lcdproc/docs/lcdproc-dev/make-driver.docbook

<chapter id="make-driver">
<title>Making a LCDproc driver</title>

<sect1 id="make-driver-intro">
<title>Introduction</title>

<para>
LCDproc is meant to be modular, it is relatively easy to add new input and output drivers to LCDproc. Actually, there are a few things that you can do to make your life easier, they are listed here.
</para>

<para>
This chapter will explain you the major steps and few gotchas of adding your own driver to LCDproc. Enjoy!
</para>

</sect1>

<sect1 id="driver-shared-files">
<title>Shared files specific for drivers</title>

<para>
Driving an LCD display is not easy; you need to address ports, to send bytes in a certain order, to respect timing, and unfortunaly no two operating system let you do this in the same way. But don't dispair! There's hope! Someone in a galaxy far far away, has allready done the dirty job for you! This dirty job has been put in shared files. These shared files are full cross platform and are automagically configured by the configure script. You only need to include them and use their functions to benefit from them.
</para>

<para>
These files are provided only for drivers, others are provided for all of LCDproc. These files are located in the shared directory, they have a dedicated chapter in this book.
</para>

<sect2 id="port.h">
<title>port.h : Parallel port I/O</title>

<para>The file port.h, located in the server/drivers/ directory provide Input/Output and port permissions for the PC compatible parallel port, also known as the LPT port.</para>

<para>
  Of course, these functions will only work if the computer where LCDproc runs has parallel port!r
  In these situations, the configure script will see this and disable drivers that need a parallel port.
</para>

<para>port.h file defines 6 static inline functions for port I/O:</para>

<sect3 id="port-in">
<title>Read a byte from port</title>

<funcsynopsis>
  <funcprototype>
    <funcdef>static inline int <function>port_in</function></funcdef>
    <paramdef>unsigned short int <parameter>port</parameter></paramdef>
  </funcprototype>
</funcsynopsis>
<para>
  Returns the content of the byte.
</para>

</sect3>
<sect3 id="port-out">
<title>Write a char(byte) 'val' to port</title>

<funcsynopsis>
  <funcprototype>
    <funcdef>static inline void <function>port_out</function></funcdef>
    <paramdef>unsigned short int <parameter>port</parameter></paramdef>
    <paramdef>unsigned char <parameter>val</parameter></paramdef>
  </funcprototype>
</funcsynopsis>
<para>
  Returns nothing (void).
</para>

</sect3>
<sect3 id="port-access">
<title>Get access to a specific port</title>

<funcsynopsis>
  <funcprototype>
    <funcdef>static inline int <function>port_access</function></funcdef>
    <paramdef>unsigned short int <parameter>port</parameter></paramdef>
  </funcprototype>
</funcsynopsis>
<para>
  Returns <literal>0</literal> if successful, <literal>-1</literal> if failed.
</para>


</sect3>
<sect3 id="port-deny">
<title>Close access to a specific port</title>

<funcsynopsis>
  <funcprototype>
    <funcdef>static inline int <function>port_deny</function></funcdef>
    <paramdef>unsigned short int <parameter>port</parameter></paramdef>
  </funcprototype>
</funcsynopsis>
<para>
  Returns <literal>0</literal> if successful, <literal>-1</literal> if failed.
</para>


</sect3>
<sect3 id="port-access-multiple">
<title>Get access to multiple sequential ports</title>

<funcsynopsis>
  <funcprototype>
    <funcdef>static inline int <function>port_access_full</function></funcdef>
    <paramdef>unsigned short int <parameter>port</parameter></paramdef>
    <paramdef>unsigned short int <parameter>count</parameter></paramdef>
  </funcprototype>
</funcsynopsis>
<para>
  Returns <literal>0</literal> if successful, <literal>-1</literal> if failed.
</para>

</sect3>
<sect3 id="port-deny-multiple">
<title>Close access to multiple sequential ports</title>

<funcsynopsis>
  <funcprototype>
    <funcdef>static inline int <function>port_deny_full</function></funcdef>
    <paramdef>unsigned short int <parameter>port</parameter></paramdef>
    <paramdef>unsigned short int <parameter>count</parameter></paramdef>
  </funcprototype>
</funcsynopsis>
<para>
  Returns <literal>0</literal> if successful, <literal>-1</literal> if failed.
</para>

</sect3>

<sect3 id="port.h-example">
<title>Example use</title>

<screen>
#include "port.h"

/* Get access to these 3 ports: 
    0x378 (CONTROL), 
    0x379 (STATUS) and 
    0x37A (DATA) 
*/
if ( -1 == port_access_multiple(0x378,3) ) {
	/* Access denied, do something */
}

