It looks like some soc_offsets.h files need to be included before
kernel_offsets, otherwise there are some header race conditions due to
the infamous soc.h. This problem is exposed if all soc.h are removed
from RISC-V arch header files (see the upcoming commits). It can be
reproduced by building rv32m1_vega_ri5cy board after applying all the
patches in this series (excluding this one, of course).
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
There is no need to sync in every xtlb invalidation. Sync only
after all tlb autofill ways invalidation.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Xtensa's arch level atomic implementation requries S32C1I
support which may not exist as Xtensa cores are highly
configurable. Implying CONFIG_ATOMIC_OPERATIONS_ARCH at
arch level has a side effect that it is enabled
regardless if the other options are enabled. Given how
the header file's #ifdef is structured, only C atomic
ops can override this, which means enabling
CONFIG_ATOMIC_OPERATIONS_BUILTIN has no effect at all.
So let the SoC or board decide which atomic ops are
being used instead.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
According to the clic specification
(https://github.com/riscv/riscv-fast-interrupt), the mnxti register has
be written, in order to clear the pending bit for non-vectored
interrupts. For vectored interrupts, this is automatically done.
From the spec:
"If the pending interrupt is edge-triggered, hardware will automatically
clear the corresponding pending bit when the CSR instruction that
accesses xnxti includes a write."
I added a kconfig `RISCV_SOC_HAS_CUSTOM_IRQ_HANDLING` to allow custom
irq handling. If enabled, `__soc_handle_all_irqs` has to be implemented.
For clic, non-vectored mode, I added a `__soc_handle_all_irqs`, that
handles the pending interrupts according to the pseudo code in the spec.
Signed-off-by: Greter Raffael <rgreter@baumer.com>
The irq priority has to be called for dynamic and direct irqs, too. For
direct isrs, this was missing completely, for direct irqs just for the
clic.
For dynamic irqs, I replaced the current implementation with
`z_riscv_irq_priority_set`. For the plic, this is exaclty the same.
Signed-off-by: Greter Raffael <rgreter@baumer.com>
For some reason, unrelated code change triggered compiler
warning about this function returns even though it is
marked nonreturn. So add CODE_UNREACHABLE to silence
the warning, possibly to catch any errors.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
wsr.UPPERCASE can lead to compiler errors when UPPERCASE matches
a macro defined in the special register header file.
Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
rsr.UPPERCASE can lead to compiler errors when UPPERCASE matches
a macro defined in the special register header file.
Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
Instead of custom SOC_MCAUSE_EXP_MASK definition. Note that SoCs
selecting RISCV_PRIVILEGED already used such config indirectly (see
changes in soc_common.h).
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
As defined in Table 3.6 of "The RISC-V Instruction Set Manual, Volume
II: Privileged Architecture". Delete all spread definitions of the same,
weirdly prefixed with "SOC".
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Initial implementation of `sched_getparam()` and `sched_getscheduler()`
POSIX APIs as a part of PSE53 `_POSIX_PRIORITY_SCHEDULING` option group.
Both functions are actually placeholders and just return `ENOSYS`
since Zephyr does not yet support processes or process scheduling.
Signed-off-by: Dmitrii Golovanov <dmitrii.golovanov@intel.com>
Because it was exclusively used by the "common" RISC-V privileged code
to build CPU idle routines that are now handled by arch level code.
Also, all platforms defaulted to "y", making it pointless in practice.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
While going to idle may require SoC specific implementations, provide a
more sensible default implementation.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
The constant used to calculate TLB entries for the way six was wrong
and causing an integer overflow. Consequently only the first 512MB where
being unmapped from the TLB.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Use the arch-cache functions instead of the sys-cache-functions
in z_arm_init_arch_hw_at_boot to ensure that the caches are
disabled even when CONFIG_CACHE_MANAGEMENT is disabled.
Signed-off-by: Benedikt Schmidt <benedikt.schmidt@embedded-solutions.at>
Use sys_cache_data_enable instead of arch_dcache_enable to enable
the cache. This will ensure that CONFIG_CACHE_MANAGEMENT
is considered correctly.
Signed-off-by: Benedikt Schmidt <benedikt.schmidt@embedded-solutions.at>
The interface to flush fpu is not unique to one architecture, make it a
generic, optional interface that can be implemented (and overriden) by a
platform.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Different architecture are doing this in custom ways and using different
naming conventions, unify this interface and make it part of the arch
implementation for SMP.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Introduce a new arch level Kconfig option to signal the implementation
of the RISCV Privileged ISA spec. This replaces
SOC_FAMILY_RISCV_PRIVILEGED, because this is not a SoC specific
property, nor a SoC family.
