Add pm cpu ops to call the platform specific implementations for
bringing up secondary cores.
Signed-off-by: Lingutla Chandrasekhar <quic_lingutla@quicinc.com>
RISC-V Spec requires minimum alignment of trap handling code to be
dependent from MTVEC.BASE field size. Minimum alignment for RISC-V
platforms is 4 bytes, but maximum is platform or application-specific.
Currently there is no common approach to align the trap handling
code for RISC-V and some platforms use custom wrappers to align
_isr_wrapper properly.
This change introduces a generic solution,
CONFIG_RISCV_TRAP_HANDLER_ALIGNMENT configuration option which sets
the alignment of a RISC-V trap handling code.
The existing custom solutions for some platforms remain operational,
since the default alignment is set to minimal possible (4 bytes)
and will be overloaded by potentially larger alignment of custom solutions.
Signed-off-by: Alexander Razinkov <alexander.razinkov@syntacore.com>
This header is private and included only in architecture code, no need for
it to be in the top of the public include directory.
Note: This might move to a more private location later. For now just
cleaning up the obvious issues.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
In print_pmp_entries(), start and end are function arguments
and yet another start and end are declared inside the for
loop. So rename the function arguments to fix shadow variables
warning.
The changes in csr_*() macros are needed to avoid shadowing
__v when nesting those functions together, for example,
csr_write(..., csr_read(...)).
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Add Kconfig RISCV_SOC_HAS_CUSTOM_SYS_IO symbol so that a riscv
SoC can set to specify that it has a custom implementation for
sys_io functions.
Signed-off-by: Yong Cong Sin <ycsin@meta.com>
This adds a few line use zephyr_syscall_header() to include
headers containing syscall function prototypes.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Some RISCV platforms shipping a CLIC have a peculiar interrupt ID
ordering / mapping.
According to the "Core-Local Interrupt Controller (CLIC) RISC-V
Privileged Architecture Extensions" Version 0.9-draft at paragraph 16.1
one of these ordering recommendations is "CLIC-mode interrupt-map for
systems retaining interrupt ID compatible with CLINT mode" that is
described how:
The CLINT-mode interrupts retain their interrupt ID in CLIC mode.
[...]
The existing CLINT software interrupt bits are primarily intended for
inter-hart interrupt signaling, and so are retained for that purpose.
[...]
CLIC interrupt inputs are allocated IDs beginning at interrupt ID
17. Any fast local interrupts that would have been connected at
interrupt ID 16 and above should now be mapped into corresponding
inputs of the CLIC.
That is a very convoluted way to say that interrupts 0 to 15 are
reserved for internal use and CLIC only controls interrupts reserved for
platform use (16 up to n + 16, where n is the maximum number of
interrupts supported).
Let's now take now into consideration this situation in the DT:
clic: interrupt-controller {
...
};
device0: some-device {
interrupt-parent = <&clic>;
interrupts = <0x1>;
...
};
and in the driver for device0:
IRQ_CONNECT(DT_IRQN(node), ...);
From the hardware prospective:
(1a) device0 is using the first IRQ line of the CLIC
(2a) the interrupt ID / exception code of the `MSTATUS` register
associated to this IRQ is 17, because the IDs 0 to 15 are reserved
From the software / Zephyr prospective:
(1b) Zephyr is installing the IRQ vector into the SW ISR table (and into
the IRQ vector table for DIRECT ISRs in case of CLIC vectored mode)
at index 0x1.
(2b) Zephyr is using the interrupt ID of the `MSTATUS` register to index
into the SW ISR table (or IRQ vector table)
It's now clear how (2a) and (2b) are in contrast with each other.
To fix this problem we have to take into account the offset introduced
by the reserved interrupts. To do so we introduce
CONFIG_RISCV_RESERVED_IRQ_ISR_TABLES_OFFSET as hidden option for the
platforms to set.
This Kconfig option is used to shift the interrupt numbers when
installing the IRQ vector into the SW ISR table and/or IRQ vector table.
