path: root/erts/doc/src/erl_driver.xml

<?xml version="1.0" encoding="latin1" ?>
<!DOCTYPE cref SYSTEM "cref.dtd">

<cref>
  <header>
    <copyright>
      <year>2001</year><year>2010</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      The contents of this file are subject to the Erlang Public License,
      Version 1.1, (the "License"); you may not use this file except in
      compliance with the License. You should have received a copy of the
      Erlang Public License along with this software. If not, it can be
      retrieved online at http://www.erlang.org/.

      Software distributed under the License is distributed on an "AS IS"
      basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
      the License for the specific language governing rights and limitations
      under the License.

    </legalnotice>

    <title>erl_driver</title>
    <prepared>Jakob Cederlund</prepared>
    <responsible>Jakob Cederlund</responsible>
    <docno>1</docno>
    <approved></approved>
    <checked></checked>
    <date>2000-11-27</date>
    <rev>PA1</rev>
    <file>erl_driver.xml</file>
  </header>
  <lib>erl_driver</lib>
  <libsummary>API functions for an Erlang driver</libsummary>
  <description>
    <p>As of erts version 5.5.3 the driver interface has been extended
      (see <seealso marker="driver_entry#extended_marker">extended marker</seealso>).
      The extended interface introduce
      <seealso marker="erl_driver#version_management">version management</seealso>,
      the possibility to pass capability flags
      (see <seealso marker="driver_entry#driver_flags">driver flags</seealso>)
      to the runtime system at driver initialization, and some new
      driver API functions.      </p>
    <note>
      <p>Old drivers (compiled with an <c>erl_driver.h</c> from an
        earlier erts version than 5.5.3) have to be recompiled
        (but does not have to use the extended interface).</p>
    </note>
    <p>The driver calls back to the emulator, using the API
      functions declared in <c>erl_driver.h</c>. They are used for
      outputting data from the driver, using timers, etc.</p>
    <p>A driver is a library with a set of function that the emulator
      calls, in response to Erlang functions and message
      sending. There may be multiple instances of a driver, each
      instance is connected to an Erlang port. Every port has a port
      owner process. Communication with the port is normally done
      through the port owner process.</p>
    <p>Most of the functions takes the <c>port</c> handle as an
      argument. This identifies the driver instance. Note that this
      port handle must be stored by the driver, it is not given when
      the driver is called from the emulator (see
      <seealso marker="driver_entry#emulator">driver_entry</seealso>).</p>
    <p>Some of the functions takes a parameter of type
      <c>ErlDrvBinary</c>, a driver binary. It should be both
      allocated and freed by the caller. Using a binary directly avoid
      one extra copying of data.</p>
    <p>Many of the output functions has a "header buffer", with
      <c>hbuf</c> and <c>hlen</c> parameters. This buffer is sent as a
      list before the binary (or list, depending on port mode) that is
      sent. This is convenient when matching on messages received from
      the port. (Although in the latest versions of Erlang, there is
      the binary syntax, that enables you to match on the beginning of
      a binary.)
            <marker id="smp_support"></marker>
</p>
    <p>In the runtime system with SMP support, drivers are locked either
      on driver level or port level (driver instance level). By default
      driver level locking will be used, i.e., only one emulator thread
      will execute code in the driver at a time. If port level locking
      is used, multiple emulator threads may execute code in the driver
      at the same time. There will only be one thread at a time calling
      driver call-backs corresponding to the same port, though. In order
      to enable port level locking set the <c>ERL_DRV_FLAG_USE_PORT_LOCKING</c>
      <seealso marker="driver_entry#driver_flags">driver flag</seealso> in
      the <seealso marker="driver_entry">driver_entry</seealso>
      used by the driver. When port level locking is used it is the
      responsibility of the driver writer to synchronize all accesses
      to data shared by the ports (driver instances).</p>
    <p>Most drivers written before the runtime system with SMP
      support existed will be able to run in the runtime system
      with SMP support without being rewritten if driver
      level locking is used.</p>
    <note>
      <p>It is assumed that drivers does not access other drivers. If
        drivers should access each other they have to provide their own
        mechanism for thread safe synchronization. Such "inter driver
        communication" is strongly discouraged.</p>
    </note>
    <p>Previously, in the runtime system without SMP support,
      specific driver call-backs were always called from the same
      thread. This is <em>not</em> the case in the runtime system
      with SMP support. Regardless of locking scheme used, calls
      to driver call-backs may be made from different threads, e.g.,
      two consecutive calls to exactly the same call-back for exactly
      the same port may be made from two different threads. This
      will for <em>most</em> drivers not be a problem, but it might.
      Drivers that depend on all call-backs being called in the
      same thread, <em>have</em> to be rewritten before being used
      in the runtime system with SMP support.</p>
    <note>
      <p>Regardless of locking scheme used, calls to driver
        call-backs may be made from different threads.</p>
    </note>
    <p>Most functions in this API are <em>not</em> thread-safe, i.e.,
      they may <em>not</em> be called from an arbitrary thread. Function
      that are not documented as thread-safe may only be called from
      driver call-backs or function calls descending from a driver
      call-back call. Note that driver call-backs may be called from
      different threads. This, however, is not a problem for any
      functions in this API, since the emulator have control over
      these threads.</p>
    <note>
      <p>Functions not explicitly documented as thread-safe are
        <em>not</em> thread-safe. Also note that some functions
        are <em>only</em> thread safe when used in a runtime
        system with SMP support.</p>
    </note>
  </description>

  <section>
    <title>FUNCTIONALITY</title>
    <p>All functions that a driver needs to do with Erlang are
      performed through driver API functions. There are functions
      for the following functionality:</p>
    <taglist>
      <tag>Timer functions</tag>
      <item>Timer functions are used to control the timer that a driver
       may use. The timer will have the emulator call the
       <seealso marker="driver_entry#timeout">timeout</seealso> entry
       function after a specified time. Only one timer is available
       for each driver instance.</item>
      <tag>Queue handling</tag>
      <item>
        <p>Every driver instance has an associated queue. This queue is a
          <c>SysIOVec</c> that works as a buffer. It's mostly used for
          the driver to buffer data that should be written to a device,
          it is a byte stream. If the port owner process closes the
          driver, and the queue is not empty, the driver will not be
          closed. This enables the driver to flush its buffers before
          closing.</p>
        <p>The queue can be manipulated from arbitrary threads if
          a port data lock is used. See documentation of the
          <seealso marker="#ErlDrvPDL">ErlDrvPDL</seealso> type for
          more information.</p>
      </item>
      <tag>Output functions</tag>
      <item>With the output functions, the driver sends data back
       the emulator. They will be received as messages by the port owner
       process, see <c>open_port/2</c>. The vector function and the
       function taking a driver binary is faster, because that avoid
       copying the data buffer. There is also a fast way of sending
       terms from the driver, without going through the binary term
       format.</item>
      <tag>Failure</tag>
      <item>The driver can exit and signal errors up to Erlang. This is
       only for severe errors, when the driver can't possibly keep
       open.</item>
      <tag>Asynchronous calls</tag>
      <item>The latest Erlang versions (R7B and later) has provision for
       asynchronous function calls, using a thread pool provided by
       Erlang. There is also a select call, that can be used for
       asynchronous drivers.</item>
      <tag><marker id="multi_threading">Multi-threading</marker></tag>
      <item>
      <p>A POSIX thread like API for multi-threading is provided. The
         Erlang driver thread API only provide a subset of the functionality
	 provided by the POSIX thread API. The subset provided is
	 more or less the basic functionality needed for multi-threaded
	 programming:
      </p>
      <list>
	 <item><seealso marker="#ErlDrvTid">Threads</seealso></item>
	 <item><seealso marker="#ErlDrvMutex">Mutexes</seealso></item>
	 <item><seealso marker="#ErlDrvCond">Condition variables</seealso></item>
	 <item><seealso marker="#ErlDrvRWLock">Read/Write locks</seealso></item>
	 <item><seealso marker="#ErlDrvTSDKey">Thread specific data</seealso></item>
      </list>
      <p>The Erlang driver thread API can be used in conjunction with
         the POSIX thread API on UN-ices and with the Windows native thread
	 API on Windows. The Erlang driver thread API has the advantage of
	 being portable, but there might exist situations where you want to
	 use functionality from the POSIX thread API or the Windows
	 native thread API.
      </p>
      <p>The Erlang driver thread API only return error codes when it is
         reasonable to recover from an error condition. If it isn't reasonable
	 to recover from an error condition, the whole runtime system is
	 terminated. For example, if a create mutex operation fails, an error
	 code is returned, but if a lock operation on a mutex fails, the
	 whole runtime system is terminated.
      </p>
      <p>Note that there exist no "condition variable wait with timeout" in
         the Erlang driver thread API. This is due to issues with
	 <c>pthread_cond_timedwait()</c>. When the system clock suddenly
	 is changed, it isn't always guaranteed that you will wake up from
	 the call as expected. An Erlang runtime system has to be able to
	 cope with sudden changes of the system clock. Therefore, we have
	 omitted it from the Erlang driver thread API. In the Erlang driver
	 case, timeouts can and should be handled with the timer functionality
	 of the Erlang driver API.
      </p>
      <p>In order for the Erlang driver thread API to function, thread
         support has to be enabled in the runtime system. An Erlang driver
	 can check if thread support is enabled by use of
	 <seealso marker="#driver_system_info">driver_system_info()</seealso>.
	 Note that some functions in the Erlang driver API are thread-safe
	 only when the runtime system has SMP support, also this
	 information can be retrieved via
	 <seealso marker="#driver_system_info">driver_system_info()</seealso>.
	 Also note that a lot of functions in the Erlang driver API are
	 <em>not</em> thread-safe regardless of whether SMP support is
	 enabled or not. If a function isn't documented as thread-safe it
	 is <em>not</em> thread-safe.
      </p>
      <p><em>NOTE</em>: When executing in an emulator thread, it is
         <em>very important</em> that you unlock <em>all</em> locks you
	 have locked before letting the thread out of your control;
	 otherwise, you are <em>very likely</em> to deadlock the whole
	 emulator. If you need to use thread specific data in an emulator
	 thread, only have the thread specific data set while the thread is
	 under your control, and clear the thread specific data before
	 you let the thread out of your control.
      </p>
      <p>In the future there will probably be debug functionality
         integrated with the Erlang driver thread API. All functions
	 that create entities take a <c>name</c> argument. Currently
	 the <c>name</c> argument is unused, but it will be used when
	 the debug functionality has been implemented. If you name all
	 entities created well, the debug functionality will be able
	 to give you better error reports.
      </p>
      </item>
      <tag>Adding / remove drivers</tag>
      <item>A driver can add and later remove drivers.</item>
      <tag>Monitoring processes</tag>
      <item>A driver can monitor a process that does not own a port.</item>
      <tag><marker id="version_management">Version management</marker></tag>
      <item>
        <p>Version management is enabled for drivers that have set the
          <seealso marker="driver_entry#extended_marker">extended_marker</seealso>
          field of their
          <seealso marker="driver_entry">driver_entry</seealso>
          to <c>ERL_DRV_EXTENDED_MARKER</c>. <c>erl_driver.h</c> defines
          <c>ERL_DRV_EXTENDED_MARKER</c>,
          <c>ERL_DRV_EXTENDED_MAJOR_VERSION</c>, and
          <c>ERL_DRV_EXTENDED_MINOR_VERSION</c>.
          <c>ERL_DRV_EXTENDED_MAJOR_VERSION</c> will be incremented when
          driver incompatible changes are made to the Erlang runtime
          system. Normally it will suffice to recompile drivers when the
          <c>ERL_DRV_EXTENDED_MAJOR_VERSION</c> has changed, but it
          could, under rare circumstances, mean that drivers have to
          be slightly modified. If so, this will of course be documented.
          <c>ERL_DRV_EXTENDED_MINOR_VERSION</c> will be incremented when
          new features are added. The runtime system use the minor version
          of the driver to determine what features to use.
          The runtime system will refuse to load a driver if the major
          versions differ, or if the major versions are equal and the
          minor version used by the driver is greater than the one used
          by the runtime system.</p>
        <p>The emulator tries to check that a driver that doesn't use the
          extended driver interface isn't incompatible when loading it.
          It can, however, not make sure that it isn't incompatible. Therefore,
          when loading a driver that doesn't use the extended driver
          interface, there is a risk that it will be loaded also when
          the driver is incompatible. When the driver use the extended driver
          interface, the emulator can verify that it isn't of an incompatible
          driver version. You are therefore advised to use the extended driver
          interface.</p>
      </item>
    </taglist>
  </section>

