static void _Thread_Make_zombie( Thread_Control *the_thread )
{
ISR_lock_Context lock_context;
Thread_Zombie_control *zombies = &_Thread_Zombies;
if ( _Thread_Owns_resources( the_thread ) ) {
_Terminate(
INTERNAL_ERROR_CORE,
false,
INTERNAL_ERROR_RESOURCE_IN_USE
);
}
_Objects_Close(
_Objects_Get_information_id( the_thread->Object.id ),
&the_thread->Object
);
_Thread_Set_state( the_thread, STATES_ZOMBIE );
_Thread_queue_Extract_with_proxy( the_thread );
_Watchdog_Remove_ticks( &the_thread->Timer );
_ISR_lock_ISR_disable_and_acquire( &zombies->Lock, &lock_context );
_Chain_Append_unprotected( &zombies->Chain, &the_thread->Object.Node );
_ISR_lock_Release_and_ISR_enable( &zombies->Lock, &lock_context );
}
bool _Objects_MP_Allocate_and_open (
Objects_Information *information,
uint32_t the_name, /* XXX -- wrong for variable */
Objects_Id the_id,
bool is_fatal_error
)
{
Objects_MP_Control *the_global_object;
the_global_object = _Objects_MP_Allocate_global_object();
if ( _Objects_MP_Is_null_global_object( the_global_object ) ) {
if ( is_fatal_error == false )
return false;
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_OUT_OF_GLOBAL_OBJECTS
);
}
_Objects_MP_Open( information, the_global_object, the_name, the_id );
return true;
}
int rtems_gxx_setspecific(__gthread_key_t key, const void *ptr)
{
pthread_key_t *pkey = key;
int eno;
if ( pkey == NULL ) {
return EINVAL;
}
eno = pthread_setspecific( *pkey, ptr );
#ifdef DEBUG_GXX_WRAPPERS
printk(
"gxx_wrappers: setspecific key=%x, ptr=%x, id=%x\n",
pkey,
ptr,
rtems_task_self()
);
#endif
if ( eno != 0 ) {
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_GXX_KEY_ADD_FAILED
);
}
return 0;
}
void _Objects_MP_Close (
Objects_Information *information,
Objects_Id the_id
)
{
Chain_Control *the_chain;
Chain_Node *the_node;
Objects_MP_Control *the_object;
the_chain = &information->global_table[ _Objects_Get_node( the_id ) ];
for ( the_node = _Chain_First( the_chain ) ;
!_Chain_Is_tail( the_chain, the_node ) ;
the_node = _Chain_Next( the_node ) ) {
the_object = (Objects_MP_Control *) the_node;
if ( _Objects_Are_ids_equal( the_object->Object.id, the_id ) ) {
_Chain_Extract( the_node );
_Objects_MP_Free_global_object( the_object );
return;
}
}
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_INVALID_GLOBAL_ID
);
}
/*
* MUTEX support
*/
void rtems_gxx_mutex_init (__gthread_mutex_t *mutex)
{
rtems_status_code status;
#ifdef DEBUG_GXX_WRAPPERS
printk( "gxx_wrappers: mutex init =%X\n", *mutex );
#endif
status = rtems_semaphore_create(
rtems_build_name ('G', 'C', 'C', '2'),
1,
RTEMS_PRIORITY|RTEMS_BINARY_SEMAPHORE|
RTEMS_INHERIT_PRIORITY|RTEMS_NO_PRIORITY_CEILING|RTEMS_LOCAL,
0,
(rtems_id *)mutex
);
if ( status != RTEMS_SUCCESSFUL ) {
#ifdef DEBUG_GXX_WRAPPERS
printk(
"gxx_wrappers: mutex init failed %s (%d)\n",
rtems_status_text(status),
status
);
#endif
