Add framework to support task groups

git-svn-id: svn://svn.code.sf.net/p/nuttx/code/trunk@5562 42af7a65-404d-4744-a932-0658087f49c3

20 files changed, 677 insertions, 141 deletions

diff --git a/nuttx/include/nuttx/sched.h b/nuttx/include/nuttx/sched.h
index 1e75b5020..4a3bae5e4 100644
--- a/nuttx/include/nuttx/sched.h
+++ b/nuttx/include/nuttx/sched.h
@@ -57,6 +57,14 @@
 /********************************************************************************
  * Pre-processor Definitions
  ********************************************************************************/
+/* Configuration ****************************************************************/
+/* Task groups currently only supported for retention of child status */
+
+#if defined(CONFIG_SCHED_HAVE_PARENT) && defined(CONFIG_SCHED_CHILD_STATUS)
+#  define HAVE_TASK_GROUP 1
+#else
+#  undef HAVE_TASK_GROUP
+#endif
 
 /* Task Management Definitins ***************************************************/
 
@@ -64,7 +72,7 @@
 
 #define MAX_LOCK_COUNT             127
 
-/* Values for the _TCB flags flag bits */
+/* Values for the _TCB flags bits */
 
 #define TCB_FLAG_TTYPE_SHIFT       (0)      /* Bits 0-1: thread type */
 #define TCB_FLAG_TTYPE_MASK        (3 << TCB_FLAG_TTYPE_SHIFT)
@@ -74,7 +82,10 @@
 #define TCB_FLAG_NONCANCELABLE     (1 << 2) /* Bit 2: Pthread is non-cancelable */
 #define TCB_FLAG_CANCEL_PENDING    (1 << 3) /* Bit 3: Pthread cancel is pending */
 #define TCB_FLAG_ROUND_ROBIN       (1 << 4) /* Bit 4: Round robin sched enabled */
-#define TCB_FLAG_NOCLDWAIT         (1 << 5) /* Bit 5: Do not retain child exit status */
+
+/* Values for struct task_group tg_flags */
+
+#define GROUP_FLAG_NOCLDWAIT       (1 << 0) /* Bit 0: Do not retain child exit status */
 
 /* Values for struct child_status_s ch_flags */
 
@@ -159,6 +170,7 @@ typedef CODE void (*onexitfunc_t)(int exitcode, FAR void *arg);
 
 typedef struct msgq_s msgq_t;
 
+/* struct environ_s **************************************************************/
 /* The structure used to maintain environment variables */
 
 #ifndef CONFIG_DISABLE_ENVIRON
@@ -173,6 +185,7 @@ typedef struct environ_s environ_t;
 # define SIZEOF_ENVIRON_T(alloc) (sizeof(environ_t) + alloc - 1)
 #endif
 
+/* struct child_status_s *********************************************************/
 /* This structure is used to maintin information about child tasks.
  * pthreads work differently, they have join information.  This is 
  * only for child tasks.
@@ -189,6 +202,7 @@ struct child_status_s
 };
 #endif
 
+/* struct dspace_s ***************************************************************/
 /* This structure describes a reference counted D-Space region.  This must be a
  * separately allocated "break-away" structure that can be owned by a task and
  * any pthreads created by the task.
@@ -214,6 +228,58 @@ struct dspace_s
 };
 #endif
 
+/* struct task_group_s ***********************************************************/
+/* All threads created by pthread_create belong in the same task group (along with
+ * the thread of the original task).  struct task_group_s is a shared, "breakaway"
+ * structure referenced by each TCB.
+ *
+ * This structure should contain *all* resources shared by tasks and threads that
+ * belong to the same task group:
+ *
+ *   Child exit status
+ *   Environment varibles
+ *   PIC data space and address environments
+ *   File descriptors
+ *   FILE streams
+ *   Sockets
+ *
+ * Currenty, however, this implementation only applies to child exit status.
+ *
+ * Each instance of struct task_group_s is reference counted. Each instance is
+ * created with a reference count of one.  The reference incremeneted when each
+ * thread joins the group and decremented when each thread exits, leaving the
+ * group.  When the refernce count decrements to zero, the struc task_group_s
+ * is free.
+ */
+
+#ifdef HAVE_TASK_GROUP
+struct task_group_s
+{
+  uint16_t tg_crefs;                      /* Count of threads sharing this data */
+  uint8_t  tg_flags;                      /* See GROUP_FLAG_* definitions       */
+
+  /* Child exit status **********************************************************/
+
+  FAR struct child_status_s *tg_children; /* Head of a list of child status     */
+
+  /* Environment varibles *******************************************************/
+  /* Not yet (see type environ_t) */
+
+  /* PIC data space and address environments */
+  /* Not yet (see struct dspace_s) */
+
+  /* File descriptors */
+  /* Not yet (see struct filelist) */
+
+  /* FILE streams */
+  /* Not yet (see streamlist) */
+
+  /* Sockets */
+  /* Not yet (see struct socketlist) */
+};
+#endif
+
+/* _TCB **************************************************************************/
 /* This is the task control block (TCB).  Each task or thread is represented by
  * a TCB.  The TCB is the heart of the NuttX task-control logic.
  */
@@ -225,14 +291,18 @@ struct _TCB
   FAR struct _TCB *flink;                /* Doubly linked list                  */
   FAR struct _TCB *blink;
 
+  /* Task Group *****************************************************************/
+
+#ifdef HAVE_TASK_GROUP
+  FAR struct task_group_s *group;        /* Pointer to shared task group data   */
+#endif
+
   /* Task Management Fields *****************************************************/
 
   pid_t    pid;                          /* This is the ID of the thread        */
 #ifdef CONFIG_SCHED_HAVE_PARENT          /* Support parent-child relationship   */
   pid_t    parent;                       /* This is the ID of the parent thread */
-#ifdef CONFIG_SCHED_CHILD_STATUS         /* Retain child thread status          */
-  FAR struct child_status_s *children;   /* Head of a list of child status      */
-#else
+#ifndef CONFIG_SCHED_CHILD_STATUS        /* Retain child thread status          */
   uint16_t nchildren;                    /* This is the number active children  */
 #endif
 #endif
diff --git a/nuttx/net/net_poll.c b/nuttx/net/net_poll.c
index 3021ac35e..1838f541e 100644
--- a/nuttx/net/net_poll.c
+++ b/nuttx/net/net_poll.c
@@ -118,11 +118,11 @@ static uint16_t poll_interrupt(FAR struct uip_driver_s *dev, FAR void *conn,
 
   nllvdbg("flags: %04x\n", flags);
 
