Subversion Repositories HelenOS-doc

Rev

Rev 61 | Rev 73 | Go to most recent revision | Blame | Compare with Previous | Last modification | View Log | Download | RSS feed

  1. <?xml version="1.0" encoding="UTF-8"?>
  2. <chapter id="ds">
  3.   <?dbhtml filename="ds.html"?>
  4.  
  5.   <title>Data structures</title>
  6.  
  7.   <para>There is lots of data that either flows through various HelenOS
  8.   subsystems or is stored directly by them. Each subsystem uses its own data
  9.   structures to represent the data. These data structures need to be kept
  10.   somewhere. In order to work efficiently, HelenOS, and especially its kernel,
  11.   deploys several house keeping data types that are designed to faciliate
  12.   managing other data structures. Most of them serve like generic
  13.   containers.</para>
  14.  
  15.   <section>
  16.     <title>Lists</title>
  17.  
  18.     <para>HelenOS uses doubly-circularly-linked lists to bind related data
  19.     together. Lists are composed of an independent sentinel node called head
  20.     and links that are always part of the object that is to be put into the
  21.     list. Adding items to a list thus doesn't require any further memory
  22.    allocations. Head and each link then contains forward and backward
  23.    pointer. An empty list is composed of a sole head whose both pointers
  24.    reference the head itself. The expense of two times bigger memory
  25.    consumption as compared to memory consumption of singly linked lists is
  26.    justified by constant insertion and removal times at random positions
  27.    within the list.</para>
  28.  
  29.    <para>Lists are frequently used to implement FIFO behaviour (e.g.
  30.    scheduler run queues or synchronization wait queues). Contrary to the FIFO
  31.    type, which is also supported by HelenOS, they don't take up any unused
  32.     space and are more general. On the other hand, they are slower than
  33.     in-array FIFOs and can be hardly used to implement buffers.</para>
  34.   </section>
  35.  
  36.   <section>
  37.     <title>FIFO queues</title>
  38.  
  39.     <para>FIFO queues are implemented as either statically or dynamically
  40.     allocated arrays<footnote>
  41.         <para>Depending on the array size.</para>
  42.       </footnote> of some generic type with two indices. The first index
  43.     points to the head of the FIFO queue and the other points to the tail
  44.     thereof. There can be as many items in the FIFO as is the number of
  45.     elements in the array and no more. The indices are taken modulo size of
  46.     the queue because as a consequence of insertions and deletions, the tail
  47.     can have numericaly lower index than the head.</para>
  48.  
  49.     <para>FIFO queues are used, for example, in ASID management code to store
  50.     inactive ASIDs or in userspace keyboard driver to buffer read
  51.     characters.</para>
  52.  
  53.     <figure>
  54.     <mediaobject id="fifo" xreflabel="">
  55.     <imageobject role="html">
  56.         <imagedata fileref="images/fifo.png" format="PNG" />
  57.     </imageobject>
  58.            
  59.     <imageobject role="fop">
  60.         <imagedata fileref="images.vector/fifo.svg" format="SVG" />
  61.     </imageobject>
  62.                    
  63.     </mediaobject>
  64.     <title>FIFO queue showing the wrap around the end of the array.</title>
  65.     </figure>
  66.  
  67.   </section>
  68.  
  69.   <section>
  70.     <title>Hash tables</title>
  71.  
  72.     <para>The kernel, as well as userspace, provides hash table data type
  73.     which uses separate chaining. The hash table type is very generic in that
  74.     it forces the user to supply methods for computing the hash index,
  75.     comparing items against a set of keys and the item removal callback
  76.     function. Besides these virtual operations, the hash table is composed of
  77.     a dynamically allocated array of list heads that represent each chain,
  78.     number of chains and the maximal number of keys.</para>
  79.   </section>
  80.  
  81.   <section>
  82.     <title>Bitmaps</title>
  83.  
  84.     <para>Several bitmap operations such as clearing or setting consecutive
  85.     bit sequences as well as copying portions of one bitmap into another one
  86.     are supported.</para>
  87.   </section>
  88.  
  89.   <section>
  90.     <title>B+trees</title>
  91.  
  92.     <para>HelenOS makes use of a variant of B-tree called B+tree. B+trees, in
  93.     HelenOS implementation, are 3-4-5 balanced trees. They are characteristic
  94.     by the fact that values are kept only in the leaf-level nodes and that
  95.     these nodes are linked together in a list. This data structure has
  96.     logaritmic search, insertion and deletion times and, thanks to the
  97.     leaf-level list, provides fantastic means of walking the nodes containing
  98.     data. Moreover, B+trees can be used for easy storing, resizing and merging
  99.     of disjunctive intervals.</para>
  100.  
  101.     <figure>
  102.     <mediaobject id="btree" xreflabel="">
  103.     <imageobject role="html">
  104.         <imagedata fileref="images/btree.png" format="PNG" />
  105.     </imageobject>
  106.            
  107.     <imageobject role="fop">
  108.         <imagedata fileref="images.vector/btree.svg" format="SVG" />
  109.     </imageobject>
  110.                    
  111.  
  112.     </mediaobject>
  113.     <title>B+tree containing keys ranging from 1 to 12.</title>
  114.     </figure>
  115.  
  116.   </section>
  117. </chapter>