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_GXX_MUTEX_INIT_FAILED
);
}
#ifdef DEBUG_GXX_WRAPPERS
printk( "gxx_wrappers: mutex init complete =%X\n", *mutex );
#endif
}
void _MPCI_Internal_packets_Process_packet (
MP_packet_Prefix *the_packet_prefix
)
{
MPCI_Internal_packet *the_packet;
uint32_t maximum_nodes;
uint32_t maximum_global_objects;
the_packet = (MPCI_Internal_packet *) the_packet_prefix;
switch ( the_packet->operation ) {
case MPCI_PACKETS_SYSTEM_VERIFY:
maximum_nodes = the_packet->maximum_nodes;
maximum_global_objects = the_packet->maximum_global_objects;
if ( maximum_nodes != _Objects_Maximum_nodes ||
maximum_global_objects != _Objects_MP_Maximum_global_objects ) {
_MPCI_Return_packet( the_packet_prefix );
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_INCONSISTENT_MP_INFORMATION
);
}
_MPCI_Return_packet( the_packet_prefix );
break;
}
}
static void _Thread_Make_zombie( Thread_Control *the_thread )
{
if ( _Thread_Owns_resources( the_thread ) ) {
_Terminate(
INTERNAL_ERROR_CORE,
false,
INTERNAL_ERROR_RESOURCE_IN_USE
);
}
_Objects_Close(
_Objects_Get_information_id( the_thread->Object.id ),
&the_thread->Object
);
_Thread_Set_state( the_thread, STATES_ZOMBIE );
_Thread_queue_Extract_with_proxy( the_thread );
_Thread_Timer_remove( the_thread );
/*
* Add the thread to the thread zombie chain before we wake up joining
* threads, so that they are able to clean up the thread immediately. This
* matters for SMP configurations.
*/
_Thread_Add_to_zombie_chain( the_thread );
_Thread_Wake_up_joining_threads( the_thread );
}
void _Thread_queue_Deadlock_fatal( Thread_Control *the_thread )
{
_Terminate(
INTERNAL_ERROR_CORE,
false,
INTERNAL_ERROR_THREAD_QUEUE_DEADLOCK
);
}
void _ISR_Handler_initialization( void )
{
_ISR_Nest_level = 0;
#if (CPU_SIMPLE_VECTORED_INTERRUPTS == TRUE)
_ISR_Vector_table = _Workspace_Allocate_or_fatal_error(
sizeof(ISR_Handler_entry) * ISR_NUMBER_OF_VECTORS
);
_CPU_Initialize_vectors();
#endif
#if ( CPU_ALLOCATE_INTERRUPT_STACK == TRUE )
{
size_t stack_size = rtems_configuration_get_interrupt_stack_size();
uint32_t max_cpus = rtems_configuration_get_maximum_processors();
uint32_t cpu;
if ( !_Stack_Is_enough( stack_size ) )
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_INTERRUPT_STACK_TOO_SMALL
);
for ( cpu = 0 ; cpu < max_cpus; ++cpu ) {
Per_CPU_Control *per_cpu = _Per_CPU_Get_by_index( cpu );
void *low = _Workspace_Allocate_or_fatal_error( stack_size );
void *high = _Addresses_Add_offset( low, stack_size );
#if (CPU_STACK_ALIGNMENT != 0)
high = _Addresses_Align_down( high, CPU_STACK_ALIGNMENT );
#endif
per_cpu->interrupt_stack_low = low;
per_cpu->interrupt_stack_high = high;
/*
* Interrupt stack might have to be aligned and/or setup in a specific
* way. Do not use the local low or high variables here since
* _CPU_Interrupt_stack_setup() is a nasty macro that might want to play
* with the real memory locations.
