ThreadCreate(), ThreadCreate_r()

Create a thread

Synopsis:

#include <sys/neutrino.h>

int ThreadCreate( 
          pid_t pid,
          void* (func)( void* ),
          void* arg,
          const struct _thread_attr* attr );

int ThreadCreate_r( 
          pid_t pid,
          void* (func)( void* ),
          void* arg,
          const struct _thread_attr* attr );

Arguments:

pid
The ID of the process that you want to create the thread in, or 0 to create the thread in the current process.
func
A pointer to the function that you want the thread to execute. The arg argument that you pass to ThreadCreate() is passed to func() as its sole argument. If func() returns, it returns to the address defined in the __exitfunc member of attr.
arg
A pointer to any data that you want to pass to func.
attr
A pointer to a _thread_attr structure that specifies the attributes for the new thread, or NULL if you want to use the default attributes. For more information, see Thread attributes,” below.
Note: If you modify the attributes after creating the thread, the thread isn't affected.

Library:

libc

Use the -l c option to qcc to link against this library. This library is usually included automatically.

Description:

These kernel calls create a new thread of execution, with attributes specified by attr, within the process specified by pid. If pid is zero, the current process is used. The ThreadCreate() and ThreadCreate_r() functions are identical, except in the way they indicate errors. See the Returns section for details.

Note:
  • Only the Process Manager can create threads in another process.
  • Instead of using these kernel calls directly, consider calling pthread_create().

The new thread shares all resources of the process in which it's created. This includes memory, timers, channels and connections. The standard C library contains mutexes to make it thread-safe.

Thread attributes

The _thread_attr structure pointed to by attr contains at least the following members:

int __flags
Zero or more of the following (the default of each pair of flags is always zero).

Joinable or detached:

  • PTHREAD_CREATE_JOINABLE (default) — put the thread into a zombie state when it terminates. It stays in this state until you retrieve its exit status or detach the thread.
  • PTHREAD_CREATE_DETACHED — create the thread in the detached state; it doesn't become a zombie. You can't call ThreadJoin() for a detached thread.

Scheduling inheritence:

  • PTHREAD_INHERIT_SCHED (default) — use the scheduling attributes of the creating thread for the new thread.
  • PTHREAD_EXPLICIT_SCHED — take the scheduling policy and parameters for the new thread from the __policy and __param members of attr.

Scheduling scope:

  • PTHREAD_SCOPE_SYSTEM (default) — schedule the thread against all threads in the system.
  • PTHREAD_SCOPE_PROCESS — don't set this flag; the QNX Neutrino RTOS implements true microkernel threads that have only a system scope.

Multisignalling:

  • PTHREAD_MULTISIG_ALLOW (default) — if the thread dies because of an unblocked, uncaught signal, terminate all threads, and hence, the process.
  • PTHREAD_MULTISIG_DISALLOW — terminate only this thread; all other threads in the process are unaffected.

Cancellation:

  • PTHREAD_CANCEL_DEFERRED (default) — cancellation occurs only at cancellation points as defined by ThreadCancel().
  • PTHREAD_CANCEL_ASYNCHRONOUS — every opcode executed by the thread is considered a cancellation point. The POSIX and C library aren't asynchronous-cancel safe.

(QNX Neutrino 7.0.1 or later) Initial state:

  • PTHREAD_CREATE_NOT_SUSPENDED (default) — when the thread is created, it isn't put into a suspended state.
  • PTHREAD_CREATE_SUSPENDED — when the thread is created, it's put into a suspended state.

    Starting a thread in a suspended state is equivalent to creating a thread and immediately calling ThreadCtlExt(), specifying the thread ID and the _NTO_TCTL_ONE_THREAD_HOLD command. You can unsuspend the thread with the _NTO_TCTL_ONE_THREAD_CONT command.

size_t __stacksize
If __stackaddr isn't NULL, then __stacksize specifies the size of the thread stack defined in __stackaddr. If the stack size is less than PTHREAD_STACK_MIN, the call fails with an error of EINVAL.

If __stackaddr is NULL, then __stacksize specifies the minimum size of stack to dynamically allocate. If you set __stacksize to zero, then a system-defined amount is used.

void* __stackaddr
NULL, or the address of a stack that you want the thread to use. Set the __stacksize member to the size of the stack.

If you provide a non-NULL __stackaddr, it's your responsibility to release the stack when the thread dies. If __stackaddr is NULL, then the kernel dynamically allocates a stack on thread creation and automatically releases it on the thread's death.

void* (__exitfunc)(void* status)
The address to return to if the thread function returns.
Note: The thread returns to __exitfunc. This means that the status variable isn't passed as a normal parameter. Instead, it appears in the return-value position dictated by the CPU's calling convention (e.g. EAX on an x86, and so on).

The __exitfunc function normally has to have compiler- and CPU-specific manipulation to access the status data (pulling it from the return register location to a proper local variable). Alternatively, you can write the __exitfunc function in assembly language for each CPU.

int __policy
The scheduling policy, as defined by the SchedSet() kernel call. This member is used only if you set the PTHREAD_EXPLICIT_SCHED flag. If you want the thread to inherit the policy, but you want to specify the scheduling parameters in the __param member, set the PTHREAD_EXPLICIT_SCHED flag and set the __policy member to SCHED_NOCHANGE.
struct sched_param __param
A sched_param structure that specifies the scheduling parameters, as defined by the SchedSet() kernel call. This member is used only if you set the PTHREAD_EXPLICIT_SCHED flag.
unsigned __guardsize
The amount of memory to use for the guard area of the thread's stack; for more information, see pthread_attr_setguardsize()
unsigned __prealloc
The amount of the memory to preallocate for a thread's MAP_LAZY stack; for more information, see pthread_attr_setstackprealloc().

Signal state

The signal state of the new thread is initialized as follows:

Local storage for private data

Each thread contains a thread local storage area for its private data. User applications don't need to directly access this area, but you can get a pointer to it by calling __tls(). The thread local storage is defined by the structure _thread_local_storage. Both of these are defined in <sys/storage.h>.

Blocking states

These calls don't block.

Returns:

The thread ID of the newly created thread. If an error occurs:

Errors:

EAGAIN
All kernel thread objects are in use.
EFAULT
A fault occurred when the kernel tried to access the buffers provided.
EINVAL
The scheduling policy or priority was invalid, or the stack size is less than PTHREAD_STACK_MIN.
ENOTSUP
PTHREAD_SCOPE_PROCESS was requested. All kernel threads are PTHREAD_SCOPE_SYSTEM.
EPERM
The calling thread doesn't have sufficient permission to create a thread in another process. Only a thread with a process ID of 1 can create threads in other processes.
ESRCH
The process indicated by pid doesn't exist.

Classification:

QNX Neutrino

Safety:  
Cancellation point No
Interrupt handler No
Signal handler Yes
Thread Yes

Caveats:

The QNX Neutrino interpretation of PTHREAD_STACK_MIN is enough memory to run a thread that does nothing:

void nothingthread( void )
{
    return;
}