So for example in the previous case and using
CONFIG_RISCV_RESERVED_IRQ_ISR_TABLES_OFFSET == 16, the IRQ vector
associated to the device0 would be correctly installed at index 17 (16 +
1), matching what is reported by the `MSTATUS` register.
CONFIG_NUM_IRQS must be increased accordingly.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Before adding support for the CLIC vectored mode, rename
CONFIG_RISCV_MTVEC_VECTORED_MODE to CONFIG_RISCV_VECTORED_MODE to be
more generic and eventually include also the CLIC vectored mode.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
This will avoid unconditionally pulling z_riscv_switch() into the build
as it is not used, reducing the resulting binary some more.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
This reverts commit f0b458a619.
This is a pointless change that simply increases footprint.
Existing code already supports compilation without multithreading.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Allow builds which has CONFIG_MULTITHREADING disabled.
This is reduce code footprint which is handy for
constrained targets as bootloaders.
Signed-off-by: Marek Matej <marek.matej@espressif.com>
This make MCUboot build as Zephyr application.
Providing optinal 2nd stage bootloader to the
IDF bootloader, which is used by default.
This provides more flexibility when building
and loading multiple images and aims to
brings better DX to users by using the sysbuild.
MCUboot and applications has now separate
linker scripts.
Signed-off-by: Marek Matej <marek.matej@espressif.com>
Let's consider CPU1 waiting on a spinlock already taken by CPU2.
It is possible for CPU2 to invoke the FPU and trigger an FPU exception
when the FPU context for CPU2 is not live on that CPU. If the FPU context
for the thread on CPU2 is still held in CPU1's FPU then an IPI is sent
to CPU1 asking to flush its FPU to memory.
But if CPU1 is spinning on a lock already taken by CPU2, it won't see
the pending IPI as IRQs are disabled. CPU2 won't get its FPU state
restored and won't complete the required work to release the lock.
Let's prevent this deadlock scenario by looking for a pending FPU IPI
from the arch_spin_relax() hook and honor it.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Enable single-threading support for the riscv architecture.
Add z_riscv_switch_to_main_no_multithreading function for
supporting single-threading.
The single-threading does not work with enabling PMP_STACK_GUARD.
It is because single-threading does not use context-switching.
But the privileged mode transition that PMP depends on implicitly
presupposes using context-switching. It is a contradiction.
Thus, disable PMP_STACK_GUARD when MULTITHREADING is not enabled.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@fujitsu.com>
use helper macros from csr.h instead of inline assembly which results
in cleaner and more maintainable code
Signed-off-by: Manojkumar Subramaniam <manoj@electrolance.com>
The init infrastructure, found in `init.h`, is currently used by:
- `SYS_INIT`: to call functions before `main`
- `DEVICE_*`: to initialize devices
They are all sorted according to an initialization level + a priority.
`SYS_INIT` calls are really orthogonal to devices, however, the required
function signature requires a `const struct device *dev` as a first
argument. The only reason for that is because the same init machinery is
used by devices, so we have something like:
```c
struct init_entry {
int (*init)(const struct device *dev);
/* only set by DEVICE_*, otherwise NULL */
const struct device *dev;
}
```
As a result, we end up with such weird/ugly pattern:
```c
static int my_init(const struct device *dev)
{
/* always NULL! add ARG_UNUSED to avoid compiler warning */
ARG_UNUSED(dev);
...
}
```
This is really a result of poor internals isolation. This patch proposes
a to make init entries more flexible so that they can accept sytem
initialization calls like this:
```c
static int my_init(void)
{
...
}
```
This is achieved using a union:
```c
union init_function {
/* for SYS_INIT, used when init_entry.dev == NULL */
int (*sys)(void);
/* for DEVICE*, used when init_entry.dev != NULL */
int (*dev)(const struct device *dev);
};
struct init_entry {
/* stores init function (either for SYS_INIT or DEVICE*)
union init_function init_fn;
/* stores device pointer for DEVICE*, NULL for SYS_INIT. Allows
* to know which union entry to call.
