A lot of files rely on include/driver.h including include/of.h (and
this including include/errno.h. include the files explicitly so we can
eventually get rid of including of.h from driver.h
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Drivers currently cannot implement explicit cache handling and rely on
running the same code before and after mmu_initcall() without crashing.
Depending on the chosen config options, the cache functions are not yet
setup and using them early on ends in a null pointer dereference.
The RPi's mailbox driver is such a case; it requires cache handling once
the MMU is fully set up and yet the RPi setup needs to use the driver to
get the memory size before mem_initcall() and hence mmu_initcall().
Fix this by checking the cache_fns pointer before dereferencing it.
Signed-off-by: Andre Heider <a.heider@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
When using CONFIG_MMU_EARLY combined with CONFIG_PBL_IMAGE, the barebox
setup reuses the MMU setup from the PBL, but doesn't setup the cache
functions.
Set these up to guarantee proper early cache handing before mmu_initcall().
Signed-off-by: Andre Heider <a.heider@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The data caches should be invalided once during startup. This should
also be done when we do not have the MMU enabled in barebox because
the Kernel does not invalidate the caches during start.
To make this sure this patch enables the arm_early_mmu_cache_invalidate
function even if MMU support is disabled. Additionally this patch adds
calls to arm_early_mmu_cache_invalidate in start.c and uncompress.c.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This adds the make infrastructure to build multiple SoC or
board specific images from a single barebox binary.
The basic idea is that we no longer have a single pbl, but instead
multiple pbls, one per image if necessary. Each pbl is defined
by its entry function so that each pbl can do exactly what a given
board needs. Additionally the pbls together with a self extracting
barebox binary can be encapsulated in specific image formats.
squashed in build fixes from Lucas Stach for make version >= 3.82:
Split Multimage Makefile rule in explicit and implicit parts
Fixes build with make version >=3.82
Frome the make 3.82 NEWS file:
* WARNING: Backward-incompatibility!
In previous versions of make it was acceptable to list one or more explicit
targets followed by one or more pattern targets in the same rule and it
worked "as expected". However, this was not documented as acceptable and if
you listed any explicit targets AFTER the pattern targets, the entire rule
would be mis-parsed. This release removes this ability completely: make
will generate an error message if you mix explicit and pattern targets in
the same rule.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Addionally to having a builtin DTB provide the possibility for
the board to provide a dtb via boarddata.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Some SoCs come up with invalid entries in the data cache. This can
lead to memory corruption when we enable them later, so invalidate
the caches early.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Tested-by: Lucas Stach <l.stach@pengutronix.de>
At least the iMX6 boot rom seems to jump into barebox with a non
invalidated d-cache which causes data corruption when
v7_mmu_cache_flush() executed by arm_early_mmu_cache_flush() overrides
stack or other valid data.
That's why the cache must be invalided for this processors explicitly
(e.g. in barebox_arm_reset_vector()). Operation differs from flush only
in one instruction so that patch modifies the existing
v7_mmu_cache_flush() function slightly by adding an optional argument.
Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Registers 'r0' till 'r3' are scratch registers and do not need to be
restored.
Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Although conclusions in 50d1b2de8e "ARM
v7: Fix register corruption in v7_mmu_cache_off" are correct, the
implemented fix is not complete because the following failure can
happen:
1. d-cache contains the cache line around 'sp'
2. v7_mmu_cache_off() disables cache
3. early v7_mmu_cache_flush() pushes 'lr' on uncached stack
4. v7_mmu_cache_flush() flushes d-cache and can override stack written
by step 3.
5. v7_mmu_cache_flush() pops 'lr' out of cache and jumps to it which
might be random data now.
Patch avoids step 3 which is easy because 'lr' is never modified by the
function. By using the 'r12' scratch register instead of 'r10', the
whole initial 'push' can be avoided.
Patch moves also the 'DMB' operation so that it is executed after data
has been pushed on stack.
Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The Kirkwood Marvell SoC uses a Marvell-specific implementation of an
ARMv5TE compatible ARM core, the Feroceon. This patch introduces a
Kconfig option that allows to select this CPU type.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We PAGE_ALIGN the size in dma_alloc_coherent so do it also when free the memory.
Use PAGE_SIZE instead of magic numbers.
Signed-off-by: Jan Weitzel <j.weitzel@phytec.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
- Add rules to generate dtb files in arch/arm/dts/
- add an initcall which unflattens and probes the internal devicetree
- Add skeleton devicetree
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The comment above barebox_arm_entry promises to preserve the boarddata
variable passed to it which can then later get back with
barebox_arm_boarddata(). This function was missing so far, add it.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
For making the same binary executable on different SoCs which have
different DRAM addresses we have to be independent of the compile
time link address.
This patch adds relocatable binary support for the ARM architecture.
With this two new functions are available. relocate_to_current_adr
will fixup the binary to continue executing from the current position.
relocate_to_adr will copy the binary to a given address, fixup the
binary and continue executing from there.
For the PBL and the real image relocatable support can be enabled
independently. This is done to (hopefully) better cope with setups
where the PBL runs from SRAM or ROM and the real binary does not.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
With relocatable binaries the vector addresses cannot be supplied by
the linker. This adds support for fixing them up during runtime.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Since recently with MMU_EARLY support it may happen that setup_c
runs with data caches enabled, so we have to make sure the caches
are flushed before we jump to the new binary.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
When we have multi cpu support compiled in we need the cpu architecture
early so that we can pick the correct cacheflush function. Make it available
as static inline function and add a comment above it that this function
normally should not be used.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
reset is confusing with the cpu reset and impossible to grep
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This optionally enabled the MMU in the PBL or during early startup for
the non PBL case. The regular MMU init code will pickup the already enabled
MMU later. This might complicate debugging early code, so this has been
made optional.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Move early mmu code to a separate file so that it can be
used from the pbl and the regular image. Disabling the mmu
can be dropped since the regular mmu code is now able to
pickup an enabled mmu.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Additionally to the generic entry point the i.MX specific ones
calculate the SDRAM size automatically so the boards do not have
to care.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Memory is a precious resource, so it makes sense to make it available as
early as possible. By definition the lowlevel init code already knows where
to find memory because it's the lowlevel init code which sets up the memory.
Until all boards are converted this new entry is just a fallback to the old
entry point.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Old cache/uncache pte flags were declared as defines.
Since these flags are determine at runtime they are static
variables.
This patch switch the naming style of these variables to
lower case which is typically used for variables.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>