  <section>
    <title>DATA TYPES</title>

    <taglist>
     <tag><marker id="ErlDrvSysInfo"/>ErlDrvSysInfo</tag>
      <item>
      <p/>
      <code type="none">
typedef struct ErlDrvSysInfo {
   int driver_major_version;
   int driver_minor_version;
   char *erts_version;
   char *otp_release;
   int thread_support;
   int smp_support;
   int async_threads;
   int scheduler_threads;
   int nif_major_version;
   int nif_minor_version;
} ErlDrvSysInfo;
      </code>

      <p>
        The <c>ErlDrvSysInfo</c> structure is used for storage of
        information about the Erlang runtime system.
        <seealso marker="#driver_system_info">driver_system_info()</seealso>
        will write the system information when passed a reference to
        a <c>ErlDrvSysInfo</c> structure. A description of the
        fields in the structure follow:
      </p>
      <taglist>
        <tag><c>driver_major_version</c></tag>
        <item>The value of
          <seealso marker="#version_management">ERL_DRV_EXTENDED_MAJOR_VERSION</seealso>
           when the runtime system was compiled. This value is the same
           as the value of
          <seealso marker="#version_management">ERL_DRV_EXTENDED_MAJOR_VERSION</seealso>
           used when compiling the driver; otherwise, the runtime system
           would have refused to load the driver.
        </item>
        <tag><c>driver_minor_version</c></tag>
        <item>The value of
          <seealso marker="#version_management">ERL_DRV_EXTENDED_MINOR_VERSION</seealso>
           when the runtime system was compiled. This value might differ
           from the value of
          <seealso marker="#version_management">ERL_DRV_EXTENDED_MINOR_VERSION</seealso>
           used when compiling the driver.
        </item>
        <tag><c>erts_version</c></tag>
        <item>A string containing the version number of the runtime system
           (the same as returned by
          <seealso marker="erts:erlang#system_info_version">erlang:system_info(version)</seealso>).
        </item>
        <tag><c>otp_release</c></tag>
        <item>A string containing the OTP release number
           (the same as returned by
          <seealso marker="erts:erlang#system_info_otp_release">erlang:system_info(otp_release)</seealso>).
        </item>
        <tag><c>thread_support</c></tag>
        <item>A value <c>!= 0</c> if the runtime system has thread support;
           otherwise, <c>0</c>.
        </item>
        <tag><c>smp_support</c></tag>
        <item>A value <c>!= 0</c> if the runtime system has SMP support;
           otherwise, <c>0</c>.
        </item>
        <tag><c>thread_support</c></tag>
        <item>A value <c>!= 0</c> if the runtime system has thread support;
           otherwise, <c>0</c>.
        </item>
        <tag><c>smp_support</c></tag>
        <item>A value <c>!= 0</c> if the runtime system has SMP support;
           otherwise, <c>0</c>.
        </item>
        <tag><c>async_threads</c></tag>
        <item>The number of async threads in the async thread pool used
           by <seealso marker="#driver_async">driver_async()</seealso>
           (the same as returned by
          <seealso marker="erts:erlang#system_info_thread_pool_size">erlang:system_info(thread_pool_size)</seealso>).
        </item>
        <tag><c>scheduler_threads</c></tag>
        <item>The number of scheduler threads used by the runtime system
           (the same as returned by
          <seealso marker="erts:erlang#system_info_schedulers">erlang:system_info(schedulers)</seealso>).
        </item>
        <tag><c>nif_major_version</c></tag>
        <item>The value of <c>ERL_NIF_MAJOR_VERSION</c> when the runtime system was compiled.
        </item>
        <tag><c>nif_minor_version</c></tag>
        <item>The value of <c>ERL_NIF_MINOR_VERSION</c> when the runtime system was compiled.
        </item>
        </taglist>
      </item>
      <tag><marker id="ErlDrvBinary"/>ErlDrvBinary</tag>
      <item>
      <p/>
      <code type="none">
typedef struct ErlDrvBinary {
   int orig_size;
   char orig_bytes[];
} ErlDrvBinary;
</code>
        <p>The <c>ErlDrvBinary</c> structure is a binary, as sent
          between the emulator and the driver. All binaries are
          reference counted; when <c>driver_binary_free</c> is called,
          the reference count is decremented, when it reaches zero,
          the binary is deallocated. The <c>orig_size</c> is the size
          of the binary, and <c>orig_bytes</c> is the buffer. The
          <c>ErlDrvBinary</c> does not have a fixed size, its size is
          <c>orig_size + 2 * sizeof(int)</c>.</p>
        <note>
          <p>The <c>refc</c> field has been removed. The reference count of
            an <c>ErlDrvBinary</c> is now stored elsewhere. The
            reference count of an <c>ErlDrvBinary</c> can be accessed via
            <seealso marker="#driver_binary_get_refc">driver_binary_get_refc()</seealso>,
            <seealso marker="#driver_binary_inc_refc">driver_binary_inc_refc()</seealso>,
            and
            <seealso marker="#driver_binary_dec_refc">driver_binary_dec_refc()</seealso>.</p>
        </note>
        <p>Some driver calls, such as <c>driver_enq_binary</c>,
          increments the driver reference count, and others, such as
          <c>driver_deq</c> decrements it.</p>
        <p>Using a driver binary instead of a normal buffer, is often
          faster, since the emulator doesn't need to copy the data,
          only the pointer is used.</p>
        <p>A driver binary allocated in the driver, with
          <c>driver_alloc_binary</c>, should be freed in the driver (unless otherwise stated),
          with <c>driver_free_binary</c>. (Note that this doesn't
          necessarily deallocate it, if the driver is still referred
          in the emulator, the ref-count will not go to zero.)</p>
        <p>Driver binaries are used in the <c>driver_output2</c> and
          <c>driver_outputv</c> calls, and in the queue. Also the
          driver call-back <seealso marker="driver_entry#outputv">outputv</seealso> uses driver
          binaries.</p>
        <p>If the driver of some reason or another, wants to keep a
          driver binary around, in a static variable for instance, the
          reference count should be incremented,
          and the binary can later be freed in the <seealso marker="driver_entry#stop">stop</seealso> call-back, with
          <c>driver_free_binary</c>.</p>
        <p>Note that since a driver binary is shared by the driver and
          the emulator, a binary received from the emulator or sent to
          the emulator, must not be changed by the driver.</p>
        <p>From erts version 5.5 (OTP release R11B), orig_bytes is 
          guaranteed to be properly aligned for storage of an array of
          doubles (usually 8-byte aligned).</p>
      </item>
      <tag>ErlDrvData</tag>
      <item>
        <p>The <c>ErlDrvData</c> is a handle to driver-specific data,
          passed to the driver call-backs. It is a pointer, and is
          most often casted to a specific pointer in the driver.</p>
      </item>
      <tag>SysIOVec</tag>
      <item>
        <p>This is a system I/O vector, as used by <c>writev</c> on
          unix and <c>WSASend</c> on Win32. It is used in
          <c>ErlIOVec</c>.</p>
      </item>
      <tag><marker id="ErlIOVec"/>ErlIOVec</tag>
      <item>
      <p/>
      <code type="none">
typedef struct ErlIOVec {
  int vsize;
  int size;
  SysIOVec* iov;
  >ErlDrvBinary** binv;
} ErlIOVec;
</code>
        <p>The I/O vector used by the emulator and drivers, is a list
          of binaries, with a <c>SysIOVec</c> pointing to the buffers
          of the binaries. It is used in <c>driver_outputv</c> and the
          <seealso marker="driver_entry#outputv">outputv</seealso>
          driver call-back. Also, the driver queue is an
          <c>ErlIOVec</c>.</p>
      </item>

      <tag>ErlDrvMonitor</tag>
      <item>
        <p>When a driver creates a monitor for a process, a
          <c>ErlDrvMonitor</c> is filled in. This is an opaque
          data-type which can be assigned to but not compared without
          using the supplied compare function (i.e. it behaves like a struct).</p>
        <p>The driver writer should provide the memory for storing the
          monitor when calling <seealso marker="#driver_monitor_process">driver_monitor_process</seealso>. The
          address of the data is not stored outside of the driver, so
          the <c>ErlDrvMonitor</c> can be used as any other datum, it
          can be copied, moved in memory, forgotten etc.</p>
      </item>
      <tag><marker id="ErlDrvNowData"/>ErlDrvNowData</tag>
      <item>
        <p>The <c>ErlDrvNowData</c> structure holds a timestamp
          consisting of three values measured from some arbitrary
          point in the past. The three structure members are:</p>
        <taglist>
          <tag>megasecs</tag>
          <item>The number of whole megaseconds elapsed since the arbitrary
           point in time</item>
          <tag>secs</tag>
          <item>The number of whole seconds elapsed since the arbitrary
           point in time</item>
          <tag>microsecs</tag>
          <item>The number of whole microseconds elapsed since the arbitrary
           point in time</item>
        </taglist>
      </item>
      <tag><marker id="ErlDrvPDL"/>ErlDrvPDL</tag>
      <item>
        <p>If certain port specific data have to be accessed from other
          threads than those calling the driver call-backs, a port data lock
          can be used in order to synchronize the operations on the data.
          Currently, the only port specific data that the emulator
          associates with the port data lock is the driver queue.</p>
        <p>Normally a driver instance does not have a port data lock. If
          the driver instance want to use a port data lock, it has to
          create the port data lock by calling
          <seealso marker="#driver_pdl_create">driver_pdl_create()</seealso>.
          <em>NOTE</em>: Once the port data lock has been created, every
          access to data associated with the port data lock have to be done
          while having the port data lock locked. The port data lock is
          locked, and unlocked, respectively, by use of
          <seealso marker="#driver_pdl_lock">driver_pdl_lock()</seealso>, and
          <seealso marker="#driver_pdl_unlock">driver_pdl_unlock()</seealso>.</p>
        <p>A port data lock is reference counted, and when the reference
          count reach zero, it will be destroyed. The emulator will at
          least increment the reference count once when the lock is
          created and decrement it once when the port associated with
          the lock terminates. The emulator will also increment the
          reference count when an async job is enqueued and decrement
          it after an async job has been invoked, or canceled. Besides
          this, it is the responsibility of the driver to ensure that
          the reference count does not reach zero before the last use
          of the lock by the driver has been made. The reference count
          can be read, incremented, and decremented, respectively, by
          use of
          <seealso marker="#driver_pdl_get_refc">driver_pdl_get_refc()</seealso>,
          <seealso marker="#driver_pdl_inc_refc">driver_pdl_inc_refc()</seealso>, and
          <seealso marker="#driver_pdl_dec_refc">driver_pdl_dec_refc()</seealso>.</p>
      </item>

      <tag><marker id="ErlDrvTid"/>ErlDrvTid</tag>
      <item>
        <p>Thread identifier.</p>
	<p>See also:
	   <seealso marker="#erl_drv_thread_create">erl_drv_thread_create()</seealso>,
	   <seealso marker="#erl_drv_thread_exit">erl_drv_thread_exit()</seealso>,
	   <seealso marker="#erl_drv_thread_join">erl_drv_thread_join()</seealso>,
	   <seealso marker="#erl_drv_thread_self">erl_drv_thread_self()</seealso>,
	   and
	   <seealso marker="#erl_drv_equal_tids">erl_drv_equal_tids()</seealso>.
	</p>
      </item>
      <tag><marker id="ErlDrvThreadOpts"/>ErlDrvThreadOpts</tag>
      <item>
      <p/>
      <code type="none">
	int suggested_stack_size;
      </code>
        <p>Thread options structure passed to
	   <seealso marker="#erl_drv_thread_create">erl_drv_thread_create()</seealso>.
	   Currently the following fields exist:
	</p>
        <taglist>
          <tag>suggested_stack_size</tag>
          <item>A suggestion, in kilo-words, on how large stack to use. A value less
	        than zero means default size.
	  </item>
        </taglist>
	<p>See also:
	   <seealso marker="#erl_drv_thread_opts_create">erl_drv_thread_opts_create()</seealso>,
	   <seealso marker="#erl_drv_thread_opts_destroy">erl_drv_thread_opts_destroy()</seealso>,
	   and
	   <seealso marker="#erl_drv_thread_create">erl_drv_thread_create()</seealso>.
	</p>
      </item>