*/
const struct device *dev;
}
```
This solution **does not increase ROM usage**, and allows to offer clean
public APIs for both SYS_INIT and DEVICE*. Note that however, init
machinery keeps a coupling with devices.
**NOTE**: This is a breaking change! All `SYS_INIT` functions will need
to be converted to the new signature. See the script offered in the
following commit.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
init: convert SYS_INIT functions to the new signature
Conversion scripted using scripts/utils/migrate_sys_init.py.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
manifest: update projects for SYS_INIT changes
Update modules with updated SYS_INIT calls:
- hal_ti
- lvgl
- sof
- TraceRecorderSource
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
tests: devicetree: devices: adjust test
Adjust test according to the recently introduced SYS_INIT
infrastructure.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
tests: kernel: threads: adjust SYS_INIT call
Adjust to the new signature: int (*init_fn)(void);
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Looks like some implementors decided not to implement the full set of
PMP range matching modes. Let's rearrange the code so that any of those
modes can be disabled.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Let's honor CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT even for kernel
stacks. This saves one global PMP slot when creating the guard area for
the IRQ stack, and some hw implementations might require that anyway.
With this changes, arch_mem_domain_max_partitions_get() becomes much
more reliable and tests/kernel/mem_protect is more likely to pass even
with the stack guard enabled.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
When CONFIG_SOC_ISR_SW_UNSTACKING is defined, it's up to the custom soc
code to remove the ESF, because the software-managed part of the ESF is
depending on the hardware. Fix this in the ISR code.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Some implementations may not capture the faulting instruction in mtval
and set it to zero when an illegal instruction fault is raised This is
notably the case with QEMU version 7.0.0 when a CSR instruction is
involved.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
The FRCSR, FSCSR, FRRM, FSRM, FSRMI, FRFLAGS, FSFLAGS and FSFLAGSI
are in fact CSR instructions targeting the fcsr, frm and fflags
registers. They should be caught as FPU instructions as well.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
- IRQ state for the interrupted context corresponds to the PIE bit not
the IE bit.
- Restoring the saved FPU state should clear the entire field before
or'ing wanted bits in.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
For RISC-V, the reg property of a cpu node in the devicetree describes
the low level unique ID of each hart. Using devicetree macro's, a list
of all cpus with status "okay" can be generated.
Using devicetree overlays, a hart or multiple harts can be marked as
"disabled", thus excluding them from the list. This allows platforms
that have non-zero indexed SMP capable harts to be functionally mapped
to Zephyr's sequential CPU numbering scheme.
On kernel init, if the application has MP_MAX_NUM_CPUS greater than 1,
generate the list of cpu nodes from the device tree with status "okay"
and map the unique hartid's to zephyr cpu's
While we are at it, as the hartid is the value that gets passed to
z_riscv_secondary_cpu_init, use that as the variable name instead of
cpu_num
Signed-off-by: Conor Paxton <conor.paxton@microchip.com>
RISC-V multi-hart systems that employ a heterogeneous core complex are
not guaranteed to have the smp capable harts starting with a unique id
of zero, matching Zephyr's sequential zero indexed cpu numbering scheme.
Add option, RV_BOOT_HART to choose the hart to boot from.
On reset, check the current hart equals RV_BOOT_HART: if so, boot first
core. else, loop in the boot secondary core and wait to be brought up.
For multi-hart systems that are not running a Zephyr mp or smp
application, park the non zephyr related harts in a wfi loop.
Signed-off-by: Conor Paxton <conor.paxton@microchip.com>
Commit 4f9b547ebd ("riscv: smp: prepare for more than one IPI type")
didn't clear pending IPI flags.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
We can leverage the FPU dirty state as an indicator for preemptively
reloading the FPU content when a thread that did use the FPU before
being scheduled out is scheduled back in. This avoids the FPU access
trap overhead when switching between multiple threads with heavy FPU
usage.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
FPU context switching is always performed on demand through the FPU
access exception handler. Actual task switching only grants or denies
FPU access depending on the current FPU owner.