/* Write a 'A' to the control port */
ort_out(0x378, 'A');

/* Read from the status port */
char status = port_in(0x379);

/* Close the 3 ports */
port_deny_multiple(0x378,3);
</screen>

</sect3>

</sect2>


<sect2 id="adv_bignum.h">
<title>adv_bignum.h : Write Big-Numbers</title>

<para>
  <filename>adv_bignum.h</filename> is the headerfile for <filename>libbignum.a</filename>
  (made from <filename>adv_bignum.c</filename>) which contains everything needed to show big-numbers,
  including the fonts for the different displays.
  (All files are located in the <filename>server/drivers/</filename> directory.) </para>
<para/>

<para>
  There are only a few requirements to the calling driver:
</para>

<itemizedlist>
  <listitem>
    <para>
      The following functions have to be implemented by the driver:
      <variablelist>
        <varlistentry>
          <term><function>height()</function></term>
          <listitem>
            <para>
              to determine the display's height and thus the maximal height of the
              big numbers to be displayed.
            </para>
          </listitem>	
        </varlistentry>

        <varlistentry>
          <term><function>get_free_chars()</function></term>
          <listitem>
            <para>
              to determine the number of user-defineable characters that can be
              used in the generation of big numbers.
            </para>
          </listitem>	
        </varlistentry>

        <varlistentry>
          <term><function>set_char()</function></term>
          <listitem>	
            <para>
              to define a character necessary to write a big number.
              Of course this is only necessary if there really are user-definieable
	      characters, i.e. only if <function>get_free_chars()</function> returns
	      a value greater <literal>0</literal>.
            </para>
          </listitem>	
        </varlistentry>

        <varlistentry>
          <term><function>chr()</function></term>
          <listitem>
            <para>
              to actually write the characters the big numbers consist of.
            </para>
          </listitem>	
        </varlistentry>
      </variablelist>  
    </para>
  </listitem>

  <listitem>
    <para>
      The display's <replaceable>cellwidth</replaceable> has to be <literal>5</literal>
      (<literal>6</literal> works also in some cases) and the <replaceable>cellheight</replaceable>
      <literal>7</literal> or <literal>8</literal>.
    </para>
  </listitem>

  <listitem>
    <para>
      The custom-characters (if any) have to be at character positions
      <replaceable>offset</replaceable><literal>+0</literal>, 
      <replaceable>offset</replaceable><literal>+1</literal>, 
      <replaceable>offset</replaceable><literal>+2</literal>, ...
      <replaceable>offset</replaceable><literal>+</literal>
      <function>get_free_chars()</function><literal>-1</literal>, 
    </para>
  </listitem>

  <listitem>
    <para>
      <replaceable>offset</replaceable><literal>+</literal>
      <function>get_free_chars()</function><literal>-1</literal> must be less than <literal>32</literal>, 
    </para>
  </listitem>
</itemizedlist>  

<para>
  The library determines the correct font, depending on the display size and the number
  of user-defined characters itself.
  So it is easy to integrate into the driver.
</para>



<sect3 id="adv_bignum.h-usage">
<title>Provided Functions</title>

<funcsynopsis>
  <funcprototype>
    <funcdef>void <function>lib_adv_bignum</function></funcdef>
    <paramdef>Driver *<parameter>drvthis</parameter></paramdef>
    <paramdef>int <parameter>x</parameter></paramdef>
    <paramdef>int <parameter>num</parameter></paramdef>
    <paramdef>int <parameter>offset</parameter></paramdef>
    <paramdef>int <parameter>do_init</parameter></paramdef>
  </funcprototype>
</funcsynopsis>

<para>
  The main thing the driver has to do is to call this function from its <function>num()</function>
  function with the parameters described below.
</para>

<variablelist>
  <varlistentry>
    <term><parameter>drvthis</parameter></term>
    <listitem>
      <para>
        the pointer pointing to the Driver structure passed to thea driver's
        <function>num()</function> function.
      </para>
    </listitem>
  </varlistentry>  

  <varlistentry>
    <term><parameter>x</parameter></term>
    <listitem>
      <para>
        the horizontal position of the top-left corner of the big-number
        (the big-numbers don't have a y position).
        The placing of the characters is done by the client, so the driver only has to forward
        the position to the lib. The bignumlib has no influence on the placing of the characters.
      </para>
    </listitem>
  </varlistentry>  