*/
#if defined(_CPU_Interrupt_stack_setup)
_CPU_Interrupt_stack_setup(
per_cpu->interrupt_stack_low,
per_cpu->interrupt_stack_high
);
#endif
}
}
#endif
#if ( CPU_HAS_HARDWARE_INTERRUPT_STACK == TRUE )
_CPU_Install_interrupt_stack();
#endif
}
Thread _MPCI_Receive_server(
uint32_t ignored
)
{
MP_packet_Prefix *the_packet;
MPCI_Packet_processor the_function;
Thread_Control *executing;
executing = _Thread_Get_executing();
for ( ; ; ) {
executing->receive_packet = NULL;
_Thread_Disable_dispatch();
_CORE_semaphore_Seize(
&_MPCI_Semaphore,
executing,
0,
true,
WATCHDOG_NO_TIMEOUT
);
_Thread_Enable_dispatch();
for ( ; ; ) {
the_packet = _MPCI_Receive_packet();
if ( !the_packet )
break;
executing->receive_packet = the_packet;
if ( !_Mp_packet_Is_valid_packet_class ( the_packet->the_class ) )
break;
the_function = _MPCI_Packet_processors[ the_packet->the_class ];
if ( !the_function )
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_BAD_PACKET
);
(*the_function)( the_packet );
}
}
return 0; /* unreached - only to remove warnings */
}
void _MPCI_Handler_initialization(
uint32_t timeout_status
)
{
CORE_semaphore_Attributes attributes;
MPCI_Control *users_mpci_table;
users_mpci_table = _Configuration_MP_table->User_mpci_table;
if ( _System_state_Is_multiprocessing && !users_mpci_table )
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_NO_MPCI
);
_MPCI_table = users_mpci_table;
if ( !_System_state_Is_multiprocessing )
return;
/*
* Register the MP Process Packet routine.
*/
_MPCI_Register_packet_processor(
MP_PACKET_MPCI_INTERNAL,
_MPCI_Internal_packets_Process_packet
);
/*
* Create the counting semaphore used by the MPCI Receive Server.
*/
attributes.discipline = CORE_SEMAPHORE_DISCIPLINES_FIFO;
_CORE_semaphore_Initialize(
&_MPCI_Semaphore,
&attributes, /* the_semaphore_attributes */
0 /* initial_value */
);
_Thread_queue_Initialize(
&_MPCI_Remote_blocked_threads,
THREAD_QUEUE_DISCIPLINE_FIFO,
STATES_WAITING_FOR_RPC_REPLY,
timeout_status
);
}
void _Objects_MP_Handler_early_initialization(void)
{
uint32_t node;
uint32_t maximum_nodes;
node = _Configuration_MP_table->node;
maximum_nodes = _Configuration_MP_table->maximum_nodes;
if ( node < 1 || node > maximum_nodes )
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_INVALID_NODE
);
_Objects_Local_node = node;
_Objects_Maximum_nodes = maximum_nodes;
}
void _Thread_Handler_initialization(void)
{
rtems_stack_allocate_init_hook stack_allocate_init_hook =
rtems_configuration_get_stack_allocate_init_hook();
#if defined(RTEMS_MULTIPROCESSING)
uint32_t maximum_proxies =
_Configuration_MP_table->maximum_proxies;
#endif
if ( rtems_configuration_get_stack_allocate_hook() == NULL ||
rtems_configuration_get_stack_free_hook() == NULL)
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_BAD_STACK_HOOK
);
if ( stack_allocate_init_hook != NULL )
(*stack_allocate_init_hook)( rtems_configuration_get_stack_space_size() );
#if defined(RTEMS_MULTIPROCESSING)
_Thread_MP_Handler_initialization( maximum_proxies );
#endif
/*
* Initialize the internal class of threads. We need an IDLE thread
* per CPU in an SMP system. In addition, if this is a loosely
* coupled multiprocessing system, account for the MPCI Server Thread.
*/
_Thread_Initialize_information(
&_Thread_Internal_information,
OBJECTS_INTERNAL_API,
OBJECTS_INTERNAL_THREADS,
_Thread_Get_maximum_internal_threads(),
false, /* true if names for this object are strings */
8 /* maximum length of each object's name */
#if defined(RTEMS_MULTIPROCESSING)
,
false /* true if this is a global object class */
#endif
);
}
MP_packet_Prefix *_MPCI_Get_packet ( void )
{
MP_packet_Prefix *the_packet;
(*_MPCI_table->get_packet)( &the_packet );
if ( the_packet == NULL )
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_OUT_OF_PACKETS
);
/*
* Put in a default timeout that will be used for
* all packets that do not otherwise have a timeout.