      <tag><marker id="ErlDrvMutex"/>ErlDrvMutex</tag>
      <item>
        <p>Mutual exclusion lock. Used for synchronizing access to shared data.
	   Only one thread at a time can lock a mutex.
	</p>
	<p>See also:
	   <seealso marker="#erl_drv_mutex_create">erl_drv_mutex_create()</seealso>,
	   <seealso marker="#erl_drv_mutex_destroy">erl_drv_mutex_destroy()</seealso>,
	   <seealso marker="#erl_drv_mutex_lock">erl_drv_mutex_lock()</seealso>,
	   <seealso marker="#erl_drv_mutex_trylock">erl_drv_mutex_trylock()</seealso>,
	   and
	   <seealso marker="#erl_drv_mutex_unlock">erl_drv_mutex_unlock()</seealso>.
	</p>
      </item>
      <tag><marker id="ErlDrvCond"/>ErlDrvCond</tag>
      <item>
        <p>Condition variable. Used when threads need to wait for a specific
	   condition to appear before continuing execution. Condition variables
	   need to be used with associated mutexes.
	</p>
	<p>See also:
	   <seealso marker="#erl_drv_cond_create">erl_drv_cond_create()</seealso>,
	   <seealso marker="#erl_drv_cond_destroy">erl_drv_cond_destroy()</seealso>,
	   <seealso marker="#erl_drv_cond_signal">erl_drv_cond_signal()</seealso>,
	   <seealso marker="#erl_drv_cond_broadcast">erl_drv_cond_broadcast()</seealso>,
	   and
	   <seealso marker="#erl_drv_cond_wait">erl_drv_cond_wait()</seealso>.
	</p>
      </item>
      <tag><marker id="ErlDrvRWLock"/>ErlDrvRWLock</tag>
      <item>
        <p>Read/write lock. Used to allow multiple threads to read shared data
	   while only allowing one thread to write the same data. Multiple threads
	   can read lock an rwlock at the same time, while only one thread can
	   read/write lock an rwlock at a time.
	</p>
	<p>See also:
	   <seealso marker="#erl_drv_rwlock_create">erl_drv_rwlock_create()</seealso>,
	   <seealso marker="#erl_drv_rwlock_destroy">erl_drv_rwlock_destroy()</seealso>,
	   <seealso marker="#erl_drv_rwlock_rlock">erl_drv_rwlock_rlock()</seealso>,
	   <seealso marker="#erl_drv_rwlock_tryrlock">erl_drv_rwlock_tryrlock()</seealso>,
	   <seealso marker="#erl_drv_rwlock_runlock">erl_drv_rwlock_runlock()</seealso>,
	   <seealso marker="#erl_drv_rwlock_rwlock">erl_drv_rwlock_rwlock()</seealso>,
	   <seealso marker="#erl_drv_rwlock_tryrwlock">erl_drv_rwlock_tryrwlock()</seealso>,
	   and
	   <seealso marker="#erl_drv_rwlock_rwunlock">erl_drv_rwlock_rwunlock()</seealso>.
	</p>
      </item>
      <tag><marker id="ErlDrvTSDKey"/>ErlDrvTSDKey</tag>
      <item>
        <p>Key which thread specific data can be associated with.</p>
	<p>See also:
	   <seealso marker="#erl_drv_tsd_key_create">erl_drv_tsd_key_create()</seealso>,
	   <seealso marker="#erl_drv_tsd_key_destroy">erl_drv_tsd_key_destroy()</seealso>,
	   <seealso marker="#erl_drv_tsd_set">erl_drv_tsd_set()</seealso>,
	   and
	   <seealso marker="#erl_drv_tsd_get">erl_drv_tsd_get()</seealso>.
	</p>
      </item>
     </taglist>
   </section>

