<!-- {{{ Banner                         -->

<!-- =============================================================== -->
<!--                                                                 -->
<!--     HAL.sgml                                                    -->
<!--                                                                 -->
<!--     eCos common HAL documentation                               -->
<!--                                                                 -->
<!-- =============================================================== -->
<!-- ####COPYRIGHTBEGIN####                                          -->
<!--                                                                 -->
<!-- =============================================================== -->
<!-- Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002 Red Hat, Inc.  -->
<!-- This material may be distributed only subject to the terms      -->
<!-- and conditions set forth in the Open Publication License, v1.0  -->
<!-- or later (the latest version is presently available at          -->
<!-- http://www.opencontent.org/openpub/)                            -->
<!-- Distribution of the work or derivative of the work in any       -->
<!-- standard (paper) book form is prohibited unless prior           -->
<!-- permission obtained from the copyright holder                   -->
<!-- =============================================================== -->
<!--                                                                 -->      
<!-- ####COPYRIGHTEND####                                            -->
<!-- =============================================================== -->
<!-- #####DESCRIPTIONBEGIN####                                       -->
<!--                                                                 -->
<!-- ####DESCRIPTIONEND####                                          -->
<!-- =============================================================== -->

<!-- }}} -->

<part id="the-ecos-hardware-abstraction-layer">
<title>The eCos Hardware Abstraction Layer (HAL)</title>

<!-- {{{ Intro -->

<chapter id=hal-introduction>
<title>Introduction</title>
<PARA>
This is an initial specification of the <EMPHASIS>eCos</EMPHASIS> <!--
<index></index> -->Hardware Abstraction Layer (HAL). The HAL abstracts
the underlying hardware of a processor architecture and/or the
platform to a level sufficient for the eCos kernel to be ported onto
that platform.
</PARA>

<note>
<title>Caveat</title>
<PARA>
This document is an informal description of the HAL capabilities and
is not intended to be full documentation, although it may be used as a
source for such. It also describes the HAL as it is currently
implemented for the architectures targeted in this release. It most
closely describes the HALs for the MIPS, I386 and PowerPC HALs. Other
architectures are similar but may not be organized precisely as
described here.
</PARA>
</note>

</chapter>

<!-- }}} -->
<!-- {{{ Architecture, Variant and Platform  -->

<CHAPTER id="hal-architecture-variant-and-platform">
<TITLE>Architecture, Variant and Platform</TITLE>

<para>
We have identified three levels at which the HAL must operate.
</para>

<itemizedlist>
  <listitem>
    <para>
    The <!-- <index></index> --><firstterm>architecture
    HAL</firstterm> abstracts the basic CPU architecture and includes
    things like interrupt delivery, context switching, CPU startup
    etc.
    </para>
  </listitem>

  <listitem>
    <para>
    The <!-- <index></index> --> <firstterm>variant HAL</firstterm>
    encapsulates features of the CPU variant such as caches, MMU and
    FPU features. It also deals with any on-chip peripherals such as
    memory and interrupt controllers. For architectural variations,
    the actual implementation of the variation is often in the
    architectural HAL, and the variant HAL simply provides the correct
    configuration definitions.
    </para>
  </listitem>

    <listitem>
    <para>
    The <!-- <index></index> --><firstterm>platform HAL</firstterm>
    abstracts the properties of the current platform and includes
    things like platform startup, timer devices, I/O register access
    and interrupt controllers.
    </para>
  </listitem>

</itemizedlist>

<para>
The boundaries between these three HAL levels are necessarily blurred
since functionality shifts between levels on a target-by-target basis.
For example caches and MMU may be either an architecture feature or a
variant feature. Similarly, memory and interrupt controllers may be
on-chip and in the variant HAL, or off-chip and in the platform HAL.
</para>
<para>
Generally there is a separate package for each of the architecture,
variant and package HALs for a target. For some of the older targets,
or where it would be essentially empty, the variant HAL is omitted.
</para>
</CHAPTER>

