<?xml version="1.0" encoding="UTF-8"?>
<chapter id="mm">
<?dbhtml filename="mm.html"?>
<title>Memory management</title>
<section>
<title>Virtual memory management</title>
<section>
<title>Address spaces</title>
<para></para>
</section>
<section>
<title>Virtual address translation</title>
<para></para>
</section>
</section>
<section>
<title>Physical memory management</title>
<section id="zones_and_frames">
<title>Zones and frames</title>
<para>Physical memory is divided into zones. Each zone represents
continuous area of physical memory frames. Allocation of frames is
handled by the <link linkend="buddy_allocator">buddy allocator</link>
associated with the zone. Zone also contains information about free and
occupied frames and its base addresss in the memory. Some of the
architectures (Mips, PPC) have only one zone, that covers whole physical
memory. Other architectures (IA32) have multiple zones.</para>
</section>
<section id="buddy_allocator">
<title>Buddy allocator</title>
<para>Physical memory allocation inside one <link
linkend="zones_and_frames">memory zone</link> is being handled by buddy
allocation system. This approach greatly reduces possibility of memory
fragmentation and helps in allocating bigger continious blocks of
physical memory aligned to their size.</para>
<graphic fileref="images/mm1.png" />
<para>Frames are grouped into bigger blocks and blocks of the size i ^ 2
are stored in the list indexed with <varname>i</varname> (so called
order index). If list contains 2 ajacent blocks (of a same size of
cause) they can be merged into the bigger one and moved into the list
with higher order index, thus making possible allocation of a bigger
block.</para>
</section>
<section>
<title>Slab allocator</title>
<para>Kernel memory allocation is handled by slab.</para>
</section>
</section>
<section>
<title>Memory sharing</title>
<para>Not implemented yet(?)</para>
</section>
</chapter>