loaded.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Link: https://lore.barebox.org/20210615141641.31577-6-s.hauer@pengutronix.de
Link: https://lore.barebox.org/20210621092802.27275-6-s.hauer@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
barebox_add_memory_bank() can fail if the to-be-added memory region
has been requested before. This can happen most easily on i.MX and
STM32MP1 boards:
- The /memory node in the device tree requests a region
- The DDR controller driver requests an overlapping region after
reading back RAM configuration.
This most often leads to error messages down the road, but it can be
difficult to pinpoint the cause. Propagate the error code from
arm_add_mem_device(), so DDR controller drivers can fail their probe
on error.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20210531071239.30653-6-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When running under UEFI, barebox should no redo PCI enumeration,
because the UEFI implementation will likely already have drivers
that won't cope with e.g. BAR addresses changing.
The user-visible effect of this is that likely the framebuffer will
stop working because the UEFI driver won't be able to access it
any longer.
Support this configuration by changing the PCI code to consult the
new pcibios_assign_all_busses().
When it's true, there is no change to previous behavior.
When it's false, reconfiguration is omitted and instead current
configuration is read back from the bus.
Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Link: https://lore.barebox.org/20210416062436.332665-3-ahmad@a3f.at
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
Other PBL-enabled architecture can benefit from compressed dtbs as well.
Move symbol and code to a comm place to be able to use it from RISC-V
in a later commit. In order not to break out of tree boards at runtime,
the old symbol name is maintained for ARM.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
There's nothing ARM specific about these functions. Move them to a
common location, so other arches can use them as well.
This also fixes a bug on ARM MMU=n configurations:
Previously these two functions these functions only respected
dma-ranges translation when compiled with MMU support. DMA address
translation is applicable regardless of CPU use of MMU though.
Now, dma-ranges should be respected unconditionally.
Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Implement initjmp() for use with the incoming bthread support.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |/
| |
| |
| |
| |
| |
| |
| |
| | |
To avoid invalid optimizations and to enable warnings, GCC must be told
that setjmp and longjmp are to be handled specially. Add the missing
attributes.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The ARM64 generic DT image is meant to be interchangeable with a
kernel image to aid in debugging and testing. To achieve this, it
reuses an externally passed device tree in x0, just like Linux does.
The ARM barebox image format used however imposes some limitations:
- Commands verifying the header before boot, like U-Boot's booti
won't boot the barebox image
- The barebox image is not fully relocatable. It has the implicit
assumption that the barebox stack can grow into the memory space
preceding the barebox image while the /memory node is parsed
from the FDT.
Adopting the Linux ARM64 header solves both issues. booti won't be
able to tell us apart and we can specify an image load offset to
reserve a stack space that won't interfere with anything else.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
| |
cpu_is_arm(cpu) masks the read_cpuid_id() return value and compares with
a CPU-specific value. The used masks for the Cortex A53/57 zeroes bits
that are set in the value, so the comparison will never return true.
In fact the compiler already optimizes it away completely, because
read_cpuid_id() is only implemented for ARM32 anyway.
Drop the dead code.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Description from the linux commit 5756e9dd0de6d5c307773f8f734c0684b3098fdd:
ARM: 6640/1: Thumb-2: Symbol manipulation macros for function body copying
In low-level board support code, there is sometimes a need to
copy a function body to another location at run-time.
A straightforward call to memcpy doesn't work in Thumb-2,
because bit 0 of external Thumb function symbols is set to 1,
indicating that the function is Thumb. Without corrective
measures, this will cause an off-by-one copy, and the copy
may be called using the wrong instruction set.
This patch adds an fncpy() macro to help with such copies.
Particular care is needed, because C doesn't guarantee any
defined behaviour when casting a function pointer to any other
type. This has been observed to lead to strange optimisation
side-effects when doing the arithmetic which is required in
order to copy/move function bodies correctly in Thumb-2.
Thanks to Russell King and Nicolas Pitre for their input
on this patch.
Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
| |
<asm-generic/io.h> expects architectures to define memcpy_toio,
memcpy_fromio and memset_io as macros as it checks for their existence
with #ifndef. Only ARM defines an own implementation, but does so
as function declaration. Add the missing macros.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds KASan support to the ARM architecture. What we are doing is:
* Add __no_sanitize_address attribute to various lowlevel functions
which do not run in a proper C environment
* Add non-instrumented variants of memset/memcpy (prefixed with '__')
* make original memcpy/memset weak symbols so strong definitions in
lib/kasan/common.c can replace them
* Use non-instrumented memcpy in early functions
* call kasan_init()
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Unlike the Linux kernel, barebox does not have a dedicated heap for
storing modules. Therefore, if the system memory configuration places
the general heap further away than can be reached by a 'bl' instruction
(24 bits of address, or 16 MiB), then the module relocations will fail
due to being out of range.
Allocate PLTs when loading modules so that jumps and calls whose
targets are too far away for their relative offsets to be encoded
in the instructions themselves can be bounced via veneers in the
module's PLT. The modules will use slightly more memory, but after
rounding up to page size, the actual memory footprint is usually
the same.
Adoption of Linux commits:
66e94ba3c8ea ARM: kernel: avoid brute force search on PLT generation
1031a7e674d1 ARM: kernel: sort relocation sections before allocating PLTs
05123fef0982 ARM: kernel: allocate PLT entries only for external symbols
35fa91eed817 ARM: kernel: merge core and init PLTs
7d485f647c1f ARM: 8220/1: allow modules outside of bl range
Signed-off-by: David Dgien <dgienda125@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| | |
Signed-off-by: David Dgien <dgienda125@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
According to Marc Zyngier, former employee at ARM, the company owns the
copyright for code created by its employees. Convert accordingly to SPDX
with the usual rearrangements. Also dropped Marc's email address which
doesn't work any more.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
| |
Converts the files that licensecheck can determine the license for. Also
convert to obvious right Copyright statements. Additionally some minor
code reformatting is done.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
For easier code-sharing, the UBIFS code still uses the Linux atomic
accessors. Make porting it to sandbox easier, by providing these
functions for all platforms.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| | |
With the generic implementations in place in <asm-generic/bitio.h>, we can
drop the now-duplicate ARM definitions.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Many architectures duplicate the same fixed type definitions
as in <asm-generic/int-ll64.h>.
Include <asm-generic/int-ll64.h> from <asm/types.h>, and remove
the duplicated code.
Note:
Linux commit 0c79a8e29b5f ("asm/types.h: Remove
include/asm-generic/int-l64.h") unified the kernel space
definition into int-ll64.h. Barebox also adopts int-ll64.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
The CREDITS file was removed from barebox in 2015 by commit 6570288f2d97
("Remove the CREDITS file"). Remove references to it from several files.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
File has been in source tree for 17 years and is unused.
Drop it. The only definition we might want out of the file is
cpu_relax(), which we already open code as barrier() with a comment in a
couple of places. Move the definition to <linux/barebox-wrapper.h>
and drop the file altogether.
No functional change.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a OP-TEE early loading function which expects a pointer to a valid
tee binary and the device tree. OP-TEE will then be started and barebox
will continue to run in normal mode.
The function start_optee_early should be used in a boards lowlevel.c
file. Ensure that barebox has been relocated and a proper c environment
has been setup beforehand. Depending on the OP-TEE configuration, the
fdt will be modified. If the internal barebox device tree is passed,
OP-TEE will overwrite barebox PBL memory during this modification. Copy
the fdt to a save memory location beforehand to avoid a corruption of
barebox PBL memory.
This also moves the OP-TEE Kconfig symbols into a separate menu.
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
| |
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| | |
ARM64 uses a different assembler mnemonic for the breakpoint.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
| |
3 level page tables only allow to resolve 39 bit addresses. Switch to 4
level page tables to add support for bigger physical address ranges.
