Subversion Repositories HelenOS-doc

Rev

Rev 62 | Rev 75 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 62 Rev 64
Line 58... Line 58...
58 
    architectures can be compiled with this optimization.</para>
 58 
    architectures can be compiled with this optimization.</para>
59 
 
 59 
 
60 
    <para>Although the kernel does not do any floating point
 60 
    <para>Although the kernel does not do any floating point
61 
    arithmetics<footnote>
 61 
    arithmetics<footnote>
62 
        <para>Some architectures (e.g. ia64) inevitably use a fixed set of
 62 
        <para>Some architectures (e.g. ia64) inevitably use a fixed set of
63 
        floating point registers to carry out its normal operations.</para>
 63 
        floating point registers to carry out their normal operations.</para>
64 
      </footnote>, it must protect FPU context of userspace threads against
 64 
      </footnote>, it must protect FPU context of userspace threads against
65 
    destruction by other threads. Moreover, only a fraction of userspace
 65 
    destruction by other threads. Moreover, only a fraction of userspace
66 
    programs use the floating point unit. HelenOS contains a generic framework
 66 
    programs use the floating point unit. HelenOS contains a generic framework
67 
    for switching FPU context only when the switch is forced.</para>
 67 
    for switching FPU context only when the switch is forced (i.e. a thread
- 
 
 68 
    uses a floating point instruction and its FPU context is not loaded in the
- 
 
 69 
    processor).</para>
68 
 
 70 
 
69 
    <para>The last member of the context family is the memory management
 71 
    <para>The last member of the context family is the memory management
70 
    context. It includes memory management registers that identify address
 72 
    context. It includes memory management registers that identify address
71 
    spaces on hardware level (i.e. ASIDs and page tables pointers).</para>
 73 
    spaces on hardware level (i.e. ASIDs and page tables pointers).</para>
72 
 
 74 
 
Line 117... Line 119...
117 
    and to the latter group as to userspace threads. Both kernel and userspace
 119 
    and to the latter group as to userspace threads. Both kernel and userspace
118 
    threads are visible to the scheduler and can become a subject of kernel
 120 
    threads are visible to the scheduler and can become a subject of kernel
119 
    preemption and thread migration during times when preemption is
 121 
    preemption and thread migration during times when preemption is
120 
    possible.</para>
 122 
    possible.</para>
121 
 
 123 
 
122 
    <figure>
 124 
    <formalpara>
123 
      <title>Transitions among thread states.</title>
 125 
      <title>Thread states</title>
124 
 
 126 
 
- 
 
 127 
      <para>In each moment, a thread exists in one of five possible thread
- 
 
 128 
      states. When the thread is created and first readied into the
- 
 
 129 
      scheduler's run queues or when a thread is migrated to a new processor,
- 
 
 130 
      it is put into the <code>Entering</code> state. After some time, the
- 
 
 131 
      scheduler picks up the thread and starts executing it. A thread being
- 
 
 132 
      currently executed on a processor is in the <code>Running</code> state.
- 
 
 133 
      From there, the thread has three possibilities. It either runs until it
- 
 
 134 
      is preemtped, in which case the state changes to <code>Ready</code>,
- 
 
 135 
      goes to the <code>Sleeping</code> state by going to sleep or enters the
- 
 
 136 
      <code>Exiting</code> state when it reaches termination.<figure>
- 
 
 137 
          <title>Transitions among thread states.</title>
- 
 
 138 
 
125 
      <mediaobject id="thread_states" xreflabel="">
 139 
          <mediaobject id="thread_states" xreflabel="">
126 
        <imageobject role="html">
 140 
            <imageobject role="html">
127 
          <imagedata fileref="images/thread_states.png" format="PNG" />
 141 
              <imagedata fileref="images/thread_states.png" format="PNG" />
128 
        </imageobject>
 142 
            </imageobject>
129 
 
 143 
 
130 
        <imageobject role="fop">
 144 
            <imageobject role="fop">
131 
          <imagedata fileref="images.vector/thread_states.svg" format="SVG" />
 145 
              <imagedata fileref="images.vector/thread_states.svg"
- 
 
 146 
                         format="SVG" />
132 
        </imageobject>
 147 
            </imageobject>
133 
      </mediaobject>
 148 
          </mediaobject>
- 
 
 149 
        </figure></para>
134 
    </figure>
 150 
    </formalpara>
- 
 
 151 
 
- 
 
 152 
    <formalpara>
- 
 
 153 
      <title>Pseudo threads</title>
135 
 
 154 
 
136 
    <para>HelenOS userspace layer knows even smaller units of execution. Each
 155 
      <para>HelenOS userspace layer knows even smaller units of execution.
137 
    userspace thread can make use of an arbitrary number of pseudo threads.
 156 
      Each userspace thread can make use of an arbitrary number of pseudo
138 
    These pseudo threads have their own synchronous register context,
 157 
      threads. These pseudo threads have their own synchronous register
139 
    userspace code and stack. They live their own life within the userspace
 158 
      context, userspace code and stack. They live their own life within the
140 
    thread and the scheduler does not have any idea about them because they
 159 
      userspace thread and the scheduler does not have any idea about them
141 
    are completely implemented by the userspace library. This implies several
 160 
      because they are completely implemented by the userspace library. This
142 
    things:</para>
 - 
 
143 
 
 - 
 
144 
    <itemizedlist>
 161 
      implies several things:<itemizedlist>
145 
      <listitem>
 162 
          <listitem>
146 
        <para>pseudothreads schedule themselves cooperatively within the time
 163 
            <para>pseudothreads schedule themselves cooperatively within the
147 
        slice given to their userspace thread,</para>
 164 
            time slice given to their userspace thread,</para>
148 
      </listitem>
 165 
          </listitem>
149 
 
 166 
 
150 
      <listitem>
 167 
          <listitem>
151 
        <para>pseudothreads share FPU context of their containing thread
 168 
            <para>pseudothreads share FPU context of their containing thread
152 
        and</para>
 169 
            and</para>
153 
      </listitem>
 170 
          </listitem>
154 
 
 171 
 
155 
      <listitem>
 172 
          <listitem>
156 
        <para>all pseudothreads of one userspace thread block when one of them
 173 
            <para>all pseudothreads of one userspace thread block when one of
157 
        goes to sleep.</para>
 174 
            them goes to sleep.</para>
158 
      </listitem>
 175 
          </listitem>
159 
    </itemizedlist>
 176 
        </itemizedlist></para>
- 
 
 177 
    </formalpara>
160 
  </section>
 178 
  </section>
161 
 
 179 
 
162 
  <section>
 180 
  <section>
163 
    <title>Scheduler</title>
 181 
    <title>Scheduler</title>
164 
 
 182