-  DEBUGASSERT(info && info->psock && info->fds);
+  DEBUGASSERT(!info || (info->psock && info->fds));
 
   /* 'priv' might be null in some race conditions (?) */
 
-  if (info->fds)
+  if (info)
     {
       pollevent_t eventset = 0;
 
diff --git a/nuttx/sched/Makefile b/nuttx/sched/Makefile
index 1ad244450..6a710ff95 100644
--- a/nuttx/sched/Makefile
+++ b/nuttx/sched/Makefile
@@ -87,6 +87,8 @@ SCHED_SRCS += sched_waitid.c sched_wait.c
 endif
 endif
 
+GRP_SRCS = group_create.c group_join.c group_leave.c
+
 ENV_SRCS  = env_getenvironptr.c env_dup.c env_share.c env_release.c
 ENV_SRCS += env_findvar.c env_removevar.c
 ENV_SRCS += env_clearenv.c env_getenv.c env_putenv.c env_setenv.c env_unsetenv.c
@@ -169,8 +171,9 @@ IRQ_SRCS = irq_initialize.c irq_attach.c irq_dispatch.c irq_unexpectedisr.c
 KMM_SRCS = kmm_initialize.c kmm_addregion.c kmm_semaphore.c
 KMM_SRCS = kmm_kmalloc.c kmm_kzalloc.c kmm_krealloc.c kmm_kfree.c
 
-CSRCS = $(MISC_SRCS) $(TSK_SRCS) $(SCHED_SRCS) $(WDOG_SRCS) $(TIME_SRCS) \
-	    $(SEM_SRCS) $(TIMER_SRCS) $(WORK_SRCS) $(PGFILL_SRCS) $(IRQ_SRCS)
+CSRCS  = $(MISC_SRCS) $(TSK_SRCS) $(GRP_SRCS) $(SCHED_SRCS) $(WDOG_SRCS)
+CSRCS += $(TIME_SRCS) $(SEM_SRCS) $(TIMER_SRCS) $(WORK_SRCS) $(PGFILL_SRCS)
+CSRCS += $(IRQ_SRCS)
 
 ifneq ($(CONFIG_DISABLE_CLOCK),y)
 CSRCS += $(CLOCK_SRCS)
diff --git a/nuttx/sched/env_dup.c b/nuttx/sched/env_dup.c
index 033348411..30c0b7773 100644
--- a/nuttx/sched/env_dup.c
+++ b/nuttx/sched/env_dup.c
@@ -1,7 +1,7 @@
 /****************************************************************************
  * sched/env_dup.c
  *
- *   Copyright (C) 2007, 2009, 2011 Gregory Nutt. All rights reserved.
+ *   Copyright (C) 2007, 2009, 2011, 2013 Gregory Nutt. All rights reserved.
  *   Author: Gregory Nutt <gnutt@nuttx.org>
  *
  * Redistribution and use in source and binary forms, with or without
@@ -68,7 +68,7 @@
  *   exact duplicate of the parent task's environment.
  *
  * Parameters:
- *   ptcb The tcb to receive the newly allocated copy of the parspecifiedent
+ *   ptcb The tcb to receive the newly allocated copy of the parent
  *        TCB's environment structure with reference count equal to one
  *
  * Return Value:
@@ -81,48 +81,42 @@
 
 int env_dup(FAR _TCB *ptcb)
 {
+  FAR _TCB  *parent = (FAR _TCB*)g_readytorun.head;
+  environ_t *envp   = NULL;
   int ret = OK;
-  if (!ptcb )
-    {
-      ret = -EINVAL;
-    }
-  else
-    {
-      FAR _TCB  *parent = (FAR _TCB*)g_readytorun.head;
-      environ_t *envp   = NULL;
 
-      /* Pre-emption must be disabled throughout the following because the
-       * environment may be shared.
-       */
+  DEBUGASSERT(ptcb);
+
+  /* Pre-emption must be disabled throughout the following because the
+   * environment may be shared.
+   */
+
+  sched_lock();
 
-      sched_lock();
+  /* Does the parent task have an environment? */
 
-      /* Does the parent task have an environment? */
+  if (parent->envp)
+    {
+      /* Yes..The parent task has an environment, duplicate it */
 
-      if (parent->envp)
+      size_t envlen =  parent->envp->ev_alloc;
+      envp          = (environ_t*)kmalloc(SIZEOF_ENVIRON_T(envlen));
+      if (!envp)
         {
-          /* Yes..The parent task has an environment, duplicate it */
-
-          size_t envlen =  parent->envp->ev_alloc;
-          envp          = (environ_t*)kmalloc(SIZEOF_ENVIRON_T( envlen ));
-          if (!envp)
-            {
-              ret = -ENOMEM;
-            }
-          else
-            {
-              envp->ev_crefs = 1;
-              envp->ev_alloc = envlen;
-              memcpy( envp->ev_env, parent->envp->ev_env, envlen );
-            }
+          ret = -ENOMEM;
         }
+      else
+        {
+          envp->ev_crefs = 1;
+          envp->ev_alloc = envlen;
+          memcpy(envp->ev_env, parent->envp->ev_env, envlen);
+        }
+    }
 
-      /* Save the cloned environment in the new TCB */
-
-      ptcb->envp = envp;
-      sched_unlock();
-  }
+  /* Save the cloned environment in the new TCB */
 