This is needed for example on Layerscape SoCs where the PCI windows are
outside the 39bit range.
The early MMU support still uses 39bit addressing. We only use a single
level page table in early MMU support and with 48bit addresses we
wouldn't have enough granularity to map the SDRAM differently then the
rest of the address space.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
System reset on the STM32MP may be done via PSCI when running TF-A
as first-stage boot loader. Provide a PSCI driver to simplify using it:
- A psci_invoke function is exported, so other code can use it
- A fixup for the PSCI device tree node is registered
- A reset and poweroff handler via PSCI is registered for PSCI >= v0.2
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
GCC9 now produces the following warning:
common.h:51:2: warning: listing the stack pointer register ‘sp’ in a clobber list is deprecated [-Wdeprecated]
51 | __asm__ __volatile__("mov sp, %0"
| ^~~~~~~
common.h:51:2: note: the value of the stack pointer after an ‘asm’ statement must be the same as it was before the statement
Stack pointer was added to clobber list in commit f9fc8254b2 ("ARM:
Mark SP as being clobbered in arm_setup_stack()") to prevent GCC from
generating code that would corrupt 'boarddata' pointer by trying to
restore it from invalid stack frame.
Interestingly enough, seemingly unrelated change in commit
64d95896cf ("ARM: aarch64: compile with general-regs-only") changed
generated code such that adding SP to clobber list became no longer
necessary.
While the above can probably be a fix by itself, it seems a better and
more future proof approach would be to address the problem at its root
and re-implement offending startup sequence in assembly.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
With the latest changes, this file is empty on all archs. Drop it.
Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Other arch-specific features are exposed in Kconfig too, so do here
likewise.
Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Other arch-specific features are exposed in Kconfig too, so do here
likewise.
Signed-off-by: Ahmad Fatoum <ahmad@a3f.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
__boot_cpu_mode is accessed from barebox_multi_pbl_start() and
barebox_single_pbl_start(). These functions may be called at an
address different from the address they are linked at. Calculating
the address of global variables can yield wrong results when it
is done before setup_c() is called. We can't make sure when the
address is calculated, OSELAS.Toolchain-2018.12.0 indeed calculates
the address after setup_c() is called, but Debian
arm-linux-gnueabihf-gcc 8.3.0 does it before setup_c() is called
and thus doesn't work.
This is solved by accessing __boot_cpu_mode with a wrapper function
which we call explicitly after setup_c() is done.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Not all toolchains use pc relative addresses for global variables.
Apparently the gcc 8.3.0 YOCTO toolchain uses absolute addresses.
This means can't simply return 0 for global_variable_offset() but
instead have to calculate the offset between the compile addresses
for global variables and their runtime address.
We do this by getting the address of a global variable pc relative
explicitely in assembly and substracting that address from the
location the C compiler thinks they are.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Incoming PSCI implementation for the STM32MP1 needs
to get_gicd_base_address as well, so move it to gic.h to
avoid code duplication.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
| |
With this option barebox will be build with breakpoint instruction
in early pbl stage.
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This adds initial support for OP-TEE, see https://www.op-tee.org/
barebox starts in secure mode as usual. When booting a kernel
the bootm code also loads the optee_os binary. Instead of jumping
into the kernel barebox jumps into the optee_os binary and puts
the kernel execution address into the lr register. OP-TEE then
jumps into the kernel in nonsecure mode.
The optee_os binary is passed with the -t option to bootm or
with global.bootm.tee.
Optionally OP-TEE can be compiled into barebox using the builtin firmware
feature. Enable the Kconfig option and place or link your tee binary as
optee.bin into the firmware directory.
The amount of SDRAM which is kept free for OP-TEE is configurable.
This patch was tested on a i.MX6 Nitrogen6x board.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
of_psci_fixup() can be used by code which doesn't use the barebox
psci implementation, but provides its own PSCI compatible firmware.
Factor it out to a separate file to compile it independently of the
barebox PSCI implementation.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|