  <funcs>
    <func>
      <name><ret>void</ret><nametext>driver_system_info(ErlDrvSysInfo *sys_info_ptr, size_t size)</nametext></name>
      <fsummary>Get information about the Erlang runtime system</fsummary>
      <desc>
        <marker id="driver_system_info"></marker>
        <p>This function will write information about the Erlang runtime
          system into the
          <seealso marker="#ErlDrvSysInfo">ErlDrvSysInfo</seealso>
          structure referred to by the first argument. The second
          argument should be the size of the
          <seealso marker="#ErlDrvSysInfo">ErlDrvSysInfo</seealso>
          structure, i.e., <c>sizeof(ErlDrvSysInfo)</c>.</p>
        <p>See the documentation of the
          <seealso marker="#ErlDrvSysInfo">ErlDrvSysInfo</seealso>
          structure for information about specific fields.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_output(ErlDrvPort port, char *buf, int len)</nametext></name>
      <fsummary>Send data from driver to port owner</fsummary>
      <desc>
        <marker id="driver_output"></marker>
        <p>The <c>driver_output</c> function is used to send data from
          the driver up to the emulator. The data will be received as
          terms or binary data, depending on how the driver port was
          opened.</p>
        <p>The data is queued in the port owner process' message
          queue. Note that this does not yield to the emulator. (Since
          the driver and the emulator runs in the same thread.)</p>
        <p>The parameter <c>buf</c> points to the data to send, and
          <c>len</c> is the number of bytes.</p>
        <p>The return value for all output functions is 0. (Unless the
          driver is used for distribution, in which case it can fail
          and return -1. For normal use, the output function always
          returns 0.)</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_output2(ErlDrvPort port, char *hbuf, int hlen, char *buf, int len)</nametext></name>
      <fsummary>Send data and binary data to port owner</fsummary>
      <desc>
        <marker id="driver_output2"></marker>
        <p>The <c>driver_output2</c> function first sends <c>hbuf</c>
          (length in <c>hlen</c>) data as a list, regardless of port
          settings. Then <c>buf</c> is sent as a binary or list.
          E.g. if <c>hlen</c> is 3 then the port owner process will
          receive <c>[H1, H2, H3 | T]</c>.</p>
        <p>The point of sending data as a list header, is to facilitate
          matching on the data received.</p>
        <p>The return value is 0 for normal use.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_output_binary(ErlDrvPort port, char *hbuf, int hlen, ErlDrvBinary* bin, int offset, int len)</nametext></name>
      <fsummary>Send data from a driver binary to port owner</fsummary>
      <desc>
        <marker id="driver_output_binary"></marker>
        <p>This function sends data to port owner process from a
          driver binary, it has a header buffer (<c>hbuf</c>
          and <c>hlen</c>) just like <c>driver_output2</c>. The
          <c>hbuf</c> parameter can be <c>NULL</c>.</p>
        <p>The parameter <c>offset</c> is an offset into the binary and
          <c>len</c> is the number of bytes to send.</p>
        <p>Driver binaries are created with <c>driver_alloc_binary</c>.</p>
        <p>The data in the header is sent as a list and the binary as
          an Erlang binary in the tail of the list.</p>
        <p>E.g. if <c>hlen</c> is 2, then the port owner process will
          receive <c><![CDATA[[H1, H2 | <<T>>]]]></c>.</p>
        <p>The return value is 0 for normal use.</p>
        <p>Note that, using the binary syntax in Erlang, the driver
          application can match the header directly from the binary,
          so the header can be put in the binary, and hlen can be set
          to 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_outputv(ErlDrvPort port, char* hbuf, int hlen,  ErlIOVec *ev, int skip)</nametext></name>
      <fsummary>Send vectorized data to port owner</fsummary>
      <desc>
        <marker id="driver_outputv"></marker>
        <p>This function sends data from an IO vector, <c>ev</c>, to
          the port owner process. It has a header buffer (<c>hbuf</c>
          and <c>hlen</c>), just like <c>driver_output2</c>.</p>
        <p>The <c>skip</c> parameter is a number of bytes to skip of
          the <c>ev</c> vector from the head.</p>
        <p>You get vectors of <c>ErlIOVec</c> type from the driver
          queue (see below), and the <seealso marker="driver_entry#outputv">outputv</seealso> driver entry
          function. You can also make them yourself, if you want to
          send several <c>ErlDrvBinary</c> buffers at once. Often
          it is faster to use <c>driver_output</c> or
          <c>driver_output_binary</c>.</p>
        <p>E.g. if <c>hlen</c> is 2 and <c>ev</c> points to an array of
          three binaries, the port owner process will receive <c><![CDATA[[H1, H2, <<B1>>, <<B2>> | <<B3>>]]]></c>.</p>
        <p>The return value is 0 for normal use.</p>
        <p>The comment for <c>driver_output_binary</c> applies for
          <c>driver_outputv</c> too.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_vec_to_buf(ErlIOVec *ev, char *buf, int len)</nametext></name>
      <fsummary>Collect data segments into a buffer</fsummary>
      <desc>
        <marker id="driver_vec_to_buf"></marker>
        <p>This function collects several segments of data, referenced
          by <c>ev</c>, by copying them in order to the buffer
          <c>buf</c>, of the size <c>len</c>.</p>
        <p>If the data is to be sent from the driver to the port owner
          process, it is faster to use <c>driver_outputv</c>.</p>
        <p>The return value is the space left in the buffer, i.e. if
          the <c>ev</c> contains less than <c>len</c> bytes it's the
          difference, and if <c>ev</c> contains <c>len</c> bytes or
          more, it's 0. This is faster if there is more than one header byte,
          since the binary syntax can construct integers directly from
          the binary.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_set_timer(ErlDrvPort port, unsigned long time)</nametext></name>
      <fsummary>Set a timer to call the driver</fsummary>
      <desc>
        <marker id="driver_set_timer"></marker>
        <p>This function sets a timer on the driver, which will count
          down and call the driver when it is timed out. The
          <c>time</c> parameter is the time in milliseconds before the
          timer expires.</p>
        <p>When the timer reaches 0 and expires, the driver entry
          function <seealso marker="driver_entry#emulator">timeout</seealso> is called.</p>
        <p>Note that there is only one timer on each driver instance;
          setting a new timer will replace an older one.</p>
        <p>Return value i 0 (-1 only when the <c>timeout</c> driver
          function is <c>NULL</c>).</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_cancel_timer(ErlDrvPort port)</nametext></name>
      <fsummary>Cancel a previously set timer</fsummary>
      <desc>
        <marker id="driver_cancel_timer"></marker>
        <p>This function cancels a timer set with
          <c>driver_set_timer</c>.</p>
        <p>The return value is 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_read_timer(ErlDrvPort port, unsigned long *time_left)</nametext></name>
      <fsummary>Read the time left before timeout</fsummary>
      <desc>
        <marker id="driver_read_timer"></marker>
        <p>This function reads the current time of a timer, and places
          the result in <c>time_left</c>. This is the time in
          milliseconds, before the timeout will occur.</p>
        <p>The return value is 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_get_now(ErlDrvNowData *now)</nametext></name>
      <fsummary>Read a system timestamp</fsummary>
      <desc>
        <marker id="driver_get_now"></marker>
        <p>This function reads a timestamp into the memory pointed to by 
          the parameter <c>now</c>. See the description of <seealso marker="#ErlDrvNowData">ErlDrvNowData</seealso> for
          specification of it's fields.  </p>
        <p>The return value is 0 unless the <c>now</c> pointer is not
          valid, in which case it is &lt; 0.  </p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_select(ErlDrvPort port, ErlDrvEvent event, int mode, int on)</nametext></name>
      <fsummary>Provide an event for having the emulator call the driver</fsummary>
      <desc>
        <marker id="driver_select"></marker>
        <p>This function is used by drivers to provide the emulator with
          events to check for. This enables the emulator to call the driver
          when something has happened asynchronously.</p>
        <p>The <c>event</c> argument identifies an OS-specific event object.
          On Unix systems, the functions <c>select</c>/<c>poll</c> are used. The
          event object must be a socket or pipe (or other object that
          <c>select</c>/<c>poll</c> can use).
          On windows, the Win32 API function <c>WaitForMultipleObjects</c>
          is used. This places other restriction on the event object.
          Refer to the Win32 SDK documentation.</p>
        <p>The <c>on</c> parameter should be <c>1</c> for setting events
          and <c>0</c> for clearing them.</p>
        <p>The <c>mode</c> argument is bitwise-or combination of
          <c>ERL_DRV_READ</c>, <c>ERL_DRV_WRITE</c> and <c>ERL_DRV_USE</c>.
          The first two specifies whether to wait for read events and/or write
          events. A fired read event will call
          <seealso marker="driver_entry#ready_input">ready_input</seealso>
          while a fired write event will call
          <seealso marker="driver_entry#ready_output">ready_output</seealso>.
          </p>
    <note>
      <p>Some OS (Windows) does not differ between read and write events.
         The call-back for a fired event then only depends on the value of <c>mode</c>.</p>
    </note>
        <p><c>ERL_DRV_USE</c> specifies if we are using the event object or if we want to close it.
           On an emulator with SMP support, it is not safe to clear all events
           and then close the event object after <c>driver_select</c> has
           returned. Another thread may still be using the event object
           internally. To safely close an event object call
           <c>driver_select</c> with <c>ERL_DRV_USE</c> and <c>on==0</c>. That
           will clear all events and then call 
           <seealso marker="driver_entry#stop_select">stop_select</seealso>
           when it is safe to close the event object.
           <c>ERL_DRV_USE</c> should be set together with the first event
           for an event object. It is harmless to set <c>ERL_DRV_USE</c>
           even though it already has been done. Clearing all events but keeping
           <c>ERL_DRV_USE</c> set will indicate that we are using the event
           object and probably will set events for it again.</p>
    <note>
      <p>ERL_DRV_USE was added in OTP release R13. Old drivers will still work
         as before. But it is recommended to update them to use <c>ERL_DRV_USE</c> and
         <c>stop_select</c> to make sure that event objects are closed in a safe way.</p>
    </note>           
        <p>The return value is 0 (Failure, -1, only if the
          <c>ready_input</c>/<c>ready_output</c> is
          <c>NULL</c>.</p>
      </desc>
    </func>
    <func>
      <name><ret>void *</ret><nametext>driver_alloc(size_t size)</nametext></name>
      <fsummary>Allocate memory</fsummary>
      <desc>
        <marker id="driver_alloc"></marker>
        <p>This function allocates a memory block of the size specified
          in <c>size</c>, and returns it. This only fails on out of
          memory, in that case <c>NULL</c> is returned. (This is most
          often a wrapper for <c>malloc</c>).</p>
        <p>Memory allocated must be explicitly freed with a corresponding
           call to <c>driver_free</c> (unless otherwise stated).</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>void *</ret><nametext>driver_realloc(void *ptr, size_t size)</nametext></name>
      <fsummary>Resize an allocated memory block</fsummary>
      <desc>
        <marker id="driver_realloc"></marker>
        <p>This function resizes a memory block, either in place, or by
          allocating a new block, copying the data and freeing the old
          block. A pointer is returned to the reallocated memory. On
          failure (out of memory), <c>NULL</c> is returned. (This is
          most often a wrapper for <c>realloc</c>.)</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>void</ret><nametext>driver_free(void *ptr)</nametext></name>
      <fsummary>Free an allocated memory block</fsummary>
      <desc>
        <marker id="driver_free"></marker>
        <p>This function frees the memory pointed to by <c>ptr</c>. The
          memory should have been allocated with
          <c>driver_alloc</c>. All allocated memory should be
          deallocated, just once. There is no garbage collection in
          drivers.</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvBinary*</ret><nametext>driver_alloc_binary(int size)</nametext></name>
      <fsummary>Allocate a driver binary</fsummary>
      <desc>
        <marker id="driver_alloc_binary"></marker>
        <p>This function allocates a driver binary with a memory block
          of at least <c>size</c> bytes, and returns a pointer to it,
          or NULL on failure (out of memory). When a driver binary has
          been sent to the emulator, it must not be altered. Every
          allocated binary should be freed by a corresponding call to
          <c>driver_free_binary</c> (unless otherwise stated).</p>
        <p>Note that a driver binary has an internal reference counter,
          this means that calling <c>driver_free_binary</c> it may not
          actually dispose of it. If it's sent to the emulator, it may
          be referenced there.</p>
        <p>The driver binary has a field, <c>orig_bytes</c>, which
          marks the start of the data in the binary.</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvBinary*</ret><nametext>driver_realloc_binary(ErlDrvBinary *bin, int size)</nametext></name>
      <fsummary>Resize a driver binary</fsummary>
      <desc>
        <marker id="driver_realloc_binary"></marker>
        <p>This function resizes a driver binary, while keeping the
          data. The resized driver binary is returned. On failure (out
          of memory), <c>NULL</c> is returned.</p>
        <p>This function is only thread-safe when the emulator with SMP
          support is used.</p>
      </desc>
    </func>
    <func>
      <name><ret>void</ret><nametext>driver_free_binary(ErlDrvBinary *bin)</nametext></name>
      <fsummary>Free a driver binary</fsummary>
      <desc>
        <marker id="driver_free_binary"></marker>
        <p>This function frees a driver binary <c>bin</c>, allocated
          previously with <c>driver_alloc_binary</c>. Since binaries
          in Erlang are reference counted, the binary may still be
          around.</p>
        <p>This function is only thread-safe when the emulator with SMP
          support is used.</p>
      </desc>
    </func>
    <func>
      <name><ret>long</ret><nametext>driver_binary_get_refc(ErlDrvBinary *bin)</nametext></name>
      <fsummary>Get the reference count of a driver binary</fsummary>
      <desc>
        <marker id="driver_binary_get_refc"></marker>
        <p>Returns current reference count on <c>bin</c>.</p>
        <p>This function is only thread-safe when the emulator with SMP
          support is used.</p>
      </desc>
    </func>
    <func>
      <name><ret>long</ret><nametext>driver_binary_inc_refc(ErlDrvBinary *bin)</nametext></name>
      <fsummary>Increment the reference count of a driver binary</fsummary>
      <desc>
        <marker id="driver_binary_inc_refc"></marker>
        <p>Increments the reference count on <c>bin</c> and returns
          the reference count reached after the increment.</p>
        <p>This function is only thread-safe when the emulator with SMP
          support is used.</p>
      </desc>
    </func>
    <func>
      <name><ret>long</ret><nametext>driver_binary_dec_refc(ErlDrvBinary *bin)</nametext></name>
      <fsummary>Decrement the reference count of a driver binary</fsummary>
      <desc>
        <marker id="driver_binary_dec_refc"></marker>
        <p>Decrements the reference count on <c>bin</c> and returns
          the reference count reached after the decrement.</p>
        <p>This function is only thread-safe when the emulator with SMP
          support is used.</p>
        <note>
          <p>You should normally decrement the reference count of a
            driver binary by calling
            <seealso marker="#driver_free_binary">driver_free_binary()</seealso>.
            <c>driver_binary_dec_refc()</c> does <em>not</em> free
            the binary if the reference count reaches zero. <em>Only</em>
            use <c>driver_binary_dec_refc()</c> when you are sure
            <em>not</em> to reach a reference count of zero.</p>
        </note>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_enq(ErlDrvPort port, char* buf, int len)</nametext></name>
      <fsummary>Enqueue data in the driver queue</fsummary>
      <desc>
        <marker id="driver_enq"></marker>
        <p>This function enqueues data in the driver queue. The data in
          <c>buf</c> is copied (<c>len</c> bytes) and placed at the
          end of the driver queue. The driver queue is normally used
          in a FIFO way.</p>
        <p>The driver queue is available to queue output from the
          emulator to the driver (data from the driver to the emulator
          is queued by the emulator in normal erlang message
          queues). This can be useful if the driver has to wait for
          slow devices etc, and wants to yield back to the
          emulator. The driver queue is implemented as an ErlIOVec.</p>
        <p>When the queue contains data, the driver won't close, until
          the queue is empty.</p>
        <p>The return value is 0.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_pushq(ErlDrvPort port, char* buf, int len)</nametext></name>
      <fsummary>Push data at the head of the driver queue</fsummary>
      <desc>
        <marker id="driver_pushq"></marker>
        <p>This function puts data at the head of the driver queue. The
          data in <c>buf</c> is copied (<c>len</c> bytes) and placed
          at the beginning of the queue.</p>
        <p>The return value is 0.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_deq(ErlDrvPort port, int size)</nametext></name>
      <fsummary>Dequeue data from the head of the driver queue</fsummary>
      <desc>
        <marker id="driver_deq"></marker>
        <p>This function dequeues data by moving the head pointer
          forward in the driver queue by <c>size</c> bytes. The data
          in the queue will be deallocated.</p>
        <p>The return value is the number of bytes remaining in the queue
