libc: newlib: Grant access to dynamic locks from all threads
Fixes: zephyrproject-rtos#67504
The following error is produced when using the following configuration
```
CONFIG_USERSPACE=y
CONFIG_NEWLIB_LIBC=y
CONFIG_NEWLIB_LIBC_NANO=n
CONFIG_DYNAMIC_OBJECTS=y
CONFIG_HEAP_MEM_POOL_SIZE=256
CONFIG_THREAD_LOCAL_STORAGE=y
CONFIG_MAIN_STACK_SIZE=2048
```
```
os: thread 0x301a2950 (-1) does not have permission on k_mutex 0x301aaca4
os: permission bitmap
01 00 |..
os: syscall z_vrfy_k_mutex_lock failed check: access denied
os: r0/a1: 0x00000000 r1/a2: 0x00000000 r2/a3: 0x00000000
os: r3/a4: 0x00000000 r12/ip: 0x00000000 r14/lr: 0x00000000
os: xpsr: 0x00000000
os: s[ 0]: 0x00000000 s[ 1]: 0x00000000 s[ 2]: 0x00000000 s[ 3]: 0x00000000
os: s[ 4]: 0x00000000 s[ 5]: 0x00000000 s[ 6]: 0x00000000 s[ 7]: 0x00000000
os: s[ 8]: 0x00000000 s[ 9]: 0x00000000 s[10]: 0x00000000 s[11]: 0x00000000
os: s[12]: 0x00000000 s[13]: 0x00000000 s[14]: 0x00000000 s[15]: 0x00000000
os: fpscr: 0x00000000
os: Faulting instruction address (r15/pc): 0xee7fdb7d
os: >>> ZEPHYR FATAL ERROR 3: Kernel oops on CPU 0
os: Current thread: 0x301a2950
```
This bug caused by a global mutex used by _vfprintf_r()
which is initialized in __sinit() and located in z_malloc_partition
not being granted access to be used my multiple user threads
despite each user thread being granted read and write permission
to the z_malloc_partition.
Here is a sample main.c to reproduce the bug
```
extern struct k_mem_partition z_libc_partition;
extern struct k_mem_partition z_malloc_partition;
static k_tid_t tids[TEST_THDS];
static struct k_thread tcbs[TEST_THDS];
static struct k_mem_domain domains[TEST_THDS];
static K_THREAD_STACK_ARRAY_DEFINE(thd_stacks, TEST_THDS, 2048);
static int forbidden_global_data = 2;
void thread_worker(void* a, void* b, void* c) {
(void)b;
(void)c;
printf("thd %d started\n", (int)(intptr_t)a);
k_sleep(K_MSEC(1000));
forbidden_global_data++; /* This should cause an MPU Fault */
}
int main(void) {
forbidden_global_data = 1;
struct k_mem_partition* share_parts[2] = {
&z_libc_partition,
&z_malloc_partition,
};
for (int i = 0l; i < TEST_THDS; i++) {
tids[i] = k_thread_create(
&tcbs[i],
thd_stacks[i],
K_THREAD_STACK_SIZEOF(thd_stacks[i]),
thread_worker,
(void*)(intptr_t)i,
NULL,
NULL,
5,
K_USER,
K_FOREVER);
k_mem_domain_init(&domains[i], 2, share_parts);
k_mem_domain_add_thread(&domains[i], tids[i]);
k_thread_start(tids[i]);
}
for (int i = 0; i < TEST_THDS; i++) {
k_thread_join(tids[i], K_FOREVER);
}
return 0;
}
```
Signed-off-by: James Ogier <jogier@meta.com>
ffa49a8240