+  ptcb->envp = envp;
+  sched_unlock();
   return ret;
 }
 
diff --git a/nuttx/sched/group_create.c b/nuttx/sched/group_create.c
new file mode 100644
index 000000000..d036b9084
--- /dev/null
+++ b/nuttx/sched/group_create.c
@@ -0,0 +1,111 @@
+/*****************************************************************************
+ * sched/group_create.c
+ *
+ *   Copyright (C) 2013 Gregory Nutt. All rights reserved.
+ *   Author: Gregory Nutt <gnutt@nuttx.org>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ * 3. Neither the name NuttX nor the names of its contributors may be
+ *    used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+ * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+ * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
+ * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
+ * OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
+ * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
+ * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ *
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Included Files
+ *****************************************************************************/
+
+#include <nuttx/config.h>
+
+#include <sched.h>
+#include <assert.h>
+#include <errno.h>
+#include <debug.h>
+
+#include <nuttx/kmalloc.h>
+
+#ifdef HAVE_TASK_GROUP
+
+/*****************************************************************************
+ * Pre-processor Definitions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Types
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Data
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Functions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Public Functions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Name: group_create
+ *
+ * Description:
+ *   Create and initialize a new task group structure for the specified TCB.
+ *   This function is called as part of the task creation sequence.
+ *
+ * Parameters:
+ *   tcb - The tcb in need of the task group.
+ *
+ * Return Value:
+ *   0 (OK) on success; a negated errno value on failure.
+ *
+ * Assumptions:
+ *   Called during task creation in a safe context.  No special precautions
+ *   are required here.
+ *
+ *****************************************************************************/
+
+int group_create(FAR _TCB *tcb)
+{
+  FAR struct task_group_s *group;
+
+  DEBUGASSERT(tcb && !tcb->group);
+
+  /* Allocate the group structure */
+
+  group = (FAR struct task_group_s *)kzalloc(sizeof(struct task_group_s));
+  if (!group)
+    {
+      return -ENOMEM;
+    }
+
+  /* Initialize the group structure and assign it to the tcb */
+
+  group->tg_crefs = 1;
+  tcb->group      = group;
+  return OK;
+}
+
+#endif /* HAVE_TASK_GROUP */
diff --git a/nuttx/sched/group_join.c b/nuttx/sched/group_join.c
new file mode 100644
index 000000000..987c1ba87
--- /dev/null
+++ b/nuttx/sched/group_join.c
@@ -0,0 +1,107 @@
+/*****************************************************************************
+ * sched/group_join.c
+ *
+ *   Copyright (C) 2013 Gregory Nutt. All rights reserved.
+ *   Author: Gregory Nutt <gnutt@nuttx.org>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ * 3. Neither the name NuttX nor the names of its contributors may be
+ *    used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+ * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+ * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
+ * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
+ * OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
+ * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
+ * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ *
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Included Files
+ *****************************************************************************/
+
+#include <nuttx/config.h>
+
+#include <sched.h>
+#include <assert.h>
+#include <debug.h>
+
+#include "os_internal.h"
+
+#ifdef HAVE_TASK_GROUP
+
+/*****************************************************************************
+ * Pre-processor Definitions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Types
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Data
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Functions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Public Functions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Name: group_join
+ *
+ * Description:
+ *   Copy the group structure reference from one TCB to another, incrementing
+ *   the refrence count on the group.  This function is called when a pthread
+ *   is produced within the task group so that the pthread can share the
+ *   resources of the task group.
+ *
+ * Parameters:
+ *   tcb - The TCB of the new "child" task that need to join the group.
+ *
+ * Return Value:
+ *   None
+ *
+ * Assumptions:
+ * - The parent task from which the group will be inherited is the task at
+ *   the thead of the ready to run list.
+ * - Called during thread creation in a safe context.  No special precautions
+ *   are required here.
+ *
+ *****************************************************************************/
+
+void group_join(FAR _TCB *tcb)
+{
+  FAR _TCB *ptcb = (FAR _TCB *)g_readytorun.head;
+
+  DEBUGASSERT(ptcb && tcb && ptcb->group && !tcb->group);
+
+  /* Copy the group reference from the parent to the child, incrementing the
+   * reference count.
+   */
+
+  tcb->group = ptcb->group;
+  ptcb->group->tg_crefs++;
+}
+
+#endif /* HAVE_TASK_GROUP */
diff --git a/nuttx/sched/group_leave.c b/nuttx/sched/group_leave.c
new file mode 100644
index 000000000..69a6ea8b2
--- /dev/null
+++ b/nuttx/sched/group_leave.c
@@ -0,0 +1,127 @@
+/*****************************************************************************
+ * sched/group_leave.c
+ *
+ *   Copyright (C) 2013 Gregory Nutt. All rights reserved.
+ *   Author: Gregory Nutt <gnutt@nuttx.org>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ * 3. Neither the name NuttX nor the names of its contributors may be
+ *    used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+ * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+ * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
+ * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
+ * OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
+ * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
+ * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ *
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Included Files
+ *****************************************************************************/
+
+#include <nuttx/config.h>
+
+#include <sched.h>
+#include <assert.h>
+#include <errno.h>
+#include <debug.h>
+
+#include "os_internal.h"
+
+#ifdef HAVE_TASK_GROUP
+
+/*****************************************************************************
+ * Pre-processor Definitions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Types
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Data
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Private Functions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Public Functions
+ *****************************************************************************/
+
+/*****************************************************************************
+ * Name: group_leave
+ *
+ * Description:
+ *   Release a reference on a group.  This function is called when a task or
+ *   thread exits.  It decrements the reference count on the group.  If the
+ *   reference count decrements to zero, then it frees the group and all of
+ *   resources contained in the group.
+ *
+ * Parameters:
+ *   tcb - The TCB of the task that is exiting.
+ *
+ * Return Value:
+ *   None.
+ *
+ * Assumptions:
+ *   Called during task deletion in a safe context.  No special precautions
+ *   are required here.
+ *
+ *****************************************************************************/
+
+void group_leave(FAR _TCB *tcb)
+{
+  FAR struct task_group_s *group;
+
+  DEBUGASSERT(tcb);
+
+  /* Make sure that we have a group */
+
+  group = tcb->group;
+  if (group)
+    {
+      /* Would the reference count decrement to zero? */
+
+      if (group->tg_crefs > 1)
+        {
+          /* No.. just decrement the reference count and return */
+
+          group->tg_crefs--;
+        }
+      else
+        {
+          /* Yes.. Release all of the resource contained within the group */
+          /* Free all un-reaped child exit status */
+
+          task_removechildren(tcb);
+
+          /* Release the group container itself */
+
+          sched_free(group);
+        }
+
+      tcb->group = NULL;
+    }
+}
+
+#endif /* HAVE_TASK_GROUP */
diff --git a/nuttx/sched/os_internal.h b/nuttx/sched/os_internal.h
index dc87cb9a4..ee5ada165 100644
--- a/nuttx/sched/os_internal.h
+++ b/nuttx/sched/os_internal.h
@@ -268,7 +268,20 @@ int  task_schedsetup(FAR _TCB *tcb, int priority, start_t start,
 int  task_argsetup(FAR _TCB *tcb, FAR const char *name, FAR const char *argv[]);
 void task_exithook(FAR _TCB *tcb, int status);
 int  task_deletecurrent(void);
+
 #ifdef CONFIG_SCHED_HAVE_PARENT
+int  task_reparent(pid_t ppid, pid_t chpid);
+
+#ifdef HAVE_TASK_GROUP
+int group_create(FAR _TCB *tcb);
+void group_join(FAR _TCB *tcb);
+void group_leave(FAR _TCB *tcb);
+#else
+# define group_create(tcb)
+# define group_join(tcb)
+# define group_leave(tcb)
+#endif
+
 #ifdef CONFIG_SCHED_CHILD_STATUS
 void weak_function task_initialize(void);
 FAR struct child_status_s *task_allocchild(void);
@@ -279,8 +292,8 @@ FAR struct child_status_s *task_findchild(FAR _TCB *tcb, pid_t pid);
 FAR struct child_status_s *task_removechild(FAR _TCB *tcb, pid_t pid);
 void task_removechildren(FAR _TCB *tcb);
 #endif
-int  task_reparent(pid_t ppid, pid_t chpid);
 #endif
+
 #ifndef CONFIG_CUSTOM_STACK
 int  kernel_thread(FAR const char *name, int priority, int stack_size,
                    main_t entry, FAR const char *argv[]);
diff --git a/nuttx/sched/os_start.c b/nuttx/sched/os_start.c
index a6d4e83b9..244aec967 100644
--- a/nuttx/sched/os_start.c
+++ b/nuttx/sched/os_start.c
@@ -202,6 +202,9 @@ const tasklist_t g_tasklisttable[NUM_TASK_STATES] =
  */
 