<!-- }}} -->
<!-- {{{ General Principles -->

<CHAPTER id="hal-general-principles">
<TITLE>General principles</TITLE>

<PARA>
The HAL has been implemented according to the following general
principles:
</PARA>
<ORDEREDLIST>
<LISTITEM>
<PARA>	      The HAL is implemented in C and assembler, although the
	      eCos kernel is largely implemented in C&plus;&plus;.
	      This is to permit the HAL the widest possible
	      applicability.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>All interfaces to the HAL are implemented by
	      CPP macros. This allows them to be implemented as inline
	      C code, inline assembler or function calls to external C
	      or assembler code. This allows the most efficient
	      implementation to be selected without affecting the
	      interface. It also allows them to be redefined if the
	      platform or variant HAL needs to replace or enhance a definition
	      from the architecture HAL.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>The HAL provides simple, portable mechanisms for dealing
with the hardware of a wide range of architectures and platforms.
It is always possible to bypass the HAL and program the hardware
directly, but this may lead to a loss of portability. </PARA>
</LISTITEM>
</ORDEREDLIST>
</CHAPTER>

<!-- }}} -->
<!-- {{{ HAL Interfaces -->

<CHAPTER id="hal-interfaces">
<TITLE><!-- <index></index> --><!-- <xref> -->HAL Interfaces</TITLE>

<para>
This section describes the main HAL interfaces. 
</para>

<!-- {{{ Base Definitions -->

<SECTION id="hal-base-definitions">
<TITLE>Base Definitions</TITLE>

<para>
These are definitions that characterize the properties of the base
architecture that are used to compile the portable parts of the
kernel. They are concerned with such things a portable type
definitions, endianness, and labeling.
</para>

<PARA>
These definitions are supplied by the
<filename>cyg/hal/basetype.h</filename> header file which is supplied
by the architecture HAL. It is included automatically by
<FILENAME>cyg/infra/cyg_type.h</FILENAME>.
</PARA>

<!-- =================================================================== -->

<SECTION>
<TITLE>Byte order</TITLE>
<VARIABLELIST>
  <VARLISTENTRY>
    <TERM><VARNAME>CYG_BYTEORDER</VARNAME></TERM>
    <LISTITEM>
      <PARA>
      This defines the byte order of the target and must be set to either
      <varname>CYG_LSBFIRST</varname> or <varname>CYG_MSBFIRST</varname>.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Label Translation</TITLE>

<VARIABLELIST>
  <VARLISTENTRY>
    <TERM><FUNCTION>CYG_LABEL_NAME(name)</FUNCTION></TERM>
    <LISTITEM>

      <PARA>
      This is a wrapper used in some C and C&plus;&plus; files which
      use labels defined in assembly code or the linker script.
      It need only be defined if the default implementation in
      <filename>cyg/infra/cyg_type.h</filename>, which passes the name
      argument unaltered, is inadequate. It should be paired with
      <function>CYG_LABEL_DEFN()</function>.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  <VARLISTENTRY>
    <TERM><FUNCTION>CYG_LABEL_DEFN(name)</FUNCTION></TERM>
    <LISTITEM>

      <PARA>
      This is a wrapper used in assembler sources and linker scripts
      which define labels.  It need only be defined if the default
      implementation in
      <filename>cyg/infra/cyg_type.h</filename>, which passes the name
      argument unaltered, is inadequate. The most usual alternative
      definition of this macro prepends an underscore to the label
      name. 
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>
</SECTION>