Note that the SoC family naming scheme will be fixed in upcoming
commits.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
This adds a kconfig to enable invalidating the TLBs related to
the incoming thread's memory domain during page table swaps.
It provides a workaround, if needed, to clear out stale TLB
entries used by the thread being swapped out. Those stale
entries may contain incorrect permissions and rings.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
There are several subsystems and boards which require a relatively large
system heap (used by k_malloc()) to function properly. This became even
more notable with the recent introduction of the ACPICA library, which
causes ACPI-using boards to require a system heap of up to several
megabytes in size.
Until now, subsystems and boards have tried to solve this by having
Kconfig overlays which modify the default value of HEAP_MEM_POOL_SIZE.
This works ok, except when applications start explicitly setting values
in their prj.conf files:
$ git grep CONFIG_HEAP_MEM_POOL_SIZE= tests samples|wc -l
157
The vast majority of values set by current sample or test applications
is much too small for subsystems like ACPI, which results in the
application not being able to run on such boards.
To solve this situation, we introduce support for subsystems to specify
their own custom system heap size requirement. Subsystems do
this by defining Kconfig options with the prefix HEAP_MEM_POOL_ADD_SIZE_.
The final value of the system heap is the sum of the custom
minimum requirements, or the value existing HEAP_MEM_POOL_SIZE option,
whichever is greater.
We also introduce a new HEAP_MEM_POOL_IGNORE_MIN Kconfig option which
applications can use to force a lower value than what subsystems have
specficied, however this behavior is disabled by default.
Whenever the minimum is greater than the requested value a CMake warning
will be issued in the build output.
This patch ends up modifying several places outside of kernel code,
since the presence of the system heap is no longer detected using a
non-zero CONFIG_HEAP_MEM_POOL_SIZE value, rather it's now detected using
a new K_HEAP_MEM_POOL_SIZE value that's evaluated at build.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This commit adds implementation of GDB stub for 32-bit ARM. It has been
tested only on the Zynq-7000 SoC and I would like to get any feedback
from others.
The stub still has these issues:
- To implement single stepping, it uses instruction address mismatch
breakpoint, as recommended in ARMv7 reference. The breakpoint control
register is configured (the state control fields) for the "PL0,
Supervisor and System modes only" option. Otherwise the breakpoint
would also halt the processor in abort mode, in which the stub loop
runs. Zephyr kernel runs in the system mode. This works well until the
kernel enables interrupts, as interrupt handlers typically run in
Supervisor mode. Single stepping therefore sometimes "catches" a
handler instead of the next application instruction. I have not tried
User mode, because Cortex-A SoCs do not appear to have the
ARCH_HAS_USERSPACE flag.
Cc: Michal Sojka <michal.sojka@cvut.cz>
Signed-off-by: Marek Vedral <marek.vedral@gmail.com>
The only difference in the two enums are some entries related to
relocation sections. However, these entries are not used in the
code, so they can be safely removed, along with the mapping function.
Use LLEXT_MEM_* to avoid confusion with low-level "section" names.
Signed-off-by: Luca Burelli <l.burelli@arduino.cc>
Change exception code to use LOG_ERR and align with all other
architectures when CONFIG_EXCEPTION_DEBUG is enabled.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
It should be possible to disable exception debug, which is enabled by
default to reduce image size. Add missing guards now that the option is
cross architecture.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Rename xtensa_asm2.c to have a more meaningful name to actually
reflect the content of the file. This file is mostly about
handling interrupts and exceptions (via the predefined vectors
in Xtensa core).
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Fold z_arch_get_next_switch_handle() into return_to(). This is
not exactly an arch interface, and is simple enough to be
moved into return_to().
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This gets rid of the z_ prefix.
Note that z_xt_*() are being used by the HAL so they cannot be
renamed.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This follows the idea to remove any z_ prefix. Since MMU has
a large number of these, separate out these changes into one
commit to ease review effort.
Since these are no longer have z_, these need proper doxygen
doc. So add them too.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
z_xtensa_dump_stack() and z_xtensa_exccause() are both arch
internal functions that should not be exposed in public API.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Header files under arch/xtensa/include are considered internal
to architecture. There is really no need for two places to
house architecture internal header files.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
... from xtensa_asm2.c. Other architectures have
z_irq_spurious() and *_irq_is_enabled() test in irq_manage.c.
So follow the trend here.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
... from xtensa_asm2.c.
Everything has been stuffed inside xtensa_asm2.c where
they are all mangled together. So extract thread related
stuff into its own file.
Note that arch_float_*() may not be thread related but
most other architectures put them into thread.c. So we
also do it here.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
arch_spin_relax() does not really fit into the scheme of
xtensa_asm2.c as it is mainly about handling interrupts
and exceptions. So move it into smp.c, similar to other
architectures which arch_spin_relax() defined.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