 static FAR _TCB g_idletcb;
+#if defined(CONFIG_SCHED_HAVE_PARENT) && defined(CONFIG_SCHED_CHILD_STATUS)
+static struct task_group_s g_idlegroup;
+#endif
 
 /* This is the name of the idle task */
 
@@ -280,13 +283,20 @@ void os_start(void)
   g_idletcb.argv[0] = (char*)g_idlename;
 #endif /* CONFIG_TASK_NAME_SIZE */
 
+  /* Join the IDLE group */
+
+#ifdef HAVE_TASK_GROUP
+  g_idlegroup.tg_crefs = 1;
+  g_idlegroup.tg_flags = GROUP_FLAG_NOCLDWAIT;
+  g_idletcb.group      = &g_idlegroup;
+#endif
+
   /* Then add the idle task's TCB to the head of the ready to run list */
 
   dq_addfirst((FAR dq_entry_t*)&g_idletcb, (FAR dq_queue_t*)&g_readytorun);
 
   /* Initialize the processor-specific portion of the TCB */
 
-  g_idletcb.flags = (TCB_FLAG_TTYPE_KERNEL | TCB_FLAG_NOCLDWAIT);
   up_initial_state(&g_idletcb);
 
   /* Initialize the semaphore facility(if in link).  This has to be done
diff --git a/nuttx/sched/pthread_create.c b/nuttx/sched/pthread_create.c
index f4d0d8fdf..e37c06892 100644
--- a/nuttx/sched/pthread_create.c
+++ b/nuttx/sched/pthread_create.c
@@ -251,6 +251,7 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
 #if CONFIG_RR_INTERVAL > 0
   int policy;
 #endif
+  int errcode;
   pid_t pid;
 
   /* If attributes were not supplied, use the default attributes */
@@ -268,6 +269,12 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
       return ENOMEM;
     }
 
+  /* Join the parent's task group */
+
+#ifdef HAVE_TASK_GROUP
+  group_join(ptcb);
+#endif
+
   /* Share the address environment of the parent task.  NOTE:  Only tasks
    * created throught the nuttx/binfmt loaders may have an address
    * environment.
@@ -277,8 +284,8 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
   ret = up_addrenv_share((FAR const _TCB *)g_readytorun.head, ptcb);
   if (ret < 0)
     {
-      sched_releasetcb(ptcb);
-      return -ret;
+      errcode = -ret;
+      goto errout_with_tcb;
     }
 #endif
 
@@ -287,8 +294,8 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
   ret = sched_setuppthreadfiles(ptcb);
   if (ret != OK)
     {
-      sched_releasetcb(ptcb);
-      return ret;
+      errcode = ret;
+      goto errout_with_tcb;
     }
 
   /* Share the parent's envionment */
@@ -300,8 +307,8 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
   pjoin = (FAR join_t*)kzalloc(sizeof(join_t));
   if (!pjoin)
     {
-      sched_releasetcb(ptcb);
-      return ENOMEM;
+      errcode = ENOMEM;
+      goto errout_with_tcb;
     }
 
   /* Allocate the stack for the TCB */
@@ -309,9 +316,8 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
   ret = up_create_stack(ptcb, attr->stacksize);
   if (ret != OK)
     {
-      sched_releasetcb(ptcb);
-      sched_free(pjoin);
-      return ENOMEM;
+      errcode = ENOMEM;
+      goto errout_with_join;
     }
 
   /* Should we use the priority and scheduler specified in the
@@ -360,9 +366,8 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
                         TCB_FLAG_TTYPE_PTHREAD);
   if (ret != OK)
     {
-      sched_releasetcb(ptcb);
-      sched_free(pjoin);
-      return EBUSY;
+      errcode = EBUSY;
+      goto errout_with_join;
     }
 