<!-- =================================================================== -->

<SECTION>
<TITLE>Base types</TITLE>
<PROGRAMLISTING>
	cyg_halint8
	cyg_halint16
	cyg_halint32
	cyg_halint64
	cyg_halcount8
	cyg_halcount16
	cyg_halcount32
	cyg_halcount64
	cyg_halbool
</PROGRAMLISTING>
<PARA>
These macros define the C base types that should be used to define
variables of the given size. They only need to be defined if the
default types specified in <filename>cyg/infra/cyg_type.h</filename>
cannot be used. Note that these are only the base types, they will be
composed with <literal>signed</literal> and
<literal>unsigned</literal> to form full type specifications.
</PARA>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Atomic types</TITLE>
<PROGRAMLISTING>
	cyg_halatomic CYG_ATOMIC
</PROGRAMLISTING>
<PARA>
These types are guaranteed to be read or written in a single
uninterruptible operation. It is architecture defined what size this
type is, but it will be at least a byte.
</PARA>
</SECTION>

</SECTION>

<!-- }}} -->
<!-- {{{ Architecture Characterization  -->

<SECTION id="hal-architecture-characterization">
<TITLE>Architecture Characterization</TITLE>

<para>
These are definition that are related to the basic architecture of the
CPU. These include the CPU context save format, context switching, bit
twiddling, breakpoints, stack sizes and address translation.
</para>

<PARA>
Most of these definition are found in
<filename>cyg/hal/hal_arch.h</filename>.  This file is supplied by the
architecture HAL. If there are variant or platform specific
definitions then these will be found in
<filename>cyg/hal/var_arch.h</filename> or
<filename>cyg/hal/plf_arch.h</filename>. These files are include
automatically by this header, so need not be included explicitly.
</PARA>

<!-- =================================================================== -->

<SECTION>
<TITLE>Register Save Format</TITLE>
<PROGRAMLISTING>
typedef struct HAL_SavedRegisters
{
    /* architecture-dependent list of registers to be saved */ 
} HAL_SavedRegisters;
</PROGRAMLISTING>
<PARA>
This structure describes the layout of a saved machine state on the
stack. Such states are saved during thread context switches,
interrupts and exceptions. Different quantities of state may be saved
during each of these, but usually a thread context state is a subset
of the interrupt state which is itself a subset of an exception state.
For debugging purposes, the same structure is used for all three
purposes, but where these states are significantly different, this
structure may contain a union of the three states.
</PARA>
</SECTION>


<!-- =================================================================== -->

<SECTION>
<TITLE>Thread Context Initialization</TITLE>

<PROGRAMLISTING>
HAL_THREAD_INIT_CONTEXT( sp, arg, entry, id )
</PROGRAMLISTING>

<PARA>
This macro initializes a thread's context so that
it may be switched to by <FUNCTION>HAL_THREAD_SWITCH_CONTEXT()</FUNCTION>.
The arguments are:
</PARA>
<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>sp</TERM>
    <LISTITEM>
      <PARA>
      A location containing the current value of the thread's stack
      pointer. This should be a variable or a structure field. The SP
      value will be read out of here and an adjusted value written
      back.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>arg</TERM>
    <LISTITEM>
      <PARA>
      A value that is passed as the first argument to the entry
      point function.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  <VARLISTENTRY>
    <TERM>entry</TERM>
    <LISTITEM>
      <PARA>
      The address of an entry point function. This will be called
      according the C calling conventions, and the value of
      <parameter>arg</parameter> will be passed as the first
      argument. This function should have the following type signature
      <function>void entry(CYG_ADDRWORD arg)</function>.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  <VARLISTENTRY>
    <TERM>id</TERM>
    <LISTITEM>
      <PARA>
      A thread id value. This is only used for debugging purposes,
      it is ORed into the initialization pattern for unused registers
      and may be used to help identify the thread from its register dump.
      The least significant 16 bits of this value should be zero to allow
      space for a register identifier.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>
</SECTION>

<!-- =================================================================== -->

<SECTION id="hal-context-switch">
<TITLE>Thread Context Switching</TITLE>

<PROGRAMLISTING>
HAL_THREAD_LOAD_CONTEXT( to )
HAL_THREAD_SWITCH_CONTEXT( from, to )
</PROGRAMLISTING>
<PARA>
These macros implement the thread switch code. The arguments are:
</PARA>