*/
the_packet->timeout = MPCI_DEFAULT_TIMEOUT;
return the_packet;
}
uint32_t _CPU_ISR_Get_level( void )
{
uint32_t status = _Nios2_Get_ctlreg_status();
uint32_t level = 0;
switch ( _Nios2_ISR_Get_status_mask() ) {
case NIOS2_ISR_STATUS_MASK_IIC:
level = (status & NIOS2_STATUS_PIE) == 0;
break;
case NIOS2_ISR_STATUS_MASK_EIC_IL:
level = (status & NIOS2_STATUS_IL_MASK) >> NIOS2_STATUS_IL_OFFSET;
break;
case NIOS2_ISR_STATUS_MASK_EIC_RSIE:
level = (status & NIOS2_STATUS_RSIE) == 0;
break;
default:
/* FIXME */
_Terminate( INTERNAL_ERROR_CORE, false, 0xdeadbeef );
break;
}
return level;
}
void _Thread_blocking_operation_Cancel(
#if defined(RTEMS_DEBUG)
Thread_blocking_operation_States sync_state,
#else
Thread_blocking_operation_States sync_state __attribute__((unused)),
#endif
Thread_Control *the_thread,
ISR_Level level
)
{
/*
* Cases that should not happen and why.
*
* THREAD_BLOCKING_OPERATION_SYNCHRONIZED:
*
* This indicates that someone did not enter a blocking
* operation critical section.
*
* THREAD_BLOCKING_OPERATION_NOTHING_HAPPENED:
*
* This indicates that there was nothing to cancel so
* we should not have been called.
*/
#if defined(RTEMS_DEBUG)
if ( (sync_state == THREAD_BLOCKING_OPERATION_SYNCHRONIZED) ||
(sync_state == THREAD_BLOCKING_OPERATION_NOTHING_HAPPENED) ) {
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_IMPLEMENTATION_BLOCKING_OPERATION_CANCEL
);
}
#endif
_Thread_blocking_operation_Finalize( the_thread, level );
}
void RTEMS_Malloc_Initialize(
const Heap_Area *areas,
size_t area_count,
Heap_Initialization_or_extend_handler extend
)
{
Heap_Control *heap = RTEMS_Malloc_Heap;
if ( !rtems_configuration_get_unified_work_area() ) {
Heap_Initialization_or_extend_handler init_or_extend = _Heap_Initialize;
uintptr_t page_size = CPU_HEAP_ALIGNMENT;
size_t i;
for (i = 0; i < area_count; ++i) {
const Heap_Area *area = &areas [i];
uintptr_t space_available = (*init_or_extend)(
heap,
area->begin,
area->size,
page_size
);
if ( space_available > 0 ) {
init_or_extend = extend;
}
}
if ( init_or_extend == _Heap_Initialize ) {
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_NO_MEMORY_FOR_HEAP
);
}
}
}
void _POSIX_Fatal_error( POSIX_Fatal_domain domain, int eno )
{
uint32_t code = ( domain << 8 ) | ( ( uint32_t ) eno & 0xffU );
_Terminate( INTERNAL_ERROR_POSIX_API, false, code );
}
void _Thread_queue_Enqueue_critical(
Thread_queue_Queue *queue,
const Thread_queue_Operations *operations,
Thread_Control *the_thread,
States_Control state,
Thread_queue_Context *queue_context
)
{
Thread_queue_Path path;
Per_CPU_Control *cpu_self;
bool success;
#if defined(RTEMS_MULTIPROCESSING)
if ( _Thread_MP_Is_receive( the_thread ) && the_thread->receive_packet ) {
the_thread = _Thread_MP_Allocate_proxy( state );
}
#endif
_Thread_Wait_claim( the_thread, queue, operations );
if ( !_Thread_queue_Path_acquire( the_thread, queue, &path ) ) {
_Thread_Wait_restore_default( the_thread );
_Thread_queue_Queue_release( queue, &queue_context->Lock_context );
_Thread_Wait_tranquilize( the_thread );
( *queue_context->deadlock_callout )( the_thread );
return;
}
( *operations->enqueue )( queue, the_thread, &path );
_Thread_queue_Path_release( &path );
the_thread->Wait.return_code = STATUS_SUCCESSFUL;
_Thread_Wait_flags_set( the_thread, THREAD_QUEUE_INTEND_TO_BLOCK );
cpu_self = _Thread_Dispatch_disable_critical( &queue_context->Lock_context );
_Thread_queue_Queue_release( queue, &queue_context->Lock_context );
if (
cpu_self->thread_dispatch_disable_level
!= queue_context->expected_thread_dispatch_disable_level
) {
_Terminate(
INTERNAL_ERROR_CORE,
false,
INTERNAL_ERROR_THREAD_QUEUE_ENQUEUE_FROM_BAD_STATE
);
}
/*
* Set the blocking state for this thread queue in the thread.