  <varlistentry>
    <term><parameter>num</parameter></term>
    <listitem>
      <para>
        the number (legal: <literal>0</literal> - <literal>9</literal>,
        and <literal>:</literal>) to be written.
      </para>
    </listitem>
  </varlistentry>  

  <varlistentry>
    <term><parameter>offset</parameter></term>
    <listitem>
      <para>  
        the character position where the user-defineable characters start
        (usually <literal>0</literal>).
        The user-defineable characters (if any) are then expected to be at the character positions
        <replaceable>offset</replaceable><literal>+0</literal>, 
        <replaceable>offset</replaceable><literal>+1</literal>, 
        <replaceable>offset</replaceable><literal>+2</literal>, ...
        <replaceable>offset</replaceable><literal>+</literal>
        <function>get_free_chars()</function><literal>-1</literal>
	and <replaceable>offset</replaceable><literal>+</literal>
       <function>get_free_chars()</function><literal>-1</literal> is required to be less than
       <literal>32</literal>. 
      </para>  
    </listitem>
  </varlistentry>  

  <varlistentry>
    <term><parameter>do_init</parameter></term>
    <listitem>
      <para>
        if not <literal>0</literal>, lib_adv_bignum will set the
        custom characters of the display for the big-numbers.
      </para>
      <para>
        The driver has to check if the custom-characters have to be set or if it is
        already done and tell it to the lib (using the <parameter>do_init</parameter> parameter).
        The common way is to use variable called <replaceable>p-&gt;ccmode</replaceable> or similar.
        In the different drivers there are some differences in the naming and handling of this variable.
        So the responsibility of checking and setting is left to the driver.
      </para>
    </listitem>
  </varlistentry>  
</variablelist>

<example id="lib_adv_bignum-example">
<title>Calling <function>lib_adv_bignum()</function></title>

<programlisting>
<emphasis>#include "adv_bignum.h"</emphasis>

MODULE_EXPORT void
myDriver_num( Driver * drvthis, int x, int num )
{
  PrivateData *p = drvthis-&gt;private_data;
  int do_init = 0;

  if (p-&gt;ccmode != CCMODE_BIGNUM){     // Are the custom-characters set up correctly? If not:
    do_init = 1;                       // Lib_adv_bignum has to set the custom-characters.
    p-&gt;ccmode = CCMODE_BIGNUM;         // Switch custom-charactermode to bignum.
  }

  // Lib_adv_bignum does everything needed to show the big-numbers.
  <emphasis>lib_adv_bignum(drvthis, x, num, 0, do_init);</emphasis>
}
</programlisting>

</example>

<para>
  All that's left to do is to add <filename>libbignum.a</filename> to the libs and
  <filename>adv_bignum.h</filename> sources of your driver in the <filename>Makefile</filename>
  (or the file that generates the
  <filename>Makefile</filename>).
</para>

<example id="adv_bignum.h-Makefile.am-example">
<title>Enabling adv_bignum support in <filename>Makefile.am</filename></title>

<programlisting>
myDriver_LDADD: libLCD.a <emphasis>libbignum.a</emphasis>

myDriver_SOURCES: lcd.h lcd_lib.h myDriver.c myDriver.h report.h <emphasis>adv_bignum.h</emphasis>
</programlisting>

</example>

</sect3>

<sect3 id="adv_bignum.h-intern">
<title>Internal Structure and Functions</title>

<para>
  The only purpose of <function>lib_adv_bignum()</function> is to determine the best
  display-dependent big-number function, based upon the values of the driver's
  <function>height()</function> and <function>get_free_chars()</function> functions,
  and call it.
</para>

<para>
  The display-dependent functions are named
  <function>adv_bignum_num_<replaceable>N_M</replaceable>()</function>,
  where <replaceable>N</replaceable> is the display's height in lines and <replaceable>M</replaceable>
  the number of used user-defineable characters.
  The bits of the user-characters are stored in <replaceable>static char bignum</replaceable>
  (take a look at the source and you will see what I mean). (On a display with a
  <replaceable>cellheight</replaceable> of 7 the lowest line stored is not shown.)
  While <replaceable>static char num_map</replaceable> defines the placing in the big number.
  (A big number is always 3 characters wide and 4 characters high.
  On a big number for 2 line displays the 2 lower lines are not in use.)
</para>

<para>
  If user-defineable characters have to be set, the driver's <function>set_char()</function> function
  will be called once for every user-idefneable character.
</para>

<para>
  Now <function>adv_bignum_write_num()</function> is called.
  This function places the 6 or 12 characters the big-number consists of in the framebuffer
  using the drivers <function>chr()</function> function.
</para>

</sect3>

</sect2>

</sect1>

</chapter>