   /* Configure the TCB for a pthread receiving on parameter
@@ -440,10 +445,18 @@ int pthread_create(FAR pthread_t *thread, FAR pthread_attr_t *attr,
       dq_rem((FAR dq_entry_t*)ptcb, (dq_queue_t*)&g_inactivetasks);
       (void)sem_destroy(&pjoin->data_sem);
       (void)sem_destroy(&pjoin->exit_sem);
-      sched_releasetcb(ptcb);
-      sched_free(pjoin);
-      ret = EIO;
+
+      errcode = EIO;
+      goto errout_with_join;
     }
 
   return ret;
+
+errout_with_join:
+  sched_free(pjoin);
+  ptcb->joininfo = NULL;
+
+errout_with_tcb:
+  sched_releasetcb(ptcb);
+  return errcode;
 }
diff --git a/nuttx/sched/sched_releasetcb.c b/nuttx/sched/sched_releasetcb.c
index 0557c829b..50505f579 100644
--- a/nuttx/sched/sched_releasetcb.c
+++ b/nuttx/sched/sched_releasetcb.c
@@ -1,7 +1,7 @@
 /************************************************************************
  * sched/sched_releasetcb.c
  *
- *   Copyright (C) 2007, 2009, 2012 Gregory Nutt. All rights reserved.
+ *   Copyright (C) 2007, 2009, 2012-2013 Gregory Nutt. All rights reserved.
  *   Author: Gregory Nutt <gnutt@nuttx.org>
  *
  * Redistribution and use in source and binary forms, with or without
@@ -177,6 +177,11 @@ int sched_releasetcb(FAR _TCB *tcb)
       ret = up_addrenv_release(tcb);
 #endif
 
+      /* Leave the group (if we did not already leady in task_exithook.c) */
+
+#ifdef HAVE_TASK_GROUP
+      group_leave(tcb);
+#endif
       /* And, finally, release the TCB itself */
 
       sched_free(tcb);
diff --git a/nuttx/sched/sched_waitid.c b/nuttx/sched/sched_waitid.c
index 56bfb36f0..e47e3c38c 100644
--- a/nuttx/sched/sched_waitid.c
+++ b/nuttx/sched/sched_waitid.c
@@ -197,9 +197,9 @@ int waitid(idtype_t idtype, id_t id, FAR siginfo_t *info, int options)
 #ifdef CONFIG_SCHED_CHILD_STATUS
   /* Does this task retain child status? */
 
-  retains = ((rtcb->flags && TCB_FLAG_NOCLDWAIT) == 0);
+  retains = ((rtcb->group->tg_flags && GROUP_FLAG_NOCLDWAIT) == 0);
 
-  if (rtcb->children == NULL && retains)
+  if (rtcb->group->tg_children == NULL && retains)
     {
       /* There are no children */
 
@@ -264,7 +264,7 @@ int waitid(idtype_t idtype, id_t id, FAR siginfo_t *info, int options)
        * instead).
        */
 
-      DEBUGASSERT(!retains || rtcb->children);
+      DEBUGASSERT(!retains || rtcb->group->tg_children);
       if (idtype == P_ALL)
         {
           /* We are waiting for any child to exit */
diff --git a/nuttx/sched/sched_waitpid.c b/nuttx/sched/sched_waitpid.c
index d7484fca9..2d0fe2e48 100644
--- a/nuttx/sched/sched_waitpid.c
+++ b/nuttx/sched/sched_waitpid.c
@@ -312,9 +312,9 @@ pid_t waitpid(pid_t pid, int *stat_loc, int options)
 #ifdef CONFIG_SCHED_CHILD_STATUS
   /* Does this task retain child status? */
 
-  retains = ((rtcb->flags && TCB_FLAG_NOCLDWAIT) == 0);
+  retains = ((rtcb->group->tg_flags && GROUP_FLAG_NOCLDWAIT) == 0);
 
-  if (rtcb->children == NULL && retains)
+  if (rtcb->group->tg_children == NULL && retains)
     {
       err = ECHILD;
       goto errout_with_errno;
@@ -381,7 +381,7 @@ pid_t waitpid(pid_t pid, int *stat_loc, int options)
            * chilren.
            */
 
-          DEBUGASSERT(!retains || rtcb->children);
+          DEBUGASSERT(!retains || rtcb->group->tg_children);
           if (retains && (child = task_exitchild(rtcb)) != NULL)
             {
               /* A child has exited.  Apparently we missed the signal.
diff --git a/nuttx/sched/sig_action.c b/nuttx/sched/sig_action.c
index 7d84b6291..b307c1fb6 100644
--- a/nuttx/sched/sig_action.c
+++ b/nuttx/sched/sig_action.c
@@ -237,7 +237,7 @@ int sigaction(int signo, FAR const struct sigaction *act, FAR struct sigaction *
 
       /* Mark that status should be not be retained */
 
-      rtcb->flags |= TCB_FLAG_NOCLDWAIT;
+      rtcb->group->tg_flags |= GROUP_FLAG_NOCLDWAIT;
 
       /* Free all pending exit status */
 
diff --git a/nuttx/sched/task_childstatus.c b/nuttx/sched/task_childstatus.c
index 6007f03e5..aff2bdf3a 100644
--- a/nuttx/sched/task_childstatus.c
+++ b/nuttx/sched/task_childstatus.c
@@ -111,10 +111,11 @@ static struct child_pool_s g_child_pool;
 static void task_dumpchildren(FAR _TCB *tcb, FAR const char *msg)
 {
   FAR struct child_status_s *child;
+  FAR struct task_group_s   *group = tcb->group;
   int i;
 