*/
_Thread_Set_state( the_thread, state );
/*
* If the thread wants to timeout, then schedule its timer.
*/
switch ( queue_context->timeout_discipline ) {
case WATCHDOG_RELATIVE:
/* A relative timeout of 0 is a special case indefinite (no) timeout */
if ( queue_context->timeout != 0 ) {
_Thread_Timer_insert_relative(
the_thread,
cpu_self,
_Thread_Timeout,
(Watchdog_Interval) queue_context->timeout
);
}
break;
case WATCHDOG_ABSOLUTE:
_Thread_Timer_insert_absolute(
the_thread,
cpu_self,
_Thread_Timeout,
queue_context->timeout
);
break;
default:
break;
}
/*
* At this point thread dispatching is disabled, however, we already released
* the thread queue lock. Thus, interrupts or threads on other processors
* may already changed our state with respect to the thread queue object.
* The request could be satisfied or timed out. This situation is indicated
* by the thread wait flags. Other parties must not modify our thread state
* as long as we are in the THREAD_QUEUE_INTEND_TO_BLOCK thread wait state,
* thus we have to cancel the blocking operation ourself if necessary.
*/
success = _Thread_Wait_flags_try_change_acquire(
the_thread,
THREAD_QUEUE_INTEND_TO_BLOCK,
THREAD_QUEUE_BLOCKED
);
if ( !success ) {
_Thread_Remove_timer_and_unblock( the_thread, queue );
}
_Thread_Update_priority( path.update_priority );
_Thread_Dispatch_enable( cpu_self );
}
void _Thread_Handler( void )
{
Thread_Control *executing = _Thread_Executing;
ISR_Level level;
/*
* Some CPUs need to tinker with the call frame or registers when the
* thread actually begins to execute for the first time. This is a
* hook point where the port gets a shot at doing whatever it requires.
*/
_Context_Initialization_at_thread_begin();
#if !defined(RTEMS_SMP)
/*
* have to put level into a register for those cpu's that use
* inline asm here
*/
level = executing->Start.isr_level;
_ISR_Set_level( level );
#endif
/*
* Initialize the floating point context because we do not come
* through _Thread_Dispatch on our first invocation. So the normal
* code path for performing the FP context switch is not hit.
*/
#if ( CPU_HARDWARE_FP == TRUE ) || ( CPU_SOFTWARE_FP == TRUE )
#if ( CPU_USE_DEFERRED_FP_SWITCH == TRUE )
if ( (executing->fp_context != NULL) &&
!_Thread_Is_allocated_fp( executing ) ) {
if ( _Thread_Allocated_fp != NULL )
_Context_Save_fp( &_Thread_Allocated_fp->fp_context );
_Thread_Allocated_fp = executing;
}
#endif
#endif
/*
* Take care that 'begin' extensions get to complete before
* 'switch' extensions can run. This means must keep dispatch
* disabled until all 'begin' extensions complete.
*/
_User_extensions_Thread_begin( executing );
/*
* At this point, the dispatch disable level BETTER be 1.
*/
#if defined(RTEMS_SMP)
{
/*
* On SMP we enter _Thread_Handler() with interrupts disabled and
* _Thread_Dispatch() obtained the per-CPU lock for us. We have to
* release it here and set the desired interrupt level of the thread.
