1. add IS_ALIGNED() definitions for NuttX;
2. replace all the ALIGN_UP() and ALIGN_DOWN() to use common
align implementation;
Signed-off-by: Bowen Wang <wangbowen6@xiaomi.com>
When using alarm_arch implementation, 64-bit time can be returned. Using unsign long will cause precision loss.
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
Correct the alert message in `esp_setup_irq()` if
irq number allocation fails, the parameter number is not
matched with format specifier.
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
Some app with same code runs on different cores in AMP mode,
need the physical core on which the function is called.
Signed-off-by: hujun5 <hujun5@xiaomi.com>
Signed-off-by: fangxinyong <fangxinyong@xiaomi.com>
riscv_internal.h is used literally everywhere, while the SBI definitions
are needed only by whomever needs the services.
Having the SBI definitions:
a) Copied from OpenSBI (why has this been done? even the names are same)
b) Presented publicly to 99% of risc-v modules
creates a build error when building with OpenSBI, due to duplicate
definitions of the SBI service identifiers:
In file included from /nuttx/arch/risc-v/src/common/riscv_internal.h:40,
from /nuttx/arch/risc-v/src/chip/chip.h:32,
from board/mpfs_domain.c:30:
/nuttx/arch/risc-v/src/common/riscv_sbi.h:36: error: "SBI_EXT_BASE" redefined [-Werror]
36 | #define SBI_EXT_BASE 0x00000010
and so forth...
Fix this by removing riscv_sbi.h i.e. not exposing the ABI publicly.
reason:
To remove the "sync pause" and decouple the critical section from the dependency on enabling interrupts,
after that we need to further implement "schedlock + spinlock".
changelist
1 Modify the implementation of critical sections to no longer involve enabling interrupts or handling synchronous pause events.
2 GIC_SMP_CPUCALL attach to pause handler to remove arch interface up_cpu_paused_restore up_cpu_paused_save
3 Completely remove up_cpu_pause, up_cpu_resume, up_cpu_paused, and up_cpu_pausereq
4 change up_cpu_pause_async to up_send_cpu_sgi
Signed-off-by: hujun5 <hujun5@xiaomi.com>
tools/configure.sh esp32c3-generic:wifi
wapi scan wlan0
has no return result after couple times
rootcause:
mq_timedsend() return failed because the time valid check error.
the ts_nsec is bigger than 1000000000.
esp_update_time() hasn't consider of the ns > 1s after the adding
calculation
Signed-off-by: ligd <liguiding1@xiaomi.com>
reason:
In the kernel, we are planning to remove all occurrences of up_cpu_pause as one of the steps to
simplify the implementation of critical sections. The goal is to enable spin_lock_irqsave to encapsulate critical sections,
thereby facilitating the replacement of critical sections(big lock) with smaller spin_lock_irqsave(small lock)
Signed-off-by: hujun5 <hujun5@xiaomi.com>
Unwinding the kernel stack did not work previously due to the way the task
startup logic works via nxtask_start and the up_task_start() system call.
After modifying the logic behind those, the kernel stack is in fact fully
unwound when return_from_exception is executed, so calling the original
hack "riscv_current_ksp" is not necessary anymore.
This removes 2 reserved system calls and replaces them with an ASM snippet.
The result removes an unnecessary ecall from the process startup logic, as
well as ensures the stacks are FULLY unwound when the user process starts.
The logic is ported from ARM64.
Port the simplification from ARM64, this removes the ugly inline assembly
trampoline "do_syscall" and replaces it with a simple table lookup and
call via function pointer.
add_compile_options(--param asan-globals=1) is recognized as
--param-lasan-globals=1, which causes compilation exception:
Signed-off-by: W-M-R <Mike_0528@163.com>
1. refactor the ghs/gcc/clang/armclang toolchain management in CMake
2. unify the cmake toolchain naming style
3. support greenhills build procedure with CMake
4. add protect build for greenhills and gnu toolchain with CMake
Signed-off-by: guoshichao <guoshichao@xiaomi.com>
Support customize idle loop by CONFIG_ARCH_IDLE_CUSTOM
as other architectures.
Then user can provide their own `up_idle()` function
with CONFIG_ARCH_IDLE_CUSTOM enabled.
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
reason:
Currently, if we need to schedule a task to another CPU, we have to completely halt the other CPU,
manipulate the scheduling linked list, and then resume the operation of that CPU. This process is both time-consuming and unnecessary.
During this process, both the current CPU and the target CPU are inevitably subjected to busyloop.
The improved strategy is to simply send a cross-core interrupt to the target CPU.
The current CPU continues to run while the target CPU responds to the interrupt, eliminating the certainty of a busyloop occurring.
Signed-off-by: hujun5 <hujun5@xiaomi.com>
This commit fixes the regression from https://github.com/apache/nuttx/pull/13561
In order to determine whether a context switch has occurred,
we can use g_running_task to store the current regs.
This allows us to compare the current register state with the previously
stored state to identify if a context switch has taken place.
Signed-off-by: hujun5 <hujun5@xiaomi.com>
We need to record the parent's integer register context upon exception
entry to a separate non-volatile area. Why?
Because xcp.regs can move due to a context switch within the fork() system
call, be it either via interrupt or a synchronization point.
Fix this by adding a "sregs" area where the saved user context is placed.
The critical section within fork() is also unnecessary.
There was an error in the fork() routine when system calls are in use:
the child context is saved on the child's user stack, which is incorrect,
the context must be saved on the kernel stack instead.
The result is a full system crash if (when) the child executes on a
different CPU which does not have the same MMU mappings active.
for the citimon stats:
thread 0: thread 1:
enter_critical (t0)
up_switch_context
note suspend thread0 (t1)
thread running
IRQ happen, in ISR:
post thread0
up_switch_context
note resume thread0 (t2)
ISR continue f1
ISR continue f2
...
ISR continue fn
leave_critical (t3)
You will see, the thread 0, critical_section time is:
(t1 - t0) + (t3 - t2)
BUT, this result contains f1 f2 .. fn time spent, it is wrong
to tell user thead0 hold the critical lots of time but actually
not belong to it.
Resolve:
change the nxsched_suspend/resume_scheduler to real hanppends
Signed-off-by: ligd <liguiding1@xiaomi.com>
If CONFIG_SMP is not enabled, riscv_cpuindex.c will not be compiled
anyway.
And for CONFIG_ARCH_FPU, if it's not enabled, riscv_fpucmp.c will not
be compiled.
So we can remove the unnecessary macro guard for up_cpu_index() and
up_fpucmp().
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
This is actually the same table as entrypoints, so just use the same data, which
can be set before booting any of the harts
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
The logical CPU index should be retrieved with this_cpu(); the
riscv_mhartid() returns the actual hart id of the SoC.
For mpfs target for example, NuttX can run on a single HART, for example on mhartid 2, but there is still just one logical CPU for the NuttX.
Signed-off-by: Jukka Laitinen <jukkax@ssrc.tii.ae>
Previously, address 0x03b70000u was written with shift bits
that only changed the bit width, not the mode. HS mode is
changed via 0x03B90100, which is required, according to Jedec
specs, for DDR mode. HS mode was not applied before. Enforce
DDR mode (50 MHz) for now.
The real boost, however, comes from removing the DMA limitation
at 0x08xxxxxx address space, which now seems unnecessary.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
hujun5