-  dbg("Parent TCB=%p: %s\n", tcb, msg);
-  for (i = 0, child = tcb->children; child; i++, child = child->flink)
+  dbg("Parent TCB=%p group=%p: %s\n", tcb, group, msg);
+  for (i = 0, child = group->tg_children; child; i++, child = child->flink)
     {
       dbg("  %d. ch_flags=%02x ch_pid=%d ch_status=%d\n",
           i, child->ch_flags, child->ch_pid, child->ch_status);
@@ -250,10 +251,12 @@ void task_freechild(FAR struct child_status_s *child)
 
 void task_addchild(FAR _TCB *tcb, FAR struct child_status_s *child)
 {
+  FAR struct task_group_s *group = tcb->group;
+
   /* Add the entry into the TCB list of children */
 
-  child->flink  = tcb->children;
-  tcb->children = child;
+  child->flink  = group->tg_children;
+  group->tg_children = child;
 
   task_dumpchildren(tcb, "task_addchild");
 }
@@ -282,11 +285,12 @@ void task_addchild(FAR _TCB *tcb, FAR struct child_status_s *child)
 
 FAR struct child_status_s *task_findchild(FAR _TCB *tcb, pid_t pid)
 {
+  FAR struct task_group_s   *group = tcb->group;
   FAR struct child_status_s *child;
 
   /* Find the status structure with the matching PID  */
 
-  for (child = tcb->children; child; child = child->flink)
+  for (child = group->tg_children; child; child = child->flink)
     {
       if (child->ch_pid == pid)
         {
@@ -318,11 +322,12 @@ FAR struct child_status_s *task_findchild(FAR _TCB *tcb, pid_t pid)
 
 FAR struct child_status_s *task_exitchild(FAR _TCB *tcb)
 {
+  FAR struct task_group_s   *group = tcb->group;
   FAR struct child_status_s *child;
 
   /* Find the status structure of any child task that has exitted. */
 
-  for (child = tcb->children; child; child = child->flink)
+  for (child = group->tg_children; child; child = child->flink)
     {
       if ((child->ch_flags & CHILD_FLAG_EXITED) != 0)
         {
@@ -357,12 +362,13 @@ FAR struct child_status_s *task_exitchild(FAR _TCB *tcb)
 
 FAR struct child_status_s *task_removechild(FAR _TCB *tcb, pid_t pid)
 {
+  FAR struct task_group_s   *group = tcb->group;
   FAR struct child_status_s *curr;
   FAR struct child_status_s *prev;
 
   /* Find the status structure with the matching PID  */
 
-  for (prev = NULL, curr = tcb->children;
+  for (prev = NULL, curr = group->tg_children;
        curr;
        prev = curr, curr = curr->flink)
     {
@@ -384,7 +390,7 @@ FAR struct child_status_s *task_removechild(FAR _TCB *tcb, pid_t pid)
         }
       else
         {
-          tcb->children = curr->flink;
+          group->tg_children = curr->flink;
         }
 
       curr->flink = NULL;
@@ -414,18 +420,19 @@ FAR struct child_status_s *task_removechild(FAR _TCB *tcb, pid_t pid)
 
 void task_removechildren(FAR _TCB *tcb)
 {
+  FAR struct task_group_s   *group = tcb->group;
   FAR struct child_status_s *curr;
   FAR struct child_status_s *next;
 
   /* Remove all child structures for the TCB and return them to the freelist  */
 
-  for (curr = tcb->children; curr; curr = next)
+  for (curr = group->tg_children; curr; curr = next)
     {
       next = curr->flink;
       task_freechild(curr);
     }
 
-  tcb->children = NULL;
+  group->tg_children = NULL;
   task_dumpchildren(tcb, "task_removechildren");
 }
 
diff --git a/nuttx/sched/task_create.c b/nuttx/sched/task_create.c
index 2ed929ab0..f2aeeeec0 100644
--- a/nuttx/sched/task_create.c
+++ b/nuttx/sched/task_create.c
@@ -108,6 +108,7 @@ static int thread_create(const char *name, uint8_t ttype, int priority,
 {
   FAR _TCB *tcb;
   pid_t pid;
+  int errcode;
   int ret;
 
   /* Allocate a TCB for the new task. */
@@ -115,15 +116,28 @@ static int thread_create(const char *name, uint8_t ttype, int priority,
   tcb = (FAR _TCB*)kzalloc(sizeof(_TCB));
   if (!tcb)
     {
+      errcode = ENOMEM;
       goto errout;
     }
 
+  /* Create a new task group */
+
+#ifdef HAVE_TASK_GROUP
+  ret = group_create(tcb);
+  if (ret < 0)
+    {
+      errcode = -ret;
+      goto errout_with_tcb;
+    }
+#endif
+
   /* Associate file descriptors with the new task */
 
 #if CONFIG_NFILE_DESCRIPTORS > 0 || CONFIG_NSOCKET_DESCRIPTORS > 0
   ret = sched_setuptaskfiles(tcb);
   if (ret != OK)
     {
+      errcode = -ret;
       goto errout_with_tcb;
     }
 #endif
@@ -138,6 +152,7 @@ static int thread_create(const char *name, uint8_t ttype, int priority,
   ret = up_create_stack(tcb, stack_size);
   if (ret != OK)
     {
+      errcode = -ret;
       goto errout_with_tcb;
     }
 #endif
@@ -147,6 +162,7 @@ static int thread_create(const char *name, uint8_t ttype, int priority,
   ret = task_schedsetup(tcb, priority, task_start, entry, ttype);
   if (ret != OK)
     {
+      errcode = -ret;
       goto errout_with_tcb;
     }
 
@@ -163,6 +179,8 @@ static int thread_create(const char *name, uint8_t ttype, int priority,
   ret = task_activate(tcb);
   if (ret != OK)
     {
+      errcode = get_errno();
+
       /* The TCB was added to the active task list by task_schedsetup() */
 
       dq_rem((FAR dq_entry_t*)tcb, (dq_queue_t*)&g_inactivetasks);
@@ -175,7 +193,7 @@ errout_with_tcb:
   sched_releasetcb(tcb);
 
 errout:
-  errno = ENOMEM;
+  set_errno(errcode);
   return ERROR;
 }
 
diff --git a/nuttx/sched/task_exithook.c b/nuttx/sched/task_exithook.c
index 1813c12ed..30ce41f71 100644
--- a/nuttx/sched/task_exithook.c
+++ b/nuttx/sched/task_exithook.c
@@ -197,39 +197,28 @@ static inline void task_onexit(FAR _TCB *tcb, int status)
  ****************************************************************************/
 