*/
Per_CPU_Control *cpu_self = _Per_CPU_Get();
_Assert( cpu_self->thread_dispatch_disable_level == 1 );
_Assert( _ISR_Get_level() != 0 );
_Thread_Debug_set_real_processor( executing, cpu_self );
cpu_self->thread_dispatch_disable_level = 0;
_Profiling_Thread_dispatch_enable( cpu_self, 0 );
level = executing->Start.isr_level;
_ISR_Set_level( level);
/*
* The thread dispatch level changed from one to zero. Make sure we lose
* no thread dispatch necessary update.
*/
_Thread_Dispatch();
}
#else
_Thread_Enable_dispatch();
#endif
/*
* RTEMS supports multiple APIs and each API can define a different
* thread/task prototype. The following code supports invoking the
* user thread entry point using the prototype expected.
*/
if ( executing->Start.prototype == THREAD_START_NUMERIC ) {
executing->Wait.return_argument =
(*(Thread_Entry_numeric) executing->Start.entry_point)(
executing->Start.numeric_argument
);
}
#if defined(RTEMS_POSIX_API)
else if ( executing->Start.prototype == THREAD_START_POINTER ) {
executing->Wait.return_argument =
(*(Thread_Entry_pointer) executing->Start.entry_point)(
executing->Start.pointer_argument
);
}
#endif
#if defined(FUNCTIONALITY_NOT_CURRENTLY_USED_BY_ANY_API)
else if ( executing->Start.prototype == THREAD_START_BOTH_POINTER_FIRST ) {
executing->Wait.return_argument =
(*(Thread_Entry_both_pointer_first) executing->Start.entry_point)(
executing->Start.pointer_argument,
executing->Start.numeric_argument
);
}
else if ( executing->Start.prototype == THREAD_START_BOTH_NUMERIC_FIRST ) {
executing->Wait.return_argument =
(*(Thread_Entry_both_numeric_first) executing->Start.entry_point)(
executing->Start.numeric_argument,
executing->Start.pointer_argument
);
}
#endif
/*
* In the switch above, the return code from the user thread body
* was placed in return_argument. This assumed that if it returned
* anything (which is not supporting in all APIs), then it would be
* able to fit in a (void *).
*/
_User_extensions_Thread_exitted( executing );
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_THREAD_EXITTED
);
}
void boot_card(const char *cmdline)
{
test_context *ctx = &test_instance;
struct timecounter *tc_soft = &ctx->tc_soft;
uint64_t soft_freq = 1000000;
struct bintime bt;
struct timeval tv;
struct timespec ts;
rtems_test_begink();
assert(time(NULL) == TOD_SECONDS_1970_THROUGH_1988);
rtems_bsd_bintime(&bt);
assert(bt.sec == TOD_SECONDS_1970_THROUGH_1988);
assert(bt.frac == 0);
rtems_bsd_getbintime(&bt);
assert(bt.sec == TOD_SECONDS_1970_THROUGH_1988);
assert(bt.frac == 0);
rtems_bsd_microtime(&tv);
assert(tv.tv_sec == TOD_SECONDS_1970_THROUGH_1988);
assert(tv.tv_usec == 0);
rtems_bsd_getmicrotime(&tv);
assert(tv.tv_sec == TOD_SECONDS_1970_THROUGH_1988);
assert(tv.tv_usec == 0);
rtems_bsd_nanotime(&ts);
assert(ts.tv_sec == TOD_SECONDS_1970_THROUGH_1988);
assert(ts.tv_nsec == 0);
rtems_bsd_getnanotime(&ts);
assert(ts.tv_sec == TOD_SECONDS_1970_THROUGH_1988);
assert(ts.tv_nsec == 0);
assert(rtems_clock_get_uptime_seconds() == 0);
assert(rtems_clock_get_uptime_nanoseconds() == 0);
rtems_bsd_binuptime(&bt);
assert(bt.sec == 1);
assert(bt.frac == 0);
