bsps: Rework work area initialization

The work area initialization was done by the BSP through
bsp_work_area_initialize(). This approach predated the system
initialization through the system initialization linker set. The
workspace and C program heap were unconditionally initialized.  The aim
is to support RTEMS application configurations which do not need the
workspace and C program heap.  In these configurations, the workspace
and C prgram heap should not get initialized.

Change all bsp_work_area_initialize() to implement _Memory_Get()
instead.  Move the dirty memory, sbrk(), per-CPU data, workspace, and
malloc() heap initialization into separate system initialization steps.
This makes it also easier to test the individual initialization steps.

This change adds a dependency to _Heap_Extend() to all BSPs.  This
dependency will be removed in a follow up change.

Update #3838.

1 files changed, 42 insertions, 0 deletions

diff --git a/cpukit/score/src/smp.c b/cpukit/score/src/smp.c
index 306d1ea4b4..b6394a8bdf 100644
--- a/cpukit/score/src/smp.c
+++ b/cpukit/score/src/smp.c
@@ -20,9 +20,14 @@
 
 #include <rtems/score/smpimpl.h>
 #include <rtems/score/assert.h>
+#include <rtems/score/memory.h>
+#include <rtems/score/percpudata.h>
 #include <rtems/score/schedulerimpl.h>
 #include <rtems/score/threadimpl.h>
 #include <rtems/config.h>
+#include <rtems/sysinit.h>
+
+#include <string.h>
 
 #if CPU_USE_DEFERRED_FP_SWITCH == TRUE
   #error "deferred FP switch not implemented for SMP"
@@ -252,3 +257,40 @@ void _SMP_Send_message_multicast(
     }
   }
 }
+
+static void _Per_CPU_Data_initialize( void )
+{
+  uintptr_t size;
+
+  size = RTEMS_LINKER_SET_SIZE( _Per_CPU_Data );
+
+  if ( size > 0 ) {
+    const Memory_Information *mem;
+    Per_CPU_Control          *cpu;
+    uint32_t                  cpu_index;
+    uint32_t                  cpu_max;
+
+    mem = _Memory_Get();
+    cpu = _Per_CPU_Get_by_index( 0 );
+    cpu->data = RTEMS_LINKER_SET_BEGIN( _Per_CPU_Data );
+
+    cpu_max = rtems_configuration_get_maximum_processors();
+
+    for ( cpu_index = 1 ; cpu_index < cpu_max ; ++cpu_index ) {
+      cpu = _Per_CPU_Get_by_index( cpu_index );
+      cpu->data = _Memory_Allocate( mem, size, CPU_CACHE_LINE_BYTES );
+
+      if( cpu->data == NULL ) {
+        _Internal_error( INTERNAL_ERROR_NO_MEMORY_FOR_PER_CPU_DATA );
+      }
+
+      memcpy( cpu->data, RTEMS_LINKER_SET_BEGIN( _Per_CPU_Data ), size);
+    }
+  }
+}
+
+RTEMS_SYSINIT_ITEM(
+  _Per_CPU_Data_initialize,
+  RTEMS_SYSINIT_PER_CPU_DATA,
+  RTEMS_SYSINIT_ORDER_MIDDLE
+);