mpu protected mode debug

[xiaotailang]

Under mpu protected mode, I added a new system call, But appeared a panic , and I tried to debug, but when using gdb tracing, only the code in kernel mode can be traced, and the code in application mode cannot be traced. So I want to know in mpu protected mode, how does gdb track the code of application state and kernel state at the same time.

[patacongo]

Can't you just add the symbol table for the user space to GDB using the add-symbol-file command?

[superxyxiao]

Is there any document about adding new system calls under the protection of mpu? I add a new system call in the following file, but when debugging, I found that the address of the new system call in the g_stublookup array is 0.

path/syscall/syscall_lookup.h:SYSCALL_LOOKUP(mycall, 0, STUB_mycall)
path/syscall/syscall_stublookup.c:uintptr_t STUB_mycall(int nbr); path/syscall/syscall.csv:"mycall","nuttx/mycall.h","","int"

cat path/include/nuttx/mycall.h

ifndef ASSEMBLY

undef EXTERN

if defined(__cplusplus)

define EXTERN extern "C"

extern "C" {

else

define EXTERN extern

endif

int mycall(void);

undef EXTERN

ifdef __cplusplus

}

endif

path/include/sys/syscall.h

define SYS_mycall (SYS_prctl + 1)

define SYS_maxsyscall (SYS_prctl + 2)

cat path/sched/mycall/mycall.c

include <nuttx/compiler.h>

include <nuttx/mycall.h>

int mycall(void) { int num=0; num=num+2; return num; }

[superxyxiao]

I use version is nuttx-8.2

[xiaoxiang781216]

@superxyxiao nuttx-8.2 is a very old release(2019/11), why not use the last dev branch?