Because RISC-V doesn't have a dedicated FPU access exception, we must
catch the Illegal Instruction exception and look for actual FP opcodes.
There is no longer a need to allocate FPU storage on the stack for every
exception making esf smaller and stack overflows less likely.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Instead of saving/restoring FPU content on every exception and task
switch, this replaces FPU sharing support with a "lazy" (on-demand)
context switching algorithm similar to the one used on ARM64.
Every thread starts with FPU access disabled. On the first access the
FPU trap is invoked to:
- flush the FPU content to the previous thread's memory storage;
- restore the current thread's FPU content from memory.
When a thread loads its data in the FPU, it becomes the FPU owner.
FPU content is preserved across task switching, however FPU access is
either allowed if the new thread is the FPU owner, or denied otherwise.
A thread may claim FPU ownership only through the FPU trap. This way,
threads that don't use the FPU won't force an FPU context switch.
If only one running thread uses the FPU, there will be no FPU context
switching to do at all.
It is possible to do FP accesses in ISRs and syscalls. This is not the
norm though, so the same principle is applied here, although exception
contexts may not own the FPU. When they access the FPU, the FPU content
is flushed and the exception context is granted FPU access for the
duration of the exception. Nested IRQs are disallowed in that case to
dispense with the need to save and restore exception's FPU context data.
This is the core implementation only to ease reviewing. It is not yet
hooked into the build.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Right now this is hardcoded to z_sched_ipi(). Make it so that other IPI
services can be added in the future.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Given the Zephyr CPU number is no longer tied to the hartid, we must
consider the actual hartid when sending an IPI to a given CPU. Since
those hartids can be anything, let's just save them in the cpu structure
as each CPU is brought online.
While at it, throw in some `get_hart_msip()` cleanups.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Currently it is assumed that Zephyr CPU numbers match their hartid
value one for one. This assumption was relied upon to efficiently
retrieve the current CPU's `struct _cpu` pointer.
People are starting to have systems with a mix of different usage for
each CPU and such assumption may no longer be true.
Let's completely decouple the hartid from the Zephyr CPU number by
stuffing each CPU's `struct _cpu` pointer in their respective scratch
register instead. `arch_curr_cpu()` becomes more efficient as well.
Since the scratch register was previously used to store userspace's
exception stack pointer, that is now moved into `struct _cpu_arch`
which implied minor user space entry code cleanup and rationalization.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
RISC-V has a modular design. Some hardware with a custom interrupt
controller needs a bit more work to lock / unlock IRQs.
Account for this hardware by introducing a set of new
z_soc_irq_* functions that can override the default behaviour.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Some RISC-V SoCs implement a mechanism for hardware supported stacking /
unstacking of registers during ISR / exceptions. What happens is that on
ISR / exception entry part of the context is automatically saved by the
hardware on the stack without software intervention, and the same part
of the context is restored by the hardware usually on mret.
This is currently not yet supported by Zephyr, where the full context
must be saved by software in the full fledged ESF. This patcheset is
trying to address exactly this case.
At least three things are needed to support in a general fashion this
problem: (1) a way to store in software only the part of the ESF not
already stacked by hardware, (2) a way to restore in software only the
part of the context that is not going to be restored by hardware and (3)
a way to define a custom ESF.
Point (3) is important because the full ESF frame is now composed by a
custom part depending on the hardware (that can choose which register to
stack / unstack and the order they are saved onto the stack) and a part
defined in software for the remaining part of the context.
In this patch a new CONFIG_RISCV_SOC_HAS_ISR_STACKING is introduced that
enables the code path supporting the three points by the mean of three
macros that must be implemented by the user in a soc_stacking.h file:
SOC_ISR_SW_STACKING, SOC_ISR_SW_UNSTACKING and SOC_ISR_STACKING_ESF
(refer to the symbol help for more details).