          or -1 on failure.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_sizeq(ErlDrvPort port)</nametext></name>
      <fsummary>Return the size of the driver queue</fsummary>
      <desc>
        <marker id="driver_sizeq"></marker>
        <p>This function returns the number of bytes currently in the
          driver queue.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_enq_bin(ErlDrvPort port, ErlDrvBinary *bin, int offset,  int len)</nametext></name>
      <fsummary>Enqueue binary in the driver queue</fsummary>
      <desc>
        <marker id="driver_enq_bin"></marker>
        <p>This function enqueues a driver binary in the driver
          queue. The data in <c>bin</c> at <c>offset</c> with length
          <c>len</c> is placed at the end of the queue. This function
          is most often faster than <c>driver_enq</c>, because the
          data doesn't have to be copied.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
        <p>The return value is 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_pushq_bin(ErlDrvPort port, ErlDrvBinary *bin, int offset, int len)</nametext></name>
      <fsummary>Push binary at the head of the driver queue</fsummary>
      <desc>
        <marker id="driver_pushq_bin"></marker>
        <p>This function puts data in the binary <c>bin</c>, at
          <c>offset</c> with length <c>len</c> at the head of the
          driver queue. It is most often faster than
          <c>driver_pushq</c>, because the data doesn't have to be
          copied.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
        <p>The return value is 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>SysIOVec*</ret><nametext>driver_peekq(ErlDrvPort port, int *vlen)</nametext></name>
      <fsummary>Get the driver queue as a vector</fsummary>
      <desc>
        <marker id="driver_peekq"></marker>
        <p>This function retrieves the driver queue as a pointer to an
          array of <c>SysIOVec</c>s. It also returns the number of
          elements in <c>vlen</c>. This is the only way to get data
          out of the queue.</p>
        <p>Nothing is remove from the queue by this function, that must be done
          with <c>driver_deq</c>.</p>
        <p>The returned array is suitable to use with the Unix system
          call <c>writev</c>.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_enqv(ErlDrvPort port, ErlIOVec *ev, int skip)</nametext></name>
      <fsummary>Enqueue vector in the driver queue</fsummary>
      <desc>
        <marker id="driver_enqv"></marker>
        <p>This function enqueues the data in <c>ev</c>, skipping the
          first <c>skip</c> bytes of it, at the end of the driver
          queue. It is faster than <c>driver_enq</c>, because the data
          doesn't have to be copied.</p>
        <p>The return value is 0.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_pushqv(ErlDrvPort port, ErlIOVec *ev, int skip)</nametext></name>
      <fsummary>Push vector at the head of the driver queue</fsummary>
      <desc>
        <marker id="driver_pushqv"></marker>
        <p>This function puts the data in <c>ev</c>, skipping the first
          <c>skip</c> bytes of it, at the head of the driver queue.
          It is faster than <c>driver_pushq</c>, because the data
          doesn't have to be copied.</p>
        <p>The return value is 0.</p>
        <p>This function can be called from an arbitrary thread if a
          <seealso marker="#ErlDrvPDL">port data lock</seealso>
          associated with the <c>port</c> is locked by the calling
          thread during the call.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvPDL</ret><nametext>driver_pdl_create(ErlDrvPort port)</nametext></name>
      <fsummary>Create a port data lock</fsummary>
      <desc>
        <marker id="driver_pdl_create"></marker>
        <p>This function creates a port data lock associated with
          the <c>port</c>. <em>NOTE</em>: Once a port data lock has
          been created, it has to be locked during all operations
          on the driver queue of the <c>port</c>.</p>
        <p>On success a newly created port data lock is returned. On
          failure <c>NULL</c> is returned. <c>driver_pdl_create()</c> will
          fail if <c>port</c> is invalid or if a port data lock already has
          been associated with the <c>port</c>.</p>
      </desc>
    </func>
    <func>
      <name><ret>void</ret><nametext>driver_pdl_lock(ErlDrvPDL pdl)</nametext></name>
      <fsummary>Lock port data lock</fsummary>
      <desc>
        <marker id="driver_pdl_lock"></marker>
        <p>This function locks the port data lock passed as argument
          (<c>pdl</c>).</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>void</ret><nametext>driver_pdl_unlock(ErlDrvPDL pdl)</nametext></name>
      <fsummary>Unlock port data lock</fsummary>
      <desc>
        <marker id="driver_pdl_unlock"></marker>
        <p>This function unlocks the port data lock passed as argument
          (<c>pdl</c>).</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>long</ret><nametext>driver_pdl_get_refc(ErlDrvPDL pdl)</nametext></name>
      <fsummary></fsummary>
      <desc>
        <marker id="driver_pdl_get_refc"></marker>
        <p>This function returns the current reference count of
          the port data lock passed as argument (<c>pdl</c>).</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>long</ret><nametext>driver_pdl_inc_refc(ErlDrvPDL pdl)</nametext></name>
      <fsummary></fsummary>
      <desc>
        <marker id="driver_pdl_inc_refc"></marker>
        <p>This function increments the reference count of
          the port data lock passed as argument (<c>pdl</c>).</p>
        <p>The current reference count after the increment has
          been performed is returned.</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>long</ret><nametext>driver_pdl_dec_refc(ErlDrvPDL pdl)</nametext></name>
      <fsummary></fsummary>
      <desc>
        <marker id="driver_pdl_dec_refc"></marker>
        <p>This function decrements the reference count of
          the port data lock passed as argument (<c>pdl</c>).</p>
        <p>The current reference count after the decrement has
          been performed is returned.</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_monitor_process(ErlDrvPort port,  ErlDrvTermData process,  ErlDrvMonitor *monitor)</nametext></name>
      <fsummary>Monitor a process from a driver</fsummary>
      <desc>
        <marker id="driver_monitor_process"></marker>
        <p>Start monitoring a process from a driver. When a process is
          monitored, a process exit will result in a call to the
          provided <seealso marker="driver_entry#process_exit">process_exit</seealso> call-back
          in the <seealso marker="driver_entry">ErlDrvEntry</seealso>
          structure. The <c>ErlDrvMonitor</c> structure is filled in, for later
          removal or compare.</p>
        <p>The <c>process</c> parameter should be the return value of an
          earlier call to <seealso marker="#driver_caller">driver_caller</seealso> or <seealso marker="#driver_connected">driver_connected</seealso> call.</p>
        <p>The function returns 0 on success, &lt; 0 if no call-back is
          provided and &gt; 0 if the process is no longer alive.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_demonitor_process(ErlDrvPort port,  const ErlDrvMonitor *monitor)</nametext></name>
      <fsummary>Stop monitoring a process from a driver</fsummary>
      <desc>
        <marker id="driver_demonitor_process"></marker>
        <p>This function cancels an monitor created earlier. </p>
        <p>The function returns 0 if a monitor was removed and &gt; 0
          if the monitor did no longer exist.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvTermData</ret><nametext>driver_get_monitored_process(ErlDrvPort port,  const ErlDrvMonitor *monitor)</nametext></name>
      <fsummary>Retrieve the process id from a monitor</fsummary>
      <desc>
        <marker id="driver_get_monitored_process"></marker>
        <p>The function returns the process id associated with a living
          monitor. It can be used in the <c>process_exit</c> call-back to
          get the process  identification for the exiting process.</p>
        <p>The function returns <c>driver_term_nil</c> if the monitor
          no longer exists.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_compare_monitors(const ErlDrvMonitor *monitor1, const ErlDrvMonitor *monitor2)</nametext></name>
      <fsummary>Compare two monitors</fsummary>
      <desc>
        <marker id="driver_compare_monitors"></marker>
        <p>This function is used to compare two <c>ErlDrvMonitor</c>s. It
          can also be used to imply some artificial order on monitors,
          for whatever reason.</p>
        <p>The function returns 0 if <c>monitor1</c> and
          <c>monitor2</c> are equal, &lt; 0 if <c>monitor1</c> is less
          than <c>monitor2</c> and &gt; 0 if <c>monitor1</c> is greater
          than <c>monitor2</c>.</p>
      </desc>
    </func>
    <func>
      <name><ret>void</ret><nametext>add_driver_entry(ErlDrvEntry *de)</nametext></name>
      <fsummary>Add a driver entry</fsummary>
      <desc>
        <marker id="add_driver_entry"></marker>
        <p>This function adds a driver entry to the list of drivers
          known by Erlang. The <seealso marker="driver_entry#init">init</seealso> function of the <c>de</c>
          parameter is called.</p>
        <note>
          <p>To use this function for adding drivers residing in
            dynamically loaded code is dangerous. If the driver code
            for the added driver resides in the same dynamically
            loaded module (i.e. <c>.so</c> file) as a normal
            dynamically loaded driver (loaded with the <c>erl_ddll</c>
            interface), the caller should call <seealso marker="#driver_lock_driver">driver_lock_driver</seealso> before
            adding driver entries.</p>
          <p>Use of this function is generally deprecated.</p>
        </note>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>remove_driver_entry(ErlDrvEntry *de)</nametext></name>
      <fsummary>Remove a driver entry</fsummary>
      <desc>
        <marker id="remove_driver_entry"></marker>
        <p>This function removes a driver entry <c>de</c> previously
          added with <c>add_driver_entry</c>.</p>
        <p>Driver entries added by the <c>erl_ddll</c> erlang interface can 
          not be removed by using this interface.</p>
      </desc>
    </func>
    <func>
      <name><ret>char*</ret><nametext>erl_errno_id(int error)</nametext></name>
      <fsummary>Get erlang error atom name from error number</fsummary>
      <desc>
        <marker id="erl_errno_id"></marker>
        <p>This function returns the atom name of the erlang error,
          given the error number in <c>error</c>. Error atoms are:
          <c>einval</c>, <c>enoent</c>, etc. It can be used to make
          error terms from the driver.</p>
      </desc>
    </func>
    <func>
      <name><ret>void</ret><nametext>set_busy_port(ErlDrvPort port, int on)</nametext></name>
      <fsummary>Signal or unsignal port as busy</fsummary>
      <desc>
        <marker id="set_busy_port"></marker>
        <p>This function set and resets the busy status of the port. If
          <c>on</c> is 1, the port is set to busy, if it's 0 the port
          is set to not busy.</p>
        <p>When the port is busy, sending to it with <c>Port ! Data</c>
          or <c>port_command/2</c>, will block the port owner process,
          until the port is signaled as not busy.</p>
	<p>If the
	   <seealso marker="driver_entry#driver_flags"><![CDATA[ERL_DRV_FLAG_SOFT_BUSY]]></seealso>
	   has been set in the
	   <seealso marker="driver_entry">driver_entry</seealso>,
	   data can be forced into the driver via
	   <seealso marker="erlang#port_command/3">port_command(Port, Data, [force])</seealso>
	   even though the driver has signaled that it is busy.
	</p>
      </desc>
    </func>
    <func>
      <name><ret>void</ret><nametext>set_port_control_flags(ErlDrvPort port, int flags)</nametext></name>
      <fsummary>Set flags on how to handle control entry function</fsummary>
      <desc>
        <marker id="set_port_control_flags"></marker>
        <p>This function sets flags for how the <seealso marker="driver_entry#control">control</seealso> driver entry
          function will return data to the port owner process. (The
          <c>control</c> function is called from <c>port_control/3</c>
          in erlang.)</p>
        <p>Currently there are only two meaningful values for
          <c>flags</c>: 0 means that data is returned in a list, and
          <c>PORT_CONTROL_FLAG_BINARY</c> means data is returned as
          a binary from <c>control</c>.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_failure_eof(ErlDrvPort port)</nametext></name>
      <fsummary>Fail with EOF</fsummary>
      <desc>
        <marker id="driver_failure_eof"></marker>
        <p>This function signals to erlang that the driver has
          encountered an EOF and should be closed, unless the port was
          opened with the <c>eof</c> option, in that case eof is sent
          to the port. Otherwise, the port is close and an
          <c>'EXIT'</c> message is sent to the port owner process.</p>
        <p>The return value is 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_failure_atom(ErlDrvPort port, char *string)</nametext></name>
      <name><ret>int</ret><nametext>driver_failure_posix(ErlDrvPort port, int error)</nametext></name>
      <name><ret>int</ret><nametext>driver_failure(ErlDrvPort port, int error)</nametext></name>
      <fsummary>Fail with error</fsummary>
      <desc>
        <marker id="driver_failure_atom"></marker>
        <marker id="driver_failure_posix"></marker>
        <marker id="driver_failure"></marker>
        <p>These functions signal to Erlang that the driver has
          encountered an error and should be closed. The port is
          closed and the tuple <c>{'EXIT', error, Err}</c>, is sent to
          the port owner process, where error is an error atom
          (<c>driver_failure_atom</c> and
          <c>driver_failure_posix</c>), or an integer
          (<c>driver_failure</c>).</p>
        <p>The driver should fail only when in severe error situations,
          when the driver cannot possibly keep open, for instance
          buffer allocation gets out of memory. Normal errors is more
          appropriate to handle with sending error codes with
          <c>driver_output</c>.</p>
        <p>The return value is 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvTermData</ret><nametext>driver_connected(ErlDrvPort port)</nametext></name>
      <fsummary>Return the port owner process</fsummary>
      <desc>
        <marker id="driver_connected"></marker>
        <p>This function returns the port owner process.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvTermData</ret><nametext>driver_caller(ErlDrvPort port)</nametext></name>
      <fsummary>Return the process making the driver call</fsummary>
      <desc>
        <marker id="driver_caller"></marker>
        <p>This function returns the process id of the process that
	  made the current call to the driver. The process id can be
	  used with <c>driver_send_term</c> to send back data to the
	  caller. <c>driver_caller()</c> only return valid data
	  when currently executing in one of the following driver
	  callbacks:</p>
	  <taglist>
	    <tag><seealso marker="driver_entry#start">start</seealso></tag>
            <item>Called from <c>open_port/2</c>.</item>
	    <tag><seealso marker="driver_entry#output">output</seealso></tag>
            <item>Called from <c>erlang:send/2</c>, and
	          <c>erlang:port_command/2</c></item>
	    <tag><seealso marker="driver_entry#outputv">outputv</seealso></tag>
            <item>Called from <c>erlang:send/2</c>, and
	          <c>erlang:port_command/2</c></item>
	    <tag><seealso marker="driver_entry#control">control</seealso></tag>
            <item>Called from <c>erlang:port_control/3</c></item>
	    <tag><seealso marker="driver_entry#call">call</seealso></tag>
            <item>Called from <c>erlang:port_call/3</c></item>
	  </taglist>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_output_term(ErlDrvPort port, ErlDrvTermData* term, int n)</nametext></name>
      <fsummary>Send term data from driver to port owner</fsummary>
      <desc>
        <marker id="driver_output_term"></marker>
        <p>This functions sends data in the special driver term
          format. This is a fast way to deliver term data from a
          driver. It also needs no binary conversion, so the port
          owner process receives data as normal Erlang terms.</p>
        <p>The <c>term</c> parameter points to an array of
          <c>ErlDrvTermData</c>, with <c>n</c> elements. This array
          contains terms described in the driver term format. Every
          term consists of one to four elements in the array. The
          term first has a term type, and then arguments.</p>
        <p>Tuple and lists (with the exception of strings, see below),
          are built in reverse polish notation, so that to build a
          tuple, the elements are given first, and then the tuple
          term, with a count. Likewise for lists.</p>
        <p>A tuple must be specified with the number of elements. (The
          elements precedes the <c>ERL_DRV_TUPLE</c> term.)</p>
        <p>A list must be specified with the number of elements,
          including the tail, which is the last term preceding
          <c>ERL_DRV_LIST</c>.</p>
        <p>The special term <c>ERL_DRV_STRING_CONS</c> is used to
          "splice" in a string in a list, a string given this way is
          not a list per se, but the elements are elements of the
          surrounding list.</p>
        <pre>
Term type            Argument(s)
===========================================
ERL_DRV_NIL          
ERL_DRV_ATOM         ErlDrvTermData atom (from driver_mk_atom(char *string))
ERL_DRV_INT          ErlDrvSInt integer
ERL_DRV_UINT         ErlDrvUInt integer
ERL_DRV_INT64        ErlDrvSInt64 *integer_ptr
ERL_DRV_UINT64       ErlDrvUInt64 *integer_ptr
ERL_DRV_PORT         ErlDrvTermData port (from driver_mk_port(ErlDrvPort port))
ERL_DRV_BINARY       ErlDrvBinary *bin, ErlDrvUInt len, ErlDrvUInt offset
ERL_DRV_BUF2BINARY   char *buf, ErlDrvUInt len
ERL_DRV_STRING       char *str, int len
ERL_DRV_TUPLE        int sz
ERL_DRV_LIST         int sz
ERL_DRV_PID          ErlDrvTermData pid (from driver_connected(ErlDrvPort port) or driver_caller(ErlDrvPort port))
ERL_DRV_STRING_CONS  char *str, int len
ERL_DRV_FLOAT        double *dbl
ERL_DRV_EXT2TERM     char *buf, ErlDrvUInt len
        </pre>
	<p>The unsigned integer data type <c>ErlDrvUInt</c> and the
	  signed integer data type <c>ErlDrvSInt</c> are 64 bits wide
	  on a 64 bit runtime system and 32 bits wide on a 32 bit
	  runtime system. They were introduced in erts version 5.6,
	  and replaced some of the <c>int</c> arguments in the list above. 
	</p>
	<p>The unsigned integer data type <c>ErlDrvUInt64</c> and the
	  signed integer data type <c>ErlDrvSInt64</c> are always 64 bits
	  wide. They were introduced in erts version 5.7.4. 
	</p>