 #ifdef CONFIG_SCHED_HAVE_PARENT
-static inline void task_sigchild(FAR _TCB *tcb, int status)
+static inline void task_sigchild(FAR _TCB *ptcb, FAR _TCB *ctcb, int status)
 {
-  FAR _TCB *ptcb;
   siginfo_t info;
 
-  /* Only exiting tasks should generate SIGCHLD. pthreads use other
-   * mechansims.
+  /* Only the final exiting thread in a task group should generate SIGCHLD.
+   * If task groups are not supported then we will report SIGCHLD when the
+   * task exits.
    */
 
-  if ((tcb->flags & TCB_FLAG_TTYPE_MASK) == TCB_FLAG_TTYPE_TASK)
+#ifdef CONFIG_SCHED_CHILD_STATUS
+  if (ctcb->group->tg_crefs == 1)
+#else
+  if ((ctcb->flags & TCB_FLAG_TTYPE_MASK) == TCB_FLAG_TTYPE_TASK)
+#endif
     {
-      /* Keep things stationary through the following */
-
-      sched_lock();
-
-      /* Get the TCB of the receiving task */
-
-      ptcb = sched_gettcb(tcb->parent);
-      if (!ptcb)
-        {
-          /* The parent no longer exists... bail */
-
-          sched_unlock();
-          return;
-        }
-
 #ifdef CONFIG_SCHED_CHILD_STATUS
-      /* Check if the parent task has suppressed retention of child exit
+      /* Check if the parent task group has suppressed retention of child exit
        * status information.  Only 'tasks' report exit status, not pthreads.
        * pthreads have a different mechanism.
        */
 
-      if ((ptcb->flags & TCB_FLAG_NOCLDWAIT) == 0)
+      if ((ptcb->group->tg_flags & GROUP_FLAG_NOCLDWAIT) == 0)
         {
           FAR struct child_status_s *child;
 
@@ -255,15 +244,6 @@ static inline void task_sigchild(FAR _TCB *tcb, int status)
       ptcb->nchildren--;
 #endif
 
-      /* Set the parent to an impossible PID.  We do this because under
-       * certain conditions, task_exithook() can be called multiple times.
-       * If this function is called again, sched_gettcb() will fail on the
-       * invalid parent PID above, nchildren will be decremented once and
-       * all will be well.
-       */
-
-      tcb->parent = INVALID_PROCESS_ID;
-
       /* Create the siginfo structure.  We don't actually know the cause.
        * That is a bug. Let's just say that the child task just exit-ted
        * for now.
@@ -272,7 +252,7 @@ static inline void task_sigchild(FAR _TCB *tcb, int status)
       info.si_signo           = SIGCHLD;
       info.si_code            = CLD_EXITED;
       info.si_value.sival_ptr = NULL;
-      info.si_pid             = tcb->pid;
+      info.si_pid             = ctcb->pid;
       info.si_status          = status;
 
       /* Send the signal.  We need to use this internal interface so that we
@@ -280,11 +260,59 @@ static inline void task_sigchild(FAR _TCB *tcb, int status)
        */
 
       (void)sig_received(ptcb, &info);
+    }
+}
+#else
+#  define task_sigchild(ptct,ctcb,status)
+#endif
+
+/****************************************************************************
+ * Name: task_leavegroup
+ *
+ * Description:
+ *   Send the SIGCHILD signal to the parent thread
+ *
+ ****************************************************************************/
+
+#ifdef CONFIG_SCHED_HAVE_PARENT
+static inline void task_leavegroup(FAR _TCB *ctcb, int status)
+{
+  FAR _TCB *ptcb;
+
+  /* Keep things stationary throughout the following */
+
+  sched_lock();
+
+  /* Get the TCB of the receiving, parent task.  We do this early to
+   * handle multiple calls to task_leavegroup.  ctcb->parent is set to an
+   * invalid value below and the following call will fail if we are
+   * called again.
+   */
+
+  ptcb = sched_gettcb(ctcb->parent);
+  if (!ptcb)
+    {
+      /* The parent no longer exists... bail */
+
       sched_unlock();
+      return;
     }
+
+  /* Send SIGCHLD to all members of the parent's task group */
+
+  task_sigchild(ptcb, ctcb, status);
+
+  /* Set the parent to an impossible PID.  We do this because under certain
+   * conditions, task_exithook() can be called multiple times. If this
+   * function is called again, sched_gettcb() will fail on the invalid
+   * parent PID above and all will be well.
+   */
+
+  ctcb->parent = INVALID_PROCESS_ID;
+  sched_unlock();
 }
 #else
-#  define task_sigchild(tcb,status)
+#  define task_leavegroup(ctcb,status)
 #endif
 
 /****************************************************************************
@@ -363,9 +391,9 @@ void task_exithook(FAR _TCB *tcb, int status)
 
   task_onexit(tcb, status);
 
-  /* Send SIGCHLD to the parent of the exit-ing task */
+  /* Leave the task group */
 
-  task_sigchild(tcb, status);
+  task_leavegroup(tcb, status);
 
   /* Wakeup any tasks waiting for this task to exit */
 
@@ -379,10 +407,12 @@ void task_exithook(FAR _TCB *tcb, int status)
   (void)lib_flushall(tcb->streams);
 #endif
 
-  /* Discard any un-reaped child status (no zombies here!) */
+  /* Leave the task group.  Perhaps discarding any un-reaped child
+   * status (no zombies here!)
+   */
 
-#if defined(CONFIG_SCHED_HAVE_PARENT) && defined(CONFIG_SCHED_CHILD_STATUS)
-  task_removechildren(tcb);
+#ifdef HAVE_TASK_GROUP
+  group_leave(tcb);
 #endif
 