This is an example of soc_isr_stacking.h for an hardware that doesn't do
any hardware stacking / unstacking but everything is managed in
software:
#ifndef __SOC_ISR_STACKING
#define __SOC_ISR_STACKING
#if !defined(_ASMLANGUAGE)
#define SOC_ISR_STACKING_ESF_DECLARE \
struct __esf { \
unsigned long ra; \
unsigned long t0; \
unsigned long t1; \
unsigned long t2; \
unsigned long t3; \
unsigned long t4; \
unsigned long t5; \
unsigned long t6; \
unsigned long a0; \
unsigned long a1; \
unsigned long a2; \
unsigned long a3; \
unsigned long a4; \
unsigned long a5; \
unsigned long a6; \
unsigned long a7; \
unsigned long mepc; \
unsigned long mstatus; \
unsigned long s0; \
} __aligned(16)
#else
#define SOC_ISR_SW_STACKING \
addi sp, sp, -__z_arch_esf_t_SIZEOF; \
DO_CALLER_SAVED(sr);
#define SOC_ISR_SW_UNSTACKING \
DO_CALLER_SAVED(lr);
#endif /* _ASMLANGUAGE */
#endif /* __SOC_ISR_STACKING */
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
This reverts commit a7b5d606c7.
The assumption behind that commit was wrong. The software-based stack
sentinel writes to the very bottom of the _writable_ stack area i.e.
right next to the actual PMP based guard area. So they are compatible.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Change for loops of the form:
for (i = 0; i < CONFIG_MP_NUM_CPUS; i++)
...
to
unsigned int num_cpus = arch_num_cpus();
for (i = 0; i < num_cpus; i++)
...
We do the call outside of the for loop so that it only happens once,
rather than on every iteration.
Signed-off-by: Kumar Gala <kumar.gala@intel.com>
Change automated searching for files using "IRQ_CONNECT()" API not
including <zephyr/irq.h>.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The interrupt stack is used as the system stack during kernel
initialization while IRQs are not yet enabled. The sp register is
set to z_interrupt_stacks + CONFIG_ISR_STACK_SIZE.
CONFIG_ISR_STACK_SIZE only represents the desired usable stack size.
This does not take into account the added guard area. Result is a stack
whose pointer is much closer to the trigger zone than expected when
CONFIG_PMP_STACK_GUARD=y, and the SMP configuration in particular pushes
it over the edge during many CI test cases.
Worse: during early init we're not quite ready to handle exceptions
yet and complete havoc ensues with no meaningful debugging output.
Make sure the early assembly code locates the actual top of the stack
by generating a constant with its true size.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
The software-based stack sentinel writes to the very bottom of the
stack area triggering the PMP stack protection. Obviously they can't
be used together.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
The IRQ stack in particular is different on each CPU, and so is its
stack guard PMP entry value. This creates 2 issues:
- The assertion ensuring the last global PMP address is the same
for each CPU does fail;
- That last global PMP address can't be relied upon to create a
single-slot per-thread TOR mapping.
Fix both issues by not remembering the actual address for the last
global entry but a dummy address instead that is guaranteed not to
match any opportunistic single-slot TOR mapping.
While at it, lock that IRQ stack guard PMP entry.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Z_THREAD_STACK_BUFFER() must not be used here. This is meant for stacks
defined with K_THREAD_STACK_ARRAY_DEFINE() whereas in this case we are
given a stack created with K_KERNEL_STACK_ARRAY_DEFINE().
If CONFIG_USERSPACE=y then K_THREAD_STACK_RESERVED gets defined with
a bigger value than K_KERNEL_STACK_RESERVED. Then Z_THREAD_STACK_BUFFER()
returns a pointer that is more advanced than expected, resulting in a
stack pointer outside its actual stack area and therefore memory
corruption ensues.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