        <p>To build the tuple <c>{tcp, Port, [100 | Binary]}</c>, the
          following call could be made.</p>
        <code type="none"><![CDATA[
    ErlDrvBinary* bin = ...
    ErlDrvPort port = ...
    ErlDrvTermData spec[] = {
        ERL_DRV_ATOM, driver_mk_atom("tcp"),
        ERL_DRV_PORT, driver_mk_port(port),
            ERL_DRV_INT, 100,
            ERL_DRV_BINARY, bin, 50, 0,
            ERL_DRV_LIST, 2,
        ERL_DRV_TUPLE, 3,
    };
    driver_output_term(port, spec, sizeof(spec) / sizeof(spec[0]));
        ]]>
      </code>
        <p>Where <c>bin</c> is a driver binary of length at least 50
          and <c>port</c> is a port handle. Note that the <c>ERL_DRV_LIST</c>
          comes after the elements of the list, likewise the
          <c>ERL_DRV_TUPLE</c>.</p>
        <p>The term <c>ERL_DRV_STRING_CONS</c> is a way to construct
          strings. It works differently from how <c>ERL_DRV_STRING</c>
          works. <c>ERL_DRV_STRING_CONS</c> builds a string list in
          reverse order, (as opposed to how <c>ERL_DRV_LIST</c>
          works), concatenating the strings added to a list. The tail
          must be given before <c>ERL_DRV_STRING_CONS</c>.</p>
        <p>The <c>ERL_DRV_STRING</c> constructs a string, and ends
          it. (So it's the same as <c>ERL_DRV_NIL</c> followed by
          <c>ERL_DRV_STRING_CONS</c>.)</p>
        <code type="none"><![CDATA[
    /* to send [x, "abc", y] to the port: */
    ErlDrvTermData spec[] = {
        ERL_DRV_ATOM, driver_mk_atom("x"),
        ERL_DRV_STRING, (ErlDrvTermData)"abc", 3,
        ERL_DRV_ATOM, driver_mk_atom("y"),
        ERL_DRV_NIL,
        ERL_DRV_LIST, 4
    };
    driver_output_term(port, spec, sizeof(spec) / sizeof(spec[0]));
        ]]></code>
        <p></p>
        <code type="none"><![CDATA[
    /* to send "abc123" to the port: */
    ErlDrvTermData spec[] = {
        ERL_DRV_NIL,        /* with STRING_CONS, the tail comes first */
        ERL_DRV_STRING_CONS, (ErlDrvTermData)"123", 3,
        ERL_DRV_STRING_CONS, (ErlDrvTermData)"abc", 3,
    };
    driver_output_term(port, spec, sizeof(spec) / sizeof(spec[0]));
        ]]></code>
	<p>The <c>ERL_DRV_EXT2TERM</c> term type is used for passing a
	term encoded with the
	<seealso marker="erl_ext_dist">external format</seealso>,
	i.e., a term that has been encoded by
	<seealso marker="erts:erlang#term_to_binary/2">erlang:term_to_binary</seealso>,
	<seealso marker="erl_interface:ei">erl_interface</seealso>, etc.
	For example, if <c>binp</c> is a pointer to an <c>ErlDrvBinary</c>
	that contains the term <c>{17, 4711}</c> encoded with the
	<seealso marker="erl_ext_dist">external format</seealso>
	and you want to wrap it in a two tuple with the tag <c>my_tag</c>,
	i.e., <c>{my_tag, {17, 4711}}</c>, you can do as follows:
	</p>
        <code type="none"><![CDATA[
    ErlDrvTermData spec[] = {
            ERL_DRV_ATOM, driver_mk_atom("my_tag"),
            ERL_DRV_EXT2TERM, (ErlDrvTermData) binp->orig_bytes, binp->orig_size
        ERL_DRV_TUPLE, 2,
    };
    driver_output_term(port, spec, sizeof(spec) / sizeof(spec[0]));
        ]]></code>
	<p>If you want to pass a binary and doesn't already have the content
	of the binary in an <c>ErlDrvBinary</c>, you can benefit from using
	<c>ERL_DRV_BUF2BINARY</c> instead of creating an <c>ErlDrvBinary</c>
	via <c>driver_alloc_binary()</c> and then pass the binary via
	<c>ERL_DRV_BINARY</c>. The runtime system will often allocate
	binaries smarter if <c>ERL_DRV_BUF2BINARY</c> is used.
	However, if the content of the binary to pass already resides in
	an <c>ErlDrvBinary</c>, it is normally better to pass the binary
	using <c>ERL_DRV_BINARY</c> and the <c>ErlDrvBinary</c> in question.
	</p>
	<p>The <c>ERL_DRV_UINT</c>, <c>ERL_DRV_BUF2BINARY</c>, and
	<c>ERL_DRV_EXT2TERM</c> term types were introduced in the 5.6
	version of erts.
	</p>
        <p>Note that this function is <em>not</em> thread-safe, not
          even when the emulator with SMP support is used.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvTermData</ret><nametext>driver_mk_atom(char* string)</nametext></name>
      <fsummary>Make an atom from a name</fsummary>
      <desc>
        <marker id="driver_mk_atom"></marker>
        <p>This function returns an atom given a name
          <c>string</c>. The atom is created and won't change, so the
          return value may be saved and reused, which is faster than
          looking up the atom several times.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvTermData</ret><nametext>driver_mk_port(ErlDrvPort port)</nametext></name>
      <fsummary>Make a erlang term port from a port</fsummary>
      <desc>
        <marker id="driver_mk_port"></marker>
        <p>This function converts a port handle to the erlang term
          format, usable in the <c>driver_output_send</c> function.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_send_term(ErlDrvPort port, ErlDrvTermData receiver, ErlDrvTermData* term, int n)</nametext></name>
      <fsummary>Send term data to other process than port owner process</fsummary>
      <desc>
        <marker id="driver_send_term"></marker>
        <p>This function is the only way for a driver to send data to
          <em>other</em> processes than the port owner process. The
          <c>receiver</c> parameter specifies the process to receive
          the data.</p>
        <p>The parameters <c>term</c> and <c>n</c> does the same thing
          as in <seealso marker="#driver_output_term">driver_output_term</seealso>.</p>
        <p>This function is only thread-safe when the emulator with SMP
          support is used.</p>
      </desc>
    </func>
    <func>
      <name><ret>long</ret><nametext>driver_async (ErlDrvPort port, unsigned int* key, void (*async_invoke)(void*), void* async_data, void (*async_free)(void*))</nametext></name>
      <fsummary>Perform an asynchronous call within a driver</fsummary>
      <desc>
        <marker id="driver_async"></marker>
        <p>This function performs an asynchronous call. The function
          <c>async_invoke</c> is invoked in a thread separate from the
          emulator thread. This enables the driver to perform
          time-consuming, blocking operations without blocking the
          emulator.</p>
        <p>Erlang is by default started without an async thread pool. The
          number of async threads that the runtime system should use
          is specified by the
          <seealso marker="erl#async_thread_pool_size">+A</seealso>
          command line argument of <seealso marker="erl">erl(1)</seealso>.
          If no async thread pool is available, the call is made
          synchronously in the thread calling <c>driver_async()</c>. The
          current number of async threads in the async thread pool can be
          retrieved via
          <seealso marker="#driver_system_info">driver_system_info()</seealso>.</p>
        <p>If there is a thread pool available, a thread will be
          used. If the <c>key</c> argument is null, the threads from the
          pool are used in a round-robin way, each call to
          <c>driver_async</c> uses the next thread in the pool. With the
          <c>key</c> argument set, this behaviour is changed. The two
          same values of <c>*key</c> always get the same thread.</p>
        <p>To make sure that a driver instance always uses the same
          thread, the following call can be used:</p>
        <p></p>
        <code type="none"><![CDATA[
    unsigned int myKey = (unsigned int) myPort;