rtems_bsd_getbinuptime(&bt);
assert(bt.sec == 1);
assert(bt.frac == 0);
rtems_bsd_microuptime(&tv);
assert(tv.tv_sec == 1);
assert(tv.tv_usec == 0);
rtems_bsd_getmicrouptime(&tv);
assert(tv.tv_sec == 1);
assert(tv.tv_usec == 0);
rtems_bsd_nanouptime(&ts);
assert(ts.tv_sec == 1);
assert(ts.tv_nsec == 0);
rtems_bsd_getnanouptime(&ts);
assert(ts.tv_sec == 1);
assert(ts.tv_nsec == 0);
/* On RTEMS time does not advance using the dummy timecounter */
rtems_bsd_binuptime(&bt);
assert(bt.sec == 1);
assert(bt.frac == 0);
rtems_timecounter_tick();
rtems_bsd_binuptime(&bt);
assert(bt.sec == 1);
assert(bt.frac == 0);
ctx->tc_soft_counter = 0;
tc_soft->tc_get_timecount = test_get_timecount_soft;
tc_soft->tc_counter_mask = 0x0fffffff;
tc_soft->tc_frequency = soft_freq;
tc_soft->tc_quality = 1234;
tc_soft->tc_priv = ctx;
_Timecounter_Install(tc_soft);
assert(ctx->tc_soft_counter == 3);
rtems_bsd_binuptime(&bt);
assert(ctx->tc_soft_counter == 4);
assert(bt.sec == 1);
assert(bt.frac == 18446744073708);
ctx->tc_soft_counter = 0xf0000000 | 3;
rtems_bsd_binuptime(&bt);
assert(ctx->tc_soft_counter == (0xf0000000 | 4));
assert(bt.sec == 1);
assert(bt.frac == 18446744073708);
/* Ensure that the fraction overflows and the second remains constant */
ctx->tc_soft_counter = (0xf0000000 | 3) + soft_freq;
rtems_bsd_binuptime(&bt);
assert(ctx->tc_soft_counter == (0xf0000000 | 4) + soft_freq);
assert(bt.sec == 1);
assert(bt.frac == 18446742522092);
rtems_test_endk();
_Terminate(RTEMS_FATAL_SOURCE_EXIT, false, 0);
}
Thread_Control *_Thread_MP_Allocate_proxy (
States_Control the_state
)
{
Thread_Proxy_control *the_proxy;
ISR_lock_Context lock_context;
_Thread_MP_Proxies_acquire( &lock_context );
the_proxy = (Thread_Proxy_control *)
_Chain_Get_unprotected( &_Thread_MP_Inactive_proxies );
if ( the_proxy != NULL ) {
Thread_Control *executing;
MP_packet_Prefix *receive_packet;
Objects_Id source_tid;
executing = _Thread_Executing;
receive_packet = _MPCI_Receive_server_tcb->receive_packet;
source_tid = receive_packet->source_tid;
executing->Wait.return_code = THREAD_STATUS_PROXY_BLOCKING;
the_proxy->receive_packet = receive_packet;
the_proxy->Object.id = source_tid;
the_proxy->current_priority = receive_packet->source_priority;
the_proxy->current_state = _States_Set( STATES_DORMANT, the_state );
the_proxy->Wait.count = executing->Wait.count;
the_proxy->Wait.return_argument = executing->Wait.return_argument;
the_proxy->Wait.return_argument_second = executing->Wait.return_argument_second;
the_proxy->Wait.option = executing->Wait.option;
the_proxy->Wait.return_code = executing->Wait.return_code;
the_proxy->Wait.timeout_code = executing->Wait.timeout_code;
the_proxy->thread_queue_callout = _Thread_queue_MP_callout_do_nothing;
_RBTree_Insert_inline(
&_Thread_MP_Active_proxies,
&the_proxy->Active,
&source_tid,
_Thread_MP_Proxy_less
);
_Thread_MP_Proxies_release( &lock_context );
return (Thread_Control *) the_proxy;
}
_Thread_MP_Proxies_release( &lock_context );
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_OUT_OF_PROXIES
);
/*
* NOTE: The following return ensures that the compiler will
* think that all paths return a value.