   /* Free all file-related resources now.  This gets called again
diff --git a/nuttx/sched/task_init.c b/nuttx/sched/task_init.c
index 0f0fdc68e..5ba6b7e7d 100644
--- a/nuttx/sched/task_init.c
+++ b/nuttx/sched/task_init.c
@@ -1,7 +1,7 @@
 /****************************************************************************
  * sched/task_init.c
  *
- *   Copyright (C) 2007, 2009 Gregory Nutt. All rights reserved.
+ *   Copyright (C) 2007, 2009, 2013 Gregory Nutt. All rights reserved.
  *   Author: Gregory Nutt <gnutt@nuttx.org>
  *
  * Redistribution and use in source and binary forms, with or without
@@ -42,6 +42,8 @@
 #include <sys/types.h>
 #include <stdint.h>
 #include <sched.h>
+#include <errno.h>
+
 #include <nuttx/arch.h>
 
 #include "os_internal.h"
@@ -102,10 +104,10 @@
  *                parameters are required, argv may be NULL.
  *
  * Return Value:
- *   OK on success; ERROR on failure.  (See task_schedsetup() for possible
- *   failure conditions).  On failure, the caller is responsible for freeing
- *   the stack memory and for calling  sched_releasetcb() to free the TCB
- *   (which could be in most any state).
+ *   OK on success; ERROR on failure with errno set appropriately.  (See
+ *   task_schedsetup() for possible failure conditions).  On failure, the
+ *   caller is responsible for freeing the stack memory and for calling
+ *   sched_releasetcb() to free the TCB (which could be in most any state).
  *
  ****************************************************************************/
 
@@ -118,14 +120,28 @@ int task_init(FAR _TCB *tcb, const char *name, int priority,
               main_t entry, const char *argv[])
 #endif
 {
+  int errcode;
   int ret;
 
+  /* Create a new task group */
+
+#ifdef HAVE_TASK_GROUP
+  ret = group_create(tcb);
+  if (ret < 0)
+    {
+      errcode = -ret;
+      goto errout;
+    }
+#endif
+
  /* Associate file descriptors with the new task */
 
 #if CONFIG_NFILE_DESCRIPTORS > 0 || CONFIG_NSOCKET_DESCRIPTORS > 0
-  if (sched_setuptaskfiles(tcb) != OK)
+  ret = sched_setuptaskfiles(tcb);
+  if (ret < 0)
     {
-      return ERROR;
+      errcode = -ret;
+      goto errout_with_group;
     }
 #endif
 
@@ -143,13 +159,27 @@ int task_init(FAR _TCB *tcb, const char *name, int priority,
 
   ret = task_schedsetup(tcb, priority, task_start, entry,
                         TCB_FLAG_TTYPE_TASK);
-  if (ret == OK)
+  if (ret < OK)
     {
-      /* Setup to pass parameters to the new task */
-
-      (void)task_argsetup(tcb, name, argv);
+      errcode = -ret;
+      goto errout_with_env;
     }
 
-  return ret;
+  /* Setup to pass parameters to the new task */
+
+  (void)task_argsetup(tcb, name, argv);
+  return OK;
+
+errout_with_env:
+  env_release(tcb);
+
+errout_with_group:
+#ifdef HAVE_TASK_GROUP
+  group_leave(tcb);
+
+errout:
+#endif
+  set_errno(errcode);
+  return ERROR;
 }
 
diff --git a/nuttx/sched/task_reparent.c b/nuttx/sched/task_reparent.c
index 3a7ece37d..0b502dcce 100644
--- a/nuttx/sched/task_reparent.c
+++ b/nuttx/sched/task_reparent.c
@@ -138,11 +138,11 @@ int task_reparent(pid_t ppid, pid_t chpid)
   child = task_removechild(otcb, chpid);
   if (child)
     {
-      /* Has the new parent supressed child exit status? */
+      /* Has the new parent's task group supressed child exit status? */
 
-      if ((ptcb->flags && TCB_FLAG_NOCLDWAIT) == 0)
+      if ((ptcb->group->tg_flags && GROUP_FLAG_NOCLDWAIT) == 0)
         {
-          /* No.. Add the child status entry to the new parent TCB */
+          /* No.. Add the child status entry to the new parent's task group */
 
           task_addchild(ptcb, child);
         }
@@ -159,11 +159,11 @@ int task_reparent(pid_t ppid, pid_t chpid)
     }
   else
     {
-      /* This would not be an error if the original parent has
+      /* This would not be an error if the original parent's task group has
        * suppressed child exit status.
        */
 
-      ret = ((otcb->flags && TCB_FLAG_NOCLDWAIT) == 0) ? -ENOENT : OK;
+      ret = ((otcb->group->tg_flags && GROUP_FLAG_NOCLDWAIT) == 0) ? -ENOENT : OK;
     }
 #else
   DEBUGASSERT(otcb->nchildren > 0);
diff --git a/nuttx/sched/task_setup.c b/nuttx/sched/task_setup.c
index 66c948cfd..7675c7481 100644
--- a/nuttx/sched/task_setup.c
+++ b/nuttx/sched/task_setup.c
@@ -1,7 +1,7 @@
 /****************************************************************************
  * sched/task_setup.c
  *
- *   Copyright (C) 2007-2012 Gregory Nutt. All rights reserved.
+ *   Copyright (C) 2007-2013 Gregory Nutt. All rights reserved.
  *   Author: Gregory Nutt <gnutt@nuttx.org>
  *
  * Redistribution and use in source and binary forms, with or without
@@ -179,13 +179,11 @@ static inline void task_saveparent(FAR _TCB *tcb, uint8_t ttype)
   tcb->parent = rtcb->pid;
 
 #ifdef CONFIG_SCHED_CHILD_STATUS
-  /* Exit status only needs to be retained for the case of tasks, however.
-   * Tasks can also suppress retention of their child status by applying
-   * the SA_NOCLDWAIT flag with sigaction()/
+  /* Tasks can also suppress retention of their child status by applying
+   * the SA_NOCLDWAIT flag with sigaction().
    */
 
-  if (ttype == TCB_FLAG_TTYPE_TASK &&
-      (rtcb->flags && TCB_FLAG_NOCLDWAIT) == 0)
+  if ((rtcb->group->tg_flags && GROUP_FLAG_NOCLDWAIT) == 0)
     {
       FAR struct child_status_s *child;