    r = driver_async(myPort, &myKey, myData, myFunc);
        ]]></code>
        <p>It is enough to initialize <c>myKey</c> once for each
        driver instance.</p>
        <p>If a thread is already working, the calls will be
          queued up and executed in order. Using the same thread for
          each driver instance ensures that the calls will be made in
          sequence.</p>
        <p>The <c>async_data</c> is the argument to the functions
          <c>async_invoke</c> and <c>async_free</c>. It's typically a
          pointer to a structure that contains a pipe or event that
          can be used to signal that the async operation completed.
          The data should be freed in <c>async_free</c>, because it's
          called if <c>driver_async_cancel</c> is called.</p>
        <p>When the async operation is done, <seealso marker="driver_entry#ready_async">ready_async</seealso> driver
          entry function is called. If <c>async_ready</c> is null in
          the driver entry, the <c>async_free</c> function is called
          instead.</p>
        <p>The return value is a handle to the asynchronous task, which
          can be used as argument to <c>driver_async_cancel</c>.</p>
        <note>
          <p>As of erts version 5.5.4.3 the default stack size for
            threads in the async-thread pool is 16 kilowords,
            i.e., 64 kilobyte on 32-bit architectures.
            This small default size has been chosen since the
            amount of async-threads might be quite large. The
            default stack size is enough for drivers delivered
            with Erlang/OTP, but might not be sufficiently large
            for other dynamically linked in drivers that use the
            driver_async() functionality. A suggested stack size
            for threads in the async-thread pool can be configured
            via the
            <seealso marker="erl#async_thread_stack_size">+a</seealso>
            command line argument of
            <seealso marker="erl">erl(1)</seealso>.</p>
        </note>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_async_cancel(long id)</nametext></name>
      <fsummary>Cancel an asynchronous call</fsummary>
      <desc>
        <marker id="driver_async_cancel"></marker>
        <p>This function cancels an asynchronous operation, by removing
          it from the queue. Only functions in the queue can be
          cancelled; if a function is executing, it's too late to
          cancel it. The <c>async_free</c> function is also called.</p>
        <p>The return value is 1 if the operation was removed from the
          queue, otherwise 0.</p>
      </desc>
    </func>
    <func>
      <name><ret>int</ret><nametext>driver_lock_driver(ErlDrvPort port)</nametext></name>
      <fsummary>Make sure the driver is never unloaded</fsummary>
      <desc>
        <marker id="driver_lock_driver"></marker>
        <p>This function locks the driver used by the port <c>port</c>
          in memory for the rest of the emulator process
          lifetime. After this call, the driver behaves as one of Erlang's 
          statically linked in drivers.</p>
      </desc>
    </func>
    <func>
      <name><ret>ErlDrvPort</ret><nametext>driver_create_port(ErlDrvPort port, ErlDrvTermData owner_pid, char* name, ErlDrvData drv_data)</nametext></name>
      <fsummary>Create a new port (driver instance)</fsummary>
      <desc>
        <p>This function creates a new port executing the same driver
          code as the port creating the new port.
          A short description of the arguments:</p>
        <taglist>
          <tag><c>port</c></tag>
          <item>The port handle of the port (driver instance) creating
           the new port.</item>
          <tag><c>owner_pid</c></tag>
          <item>The process id of the Erlang process which will be
           owner of the new port. This process will be linked
           to the new port. You usually want to use
          <c>driver_caller(port)</c> as <c>owner_pid</c>.</item>
          <tag><c>name</c></tag>
          <item>The port name of the new port. You usually want to
           use the same port name as the driver name
           (<seealso marker="driver_entry#driver_name">driver_name</seealso>
           field of the
          <seealso marker="driver_entry">driver_entry</seealso>).</item>
          <tag><c>drv_data</c></tag>
          <item>The driver defined handle that will be passed in subsequent
           calls to driver call-backs. Note, that the 
          <seealso marker="driver_entry#start">driver start call-back</seealso>
           will not be called for this new driver instance.
           The driver defined handle is normally created in the
          <seealso marker="driver_entry#start">driver start call-back</seealso>
           when a port is created via
          <seealso marker="erts:erlang#open_port/2">erlang:open_port/2</seealso>. </item>
        </taglist>
        <p>The caller of <c>driver_create_port()</c> is allowed to
          manipulate the newly created port when <c>driver_create_port()</c>
          has returned. When
          <seealso marker="#smp_support">port level locking</seealso>
          is used, the creating port is, however, only allowed to
          manipulate the newly created port until the current driver
          call-back that was called by the emulator returns.</p>
        <note>
          <p>When
            <seealso marker="#smp_support">port level locking</seealso>
            is used, the creating port is only allowed to manipulate
            the newly created port until the current driver call-back
            returns.</p>
        </note>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_thread_create(char *name,
				      ErlDrvTid *tid,
				      void * (*func)(void *),
				      void *arg,
				      ErlDrvThreadOpts *opts)</nametext></name>
      <fsummary>Create a thread</fsummary>
      <desc>
        <marker id="erl_drv_thread_create"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>name</c></tag>
          <item>A string identifying the created thread. It will be used
	        to identify the thread in planned future debug
		functionality.
	  </item>
          <tag><c>tid</c></tag>
          <item>A pointer to a thread identifier variable.</item>
          <tag><c>func</c></tag>
          <item>A pointer to a function to execute in the created thread.</item>
          <tag><c>arg</c></tag>
          <item>A pointer to argument to the <c>func</c> function.</item>
          <tag><c>opts</c></tag>
          <item>A pointer to thread options to use or <c>NULL</c>.</item>
        </taglist>
        <p>This function creates a new thread. On success <c>0</c> is returned;
	   otherwise, an <c>errno</c> value is returned to indicate the error.
	   The newly created thread will begin executing in the function pointed
	   to by <c>func</c>, and <c>func</c> will be passed <c>arg</c> as
	   argument. When <c>erl_drv_thread_create()</c> returns the thread
	   identifier of the newly created thread will be available in
	   <c>*tid</c>. <c>opts</c> can be either a <c>NULL</c> pointer, or a
	   pointer to an
	   <seealso marker="#ErlDrvThreadOpts">ErlDrvThreadOpts</seealso>
	   structure. If <c>opts</c> is a <c>NULL</c> pointer, default options
	   will be used; otherwise, the passed options will be used.
	</p>
	<warning><p>You are not allowed to allocate the
	            <seealso marker="#ErlDrvThreadOpts">ErlDrvThreadOpts</seealso>
	            structure by yourself. It has to be allocated and
		    initialized by
		    <seealso marker="#erl_drv_thread_opts_create">erl_drv_thread_opts_create()</seealso>.
	</p></warning>
	<p>The created thread will terminate either when <c>func</c> returns
	   or if
	   <seealso marker="#erl_drv_thread_exit">erl_drv_thread_exit()</seealso>
	   is called by the thread. The exit value of the thread is either
	   returned from <c>func</c> or passed as argument to
	   <seealso marker="#erl_drv_thread_exit">erl_drv_thread_exit()</seealso>.
	   The driver creating the thread has the responsibility of joining the
	   thread, via
	   <seealso marker="#erl_drv_thread_join">erl_drv_thread_join()</seealso>,
	   before the driver is unloaded. It is not possible to create
	   "detached" threads, i.e., threads that don't need to be joined.
	</p>
	<warning><p>All created threads need to be joined by the driver before
	            it is unloaded.  If the driver fails to join all threads
		    created before it is unloaded, the runtime system will
		    most likely crash when the code of the driver is unloaded.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>ErlDrvThreadOpts *</ret><nametext>erl_drv_thread_opts_create(char *name)</nametext></name>
      <fsummary>Create thread options</fsummary>
      <desc>
        <marker id="erl_drv_thread_opts_create"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>name</c></tag>
          <item>A string identifying the created thread options. It will be used
	        to identify the thread options in planned future debug
		functionality.
	  </item>
        </taglist>
        <p>This function allocates and initialize a thread option
	   structure. On failure <c>NULL</c> is returned. A thread option
	   structure is used for passing options to
	   <seealso marker="#erl_drv_thread_create">erl_drv_thread_create()</seealso>.
	   If the structure isn't modified before it is passed to
	   <seealso marker="#erl_drv_thread_create">erl_drv_thread_create()</seealso>,
	   the default values will be used.
	</p>
	<warning><p>You are not allowed to allocate the
	            <seealso marker="#ErlDrvThreadOpts">ErlDrvThreadOpts</seealso>
	            structure by yourself. It has to be allocated and
		    initialized by <c>erl_drv_thread_opts_create()</c>.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_thread_opts_destroy(ErlDrvThreadOpts *opts)</nametext></name>
      <fsummary>Destroy thread options</fsummary>
      <desc>
        <marker id="erl_drv_thread_opts_destroy"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>opts</c></tag>
          <item>A pointer to thread options to destroy.</item>
        </taglist>
        <p>This function destroys thread options previously created by
	   <seealso marker="#erl_drv_thread_opts_create">erl_drv_thread_opts_create()</seealso>.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_thread_exit(void *exit_value)</nametext></name>
      <fsummary>Terminate calling thread</fsummary>
      <desc>
        <marker id="erl_drv_thread_exit"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>exit_value</c></tag>
          <item>A pointer to an exit value or <c>NULL</c>.</item>
        </taglist>
        <p>This function terminates the calling thread with the exit
	   value passed as argument. You are only allowed to terminate
	   threads created with
	   <seealso marker="#erl_drv_thread_create">erl_drv_thread_create()</seealso>.
	   The exit value can later be retrieved by another thread via
	   <seealso marker="#erl_drv_thread_join">erl_drv_thread_join()</seealso>.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_thread_join(ErlDrvTid tid, void **exit_value)</nametext></name>
      <fsummary>Join with another thread</fsummary>
      <desc>
        <marker id="erl_drv_thread_join"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>tid</c></tag>
          <item>The thread identifier of the thread to join.</item>
          <tag><c>exit_value</c></tag>
          <item>A pointer to a pointer to an exit value, or <c>NULL</c>.</item>
        </taglist>
        <p>This function joins the calling thread with another thread, i.e.,
	   the calling thread is blocked until the thread identified by
	   <c>tid</c> has terminated. On success <c>0</c> is returned;
	   otherwise, an <c>errno</c> value is returned to indicate the error.
	   A thread can only be joined once. The behavior of joining
	   more than once is undefined, an emulator crash is likely. If
	   <c>exit_value == NULL</c>, the exit value of the terminated thread
	   will be ignored; otherwise, the exit value of the terminated thread
	   will be stored at <c>*exit_value</c>.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>ErlDrvTid</ret><nametext>erl_drv_thread_self(void)</nametext></name>
      <fsummary>Get the thread identifier of the current thread</fsummary>
      <desc>
        <marker id="erl_drv_thread_self"></marker>
        <p>This function returns the thread identifier of the
	   calling thread.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_equal_tids(ErlDrvTid tid1, ErlDrvTid tid2)</nametext></name>
      <fsummary>Compare thread identifiers for equality</fsummary>
      <desc>
        <marker id="erl_drv_equal_tids"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>tid1</c></tag>
          <item>A thread identifier.</item>
          <tag><c>tid2</c></tag>
          <item>A thread identifier.</item>
        </taglist>
        <p>This function compares two thread identifiers for equality,
	   and returns <c>0</c> it they aren't equal, and 
	   a value not equal to <c>0</c> if they are equal.</p>
	<note><p>A Thread identifier may be reused very quickly after
	         a thread has terminated. Therefore, if a thread
		 corresponding to one of the involved thread identifiers
		 has terminated since the thread identifier was saved,
		 the result of <c>erl_drv_equal_tids()</c> might not give
		 expected result.
	</p></note>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>ErlDrvMutex *</ret><nametext>erl_drv_mutex_create(char *name)</nametext></name>
      <fsummary>Create a mutex</fsummary>
      <desc>
        <marker id="erl_drv_mutex_create"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>name</c></tag>
          <item>A string identifying the created mutex. It will be used
	        to identify the mutex in planned future debug functionality.
	  </item>
        </taglist>
        <p>This function creates a mutex and returns a pointer to it. On
	   failure <c>NULL</c> is returned. The driver creating the mutex
	   has the responsibility of destroying it before the driver is
	   unloaded.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_mutex_destroy(ErlDrvMutex *mtx)</nametext></name>
      <fsummary>Destroy a mutex</fsummary>
      <desc>
        <marker id="erl_drv_mutex_destroy"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>mtx</c></tag>
          <item>A pointer to a mutex to destroy.</item>
        </taglist>
        <p>This function destroys a mutex previously created by
	   <seealso marker="#erl_drv_mutex_create">erl_drv_mutex_create()</seealso>.
	   The mutex has to be in an unlocked state before being
	   destroyed.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_mutex_lock(ErlDrvMutex *mtx)</nametext></name>
      <fsummary>Lock a mutex</fsummary>
      <desc>
        <marker id="erl_drv_mutex_lock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>mtx</c></tag>
          <item>A pointer to a mutex to lock.</item>
        </taglist>
        <p>This function locks a mutex. The calling thread will be
	   blocked until the mutex has been locked. A thread
	   which currently has locked the mutex may <em>not</em> lock
	   the same mutex again.
	</p>
	<warning><p>If you leave a mutex locked in an emulator thread
	            when you let the thread out of your control, you will
		    <em>very likely</em> deadlock the whole emulator.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_mutex_trylock(ErlDrvMutex *mtx)</nametext></name>
      <fsummary>Try lock a mutex</fsummary>
      <desc>
        <marker id="erl_drv_mutex_trylock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>mtx</c></tag>
          <item>A pointer to a mutex to try to lock.</item>
        </taglist>
        <p>This function tries to lock a mutex. If successful <c>0</c>,
	   is returned; otherwise, <c>EBUSY</c> is returned. A thread
	   which currently has locked the mutex may <em>not</em> try to
	   lock the same mutex again.
	</p>
	<warning><p>If you leave a mutex locked in an emulator thread
	            when you let the thread out of your control, you will
		    <em>very likely</em> deadlock the whole emulator.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_mutex_unlock(ErlDrvMutex *mtx)</nametext></name>
      <fsummary>Unlock a mutex</fsummary>
      <desc>
        <marker id="erl_drv_mutex_unlock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>mtx</c></tag>
          <item>A pointer to a mutex to unlock.</item>
        </taglist>
        <p>This function unlocks a mutex. The mutex currently has to be
	   locked by the calling thread.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>ErlDrvCond *</ret><nametext>erl_drv_cond_create(char *name)</nametext></name>
      <fsummary>Create a condition variable</fsummary>
      <desc>
        <marker id="erl_drv_cond_create"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>name</c></tag>
          <item>A string identifying the created condition variable. It
	        will be used to identify the condition variable in planned
		future debug functionality.
	  </item>
        </taglist>
        <p>This function creates a condition variable and returns a
	   pointer to it. On failure <c>NULL</c> is returned. The driver
	   creating the condition variable has the responsibility of
	   destroying it before the driver is unloaded.</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_cond_destroy(ErlDrvCond *cnd)</nametext></name>
      <fsummary>Destroy a condition variable</fsummary>
      <desc>
        <marker id="erl_drv_cond_destroy"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>cnd</c></tag>
          <item>A pointer to a condition variable to destroy.</item>
        </taglist>
        <p>This function destroys a condition variable previously
	   created by
	   <seealso marker="#erl_drv_cond_create">erl_drv_cond_create()</seealso>.
	</p>
      <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_cond_signal(ErlDrvCond *cnd)</nametext></name>
      <fsummary>Signal on a condition variable</fsummary>
      <desc>
        <marker id="erl_drv_cond_signal"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>cnd</c></tag>
          <item>A pointer to a condition variable to signal on.</item>
        </taglist>
        <p>This function signals on a condition variable. That is, if
	   other threads are waiting on the condition variable being
	   signaled, <em>one</em> of them will be woken.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_cond_broadcast(ErlDrvCond *cnd)</nametext></name>
      <fsummary>Broadcast on a condition variable</fsummary>
      <desc>
        <marker id="erl_drv_cond_broadcast"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>cnd</c></tag>
          <item>A pointer to a condition variable to broadcast on.</item>
        </taglist>
        <p>This function broadcasts on a condition variable. That is, if
	   other threads are waiting on the condition variable being
	   broadcasted on, <em>all</em> of them will be woken.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_cond_wait(ErlDrvCond *cnd, ErlDrvMutex *mtx)</nametext></name>
      <fsummary>Wait on a condition variable</fsummary>
      <desc>
        <marker id="erl_drv_cond_wait"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>cnd</c></tag>
          <item>A pointer to a condition variable to wait on.</item>
          <tag><c>mtx</c></tag>
          <item>A pointer to a mutex to unlock while waiting.</item>
          <tag><c></c></tag>
          <item></item>
        </taglist>
        <p>This function waits on a condition variable. The calling
	   thread is blocked until another thread wakes it by signaling
	   or broadcasting on the condition variable. Before the calling
	   thread is blocked it unlocks the mutex passed as argument, and
	   when the calling thread is woken it locks the same mutex before
	   returning. That is, the mutex currently has to be locked by
	   the calling thread when calling this function.
	</p>
	<note><p><c>erl_drv_cond_wait()</c> might return even though
	         no-one has signaled or broadcasted on the condition
		 variable. Code calling <c>erl_drv_cond_wait()</c> should
		 always be prepared for <c>erl_drv_cond_wait()</c>
		 returning even though the condition that the thread was
		 waiting for hasn't occurred. That is, when returning from
		 <c>erl_drv_cond_wait()</c> always check if the condition
		 has occurred, and if not call <c>erl_drv_cond_wait()</c>
		 again.
	</p></note>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>ErlDrvRWLock *</ret><nametext>erl_drv_rwlock_create(char *name)</nametext></name>
      <fsummary>Create an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_create"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>name</c></tag>
          <item>A string identifying the created rwlock. It will be used to
		identify the rwlock in planned future debug functionality.
	  </item>
        </taglist>
        <p>This function creates an rwlock and returns a pointer to it. On
	   failure <c>NULL</c> is returned. The driver creating the rwlock
	   has the responsibility of destroying it before the driver is
	   unloaded.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_rwlock_destroy(ErlDrvRWLock *rwlck)</nametext></name>
      <fsummary>Destroy an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_destroy"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>rwlck</c></tag>
          <item>A pointer to an rwlock to destroy.</item>
        </taglist>
        <p>This function destroys an rwlock previously created by
	   <seealso marker="#erl_drv_rwlock_create">erl_drv_rwlock_create()</seealso>.
	   The rwlock has to be in an unlocked state before being destroyed.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_rwlock_rlock(ErlDrvRWLock *rwlck)</nametext></name>
      <fsummary>Read lock an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_rlock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>rwlck</c></tag>
          <item>A pointer to an rwlock to read lock.</item>
        </taglist>
        <p>This function read locks an rwlock. The calling thread will be
	   blocked until the rwlock has been read locked. A thread
	   which currently has read or read/write locked the rwlock may
	   <em>not</em> lock the same rwlock again.
	</p>
	<warning><p>If you leave an rwlock locked in an emulator thread
	            when you let the thread out of your control, you will
		    <em>very likely</em> deadlock the whole emulator.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_rwlock_tryrlock(ErlDrvRWLock *rwlck)</nametext></name>
      <fsummary>Try to read lock an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_tryrlock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>rwlck</c></tag>
          <item>A pointer to an rwlock to try to read lock.</item>
        </taglist>
        <p>This function tries to read lock an rwlock. If successful
	   <c>0</c>, is returned; otherwise, <c>EBUSY</c> is returned.
	   A thread which currently has read or read/write locked the
	   rwlock may <em>not</em> try to lock the same rwlock again.
	</p>
	<warning><p>If you leave an rwlock locked in an emulator thread
	            when you let the thread out of your control, you will
		    <em>very likely</em> deadlock the whole emulator.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_rwlock_runlock(ErlDrvRWLock *rwlck)</nametext></name>
      <fsummary>Read unlock an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_runlock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>rwlck</c></tag>
          <item>A pointer to an rwlock to read unlock.</item>
        </taglist>
        <p>This function read unlocks an rwlock. The rwlock currently
	   has to be read locked by the calling thread.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_rwlock_rwlock(ErlDrvRWLock *rwlck)</nametext></name>
      <fsummary>Read/Write lock an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_rwlock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>rwlck</c></tag>
          <item>A pointer to an rwlock to read/write lock.</item>
        </taglist>
        <p>This function read/write locks an rwlock. The calling thread
	   will be blocked until the rwlock has been read/write locked.
	   A thread which currently has read or read/write locked the
	   rwlock may <em>not</em> lock the same rwlock again.
	</p>
	<warning><p>If you leave an rwlock locked in an emulator thread
	            when you let the thread out of your control, you will
		    <em>very likely</em> deadlock the whole emulator.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_rwlock_tryrwlock(ErlDrvRWLock *rwlck)</nametext></name>
      <fsummary>Try to read/write lock an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_tryrwlock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>rwlck</c></tag>
          <item>A pointer to an rwlock to try to read/write lock.</item>
        </taglist>
        <p>This function tries to read/write lock an rwlock. If successful
	   <c>0</c>, is returned; otherwise, <c>EBUSY</c> is returned.
	   A thread which currently has read or read/write locked the
	   rwlock may <em>not</em> try to lock the same rwlock again.
	</p>
	<warning><p>If you leave an rwlock locked in an emulator thread
	            when you let the thread out of your control, you will
		    <em>very likely</em> deadlock the whole emulator.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_rwlock_rwunlock(ErlDrvRWLock *rwlck)</nametext></name>
      <fsummary>Read/Write unlock an rwlock</fsummary>
      <desc>
        <marker id="erl_drv_rwlock_rwunlock"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>rwlck</c></tag>
          <item>A pointer to an rwlock to read/write unlock.</item>
        </taglist>
        <p>This function read/write unlocks an rwlock. The rwlock
	   currently has to be read/write locked by the calling thread.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_tsd_key_create(char *name, ErlDrvTSDKey *key)</nametext></name>
      <fsummary>Create a thread specific data key</fsummary>
      <desc>
        <marker id="erl_drv_tsd_key_create"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>name</c></tag>
          <item>A string identifying the created key. It will be used
	        to identify the key in planned future debug
		functionality.
	  </item>
          <tag><c>key</c></tag>
          <item>A pointer to a thread specific data key variable.</item>
        </taglist>
        <p>This function creates a thread specific data key. On success
	   <c>0</c> is returned; otherwise, an <c>errno</c> value is returned
	   to indicate the error. The driver creating the key has the
	   responsibility of destroying it before the driver is unloaded.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_tsd_key_destroy(ErlDrvTSDKey key)</nametext></name>
      <fsummary>Destroy a thread specific data key</fsummary>
      <desc>
        <marker id="erl_drv_tsd_key_destroy"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>key</c></tag>
          <item>A thread specific data key to destroy.</item>
        </taglist>
        <p>This function destroys a thread specific data key
	   previously created by
	   <seealso marker="#erl_drv_tsd_key_create">erl_drv_tsd_key_create()</seealso>.
	   All thread specific data using this key in all threads
	   have to be cleared (see
	   <seealso marker="#erl_drv_tsd_set">erl_drv_tsd_set()</seealso>)
	   prior to the call to <c>erl_drv_tsd_key_destroy()</c>.
	</p>
	<warning><p>A destroyed key is very likely to be reused soon.
		    Therefore, if you fail to clear the thread specific
		    data using this key in a thread prior to destroying
		    the key, you will <em>very likely</em> get unexpected
		    errors in other parts of the system.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void</ret><nametext>erl_drv_tsd_set(ErlDrvTSDKey key, void *data)</nametext></name>
      <fsummary>Set thread specific data</fsummary>
      <desc>
        <marker id="erl_drv_tsd_set"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>key</c></tag>
          <item>A thread specific data key.</item>
          <tag><c>data</c></tag>
          <item>A pointer to data to associate with <c>key</c>
	        in calling thread.
	  </item>
        </taglist>
        <p>This function sets thread specific data associated with
	   <c>key</c> for the calling thread. You are only allowed to set
	   thread specific data for threads while they are fully under your
	   control. For example, if you set thread specific data in a thread
	   calling a driver call-back function, it has to be cleared, i.e.
	   set to <c>NULL</c>, before returning from the driver call-back
	   function.
	</p>
	<warning><p>If you fail to clear thread specific data in an
	            emulator thread before letting it out of your control,
		    you might not ever be able to clear this data with
		    later unexpected errors in other parts of the system as
		    a result.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>void *</ret><nametext>erl_drv_tsd_get(ErlDrvTSDKey key)</nametext></name>
      <fsummary>Get thread specific data</fsummary>
      <desc>
        <marker id="erl_drv_tsd_get"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>key</c></tag>
          <item>A thread specific data key.</item>
        </taglist>
        <p>This function returns the thread specific data
	   associated with <c>key</c> for the calling thread.
	   If no data has been associated with <c>key</c> for
	   the calling thread, <c>NULL</c> is returned.
	</p>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_putenv(char *key, char *value)</nametext></name>
      <fsummary>Set the value of an environment variable</fsummary>
      <desc>
        <marker id="erl_drv_putenv"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>key</c></tag>
          <item>A null terminated string containing the
	        name of the environment variable.</item>
          <tag><c>value</c></tag>
          <item>A null terminated string containing the
	        new value of the environment variable.</item>
        </taglist>
        <p>This function sets the value of an environment variable.
	   It returns <c>0</c> on success, and a value <c>!= 0</c> on
	   failure.
	</p>
	<note><p>The result of passing the empty string ("") as a value
	      is platform dependent. On some platforms the value of the
	      variable is set to the empty string, on others, the
	      environment variable is removed.</p>
	</note>
	<warning><p>Do <em>not</em> use libc's <c>putenv</c> or similar
	         C library interfaces from a driver.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>