*/
return NULL;
}
void _RTEMS_tasks_Initialize_user_tasks_body( void )
{
uint32_t index;
uint32_t maximum;
rtems_id id;
rtems_status_code return_value;
rtems_initialization_tasks_table *user_tasks;
bool register_global_construction;
rtems_task_entry entry_point;
/*
* Move information into local variables
*/
user_tasks = Configuration_RTEMS_API.User_initialization_tasks_table;
maximum = Configuration_RTEMS_API.number_of_initialization_tasks;
/*
* Verify that we have a set of user tasks to iterate
*/
if ( !user_tasks )
return;
register_global_construction = true;
/*
* Now iterate over the initialization tasks and create/start them.
*/
for ( index=0 ; index < maximum ; index++ ) {
return_value = rtems_task_create(
user_tasks[ index ].name,
user_tasks[ index ].initial_priority,
user_tasks[ index ].stack_size,
user_tasks[ index ].mode_set,
user_tasks[ index ].attribute_set,
&id
);
if ( !rtems_is_status_successful( return_value ) )
_Terminate( INTERNAL_ERROR_RTEMS_API, true, return_value );
entry_point = user_tasks[ index ].entry_point;
if ( entry_point == NULL ) {
_Terminate(
INTERNAL_ERROR_CORE,
false,
INTERNAL_ERROR_RTEMS_INIT_TASK_ENTRY_IS_NULL
);
}
if ( register_global_construction ) {
register_global_construction = false;
entry_point = _RTEMS_Global_construction;
}
return_value = rtems_task_start(
id,
entry_point,
user_tasks[ index ].argument
);
_Assert( rtems_is_status_successful( return_value ) );
(void) return_value;
}
}
void _Thread_Handler( void )
{
Thread_Control *executing = _Thread_Executing;
ISR_Level level;
Per_CPU_Control *cpu_self;
/*
* Some CPUs need to tinker with the call frame or registers when the
* thread actually begins to execute for the first time. This is a
* hook point where the port gets a shot at doing whatever it requires.
*/
_Context_Initialization_at_thread_begin();
/* On SMP we enter _Thread_Handler() with interrupts disabled */
_SMP_Assert( _ISR_Get_level() != 0 );
/*
* have to put level into a register for those cpu's that use
* inline asm here
*/
level = executing->Start.isr_level;
_ISR_Set_level( level );
/*
* Initialize the floating point context because we do not come
* through _Thread_Dispatch on our first invocation. So the normal
* code path for performing the FP context switch is not hit.
*/
_Thread_Restore_fp( executing );
/*
* Do not use the level of the thread control block, since it has a
* different format.
*/
_ISR_Local_disable( level );
/*
* At this point, the dispatch disable level BETTER be 1.
*/
cpu_self = _Per_CPU_Get();
_Assert( cpu_self->thread_dispatch_disable_level == 1 );
/*
* Make sure we lose no thread dispatch necessary update and execute the
* post-switch actions. As a side-effect change the thread dispatch level
* from one to zero. Do not use _Thread_Enable_dispatch() since there is no
* valid thread dispatch necessary indicator in this context.
*/
_Thread_Do_dispatch( cpu_self, level );
/*
* Invoke the thread begin extensions in the context of the thread entry
* function with thread dispatching enabled. This enables use of dynamic
* memory allocation, creation of POSIX keys and use of C++ thread local
* storage. Blocking synchronization primitives are allowed also.
*/
_User_extensions_Thread_begin( executing );
/*
* RTEMS supports multiple APIs and each API can define a different
* thread/task prototype. The following code supports invoking the
* user thread entry point using the prototype expected.
*/
( *executing->Start.Entry.adaptor )( executing );
/*
* In the call above, the return code from the user thread body which return
* something was placed in return_argument. This assumed that if it
* returned anything (which is not supporting in all APIs), then it would be
* able to fit in a (void *).
*/
_User_extensions_Thread_exitted( executing );
_Terminate(
INTERNAL_ERROR_CORE,
true,
INTERNAL_ERROR_THREAD_EXITTED
);
}