    <func>
      <name><ret>int</ret><nametext>erl_drv_getenv(char *key, char *value, size_t *value_size)</nametext></name>
      <fsummary>Get the value of an environment variable</fsummary>
      <desc>
        <marker id="erl_drv_getenv"></marker>
        <p>Arguments:</p>
        <taglist>
          <tag><c>key</c></tag>
          <item>A null terminated string containing the
	        name of the environment variable.</item>
          <tag><c>value</c></tag>
          <item>A pointer to an output buffer.</item>
          <tag><c>value_size</c></tag>
          <item>A pointer to an integer. The integer is both used for
	        passing input and output sizes (see below). 
		</item>
        </taglist>
	<p>This function retrieves the value of an environment variable.
	When called, <c>*value_size</c> should contain the size of
	the <c>value</c> buffer. On success <c>0</c> is returned,
	the value of the environment variable has been written to
	the <c>value</c> buffer, and <c>*value_size</c> contains the
	string length (excluding the terminating null character) of
	the value written to the <c>value</c> buffer. On failure,
	i.e., no such environment variable was found, a value less than
	<c>0</c> is returned. When the size of the <c>value</c>
	buffer is too small, a value greater than <c>0</c> is returned
	and <c>*value_size</c> has been set to the buffer size needed.
	</p>
	<warning><p>Do <em>not</em> use libc's <c>getenv</c> or similar
	         C library interfaces from a driver.
	</p></warning>
        <p>This function is thread-safe.</p>
      </desc>
    </func>
  </funcs>

  <section>
    <title>SEE ALSO</title>
    <p><seealso marker="driver_entry">driver_entry(3)</seealso>,
      <seealso marker="kernel:erl_ddll">erl_ddll(3)</seealso>,
      <seealso marker="erts:erlang">erlang(3)</seealso></p>
    <p>An Alternative Distribution Driver (ERTS User's
      Guide Ch. 3)</p>
  </section>
</cref>