<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>from</TERM>
    <LISTITEM>
      <PARA>
      A pointer to a location where the stack pointer of the current
      thread will be stored.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  <VARLISTENTRY>
    <TERM>to</TERM>
    <LISTITEM>
      <PARA>
      A pointer to a location from where the stack pointer of the next
      thread will be read.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>

<para>
For <function>HAL_THREAD_LOAD_CONTEXT()</function> the current CPU
state is discarded and the state of the destination thread is
loaded. This is only used once, to load the first thread when the
scheduler is started.
</para>

<PARA>
For <function>HAL_THREAD_SWITCH_CONTEXT()</function> the state of the
current thread is saved onto its stack, using the current value of the
stack pointer, and the address of the saved state placed in
<parameter>*from</parameter>.  The value in
<parameter>*to</parameter> is then read and the state of the new
thread is loaded from it.
</PARA>

<para>
While these two operations may be implemented with inline assembler,
they are normally implemented as calls to assembly code functions in
the HAL. There are two advantages to doing it this way. First, the
return link of the call provides a convenient PC value to be used in
the saved context. Second, the calling conventions mean that the
compiler will have already saved the caller-saved registers before the
call, so the HAL need only save the callee-saved registers.
</para>

<para>
The implementation of <function>HAL_THREAD_SWITCH_CONTEXT()</function>
saves the current CPU state on the stack, including the current
interrupt state (or at least the register that contains it). For
debugging purposes it is useful to save the entire register set, but
for performance only the ABI-defined callee-saved registers need be
saved. If it is implemented, the option
<literal>CYGDBG_HAL_COMMON_CONTEXT_SAVE_MINIMUM</literal> controls
how many registers are saved.
</para>

<para>
The implementation of <function>HAL_THREAD_LOAD_CONTEXT()</function>
loads a thread context, destroying the current context. With a little
care this can be implemented by sharing code with
<function>HAL_THREAD_SWITCH_CONTEXT()</function>. To load a thread
context simply requires the saved registers to be restored from the
stack and a jump or return made back to the saved PC.
</para>

<PARA>
Note that interrupts are not disabled during this process, any
interrupts that occur will be delivered onto the stack to which the
current CPU stack pointer points. Hence the stack pointer
should never be invalid, or loaded with a value that might cause the
saved state to become corrupted by an interrupt. However, the current
interrupt state is saved and restored as part of the thread
context. If a thread disables interrupts and does something to cause a
context switch, interrupts may be re-enabled on switching to another
thread. Interrupts will be disabled again when the original thread
regains control.
</PARA>

</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Bit indexing</TITLE>

<PROGRAMLISTING>
HAL_LSBIT_INDEX( index, mask )
HAL_MSBIT_INDEX( index, mask )
</PROGRAMLISTING>

<PARA>
These macros place in <parameter>index</parameter> the bit index of
the least significant bit in <parameter>mask</parameter>. Some
architectures have instruction level support for one or other of these
operations. If no architectural support is available, then these
macros may call C functions to do the job.
</PARA>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Idle thread activity</TITLE>

<PROGRAMLISTING>
HAL_IDLE_THREAD_ACTION( count )
</PROGRAMLISTING>

<PARA>
It may be necessary under some circumstances for the HAL to execute
code in the kernel idle thread's loop. An example might be to execute
a processor halt instruction. This macro provides a portable way of
doing this. The argument is a copy of the idle thread's loop counter,
and may be used to trigger actions at longer intervals than every
loop.
</PARA>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Reorder barrier</TITLE>

<PROGRAMLISTING>
HAL_REORDER_BARRIER()
</PROGRAMLISTING>

<PARA>
When optimizing the compiler can reorder code. In some parts of
multi-threaded systems, where the order of actions is vital, this can
sometimes cause problems. This macro may be inserted into places where
reordering should not happen and prevents code being migrated across
it by the compiler optimizer. It should be placed between statements
that must be executed in the order written in the code.
</PARA>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Breakpoint support</TITLE>

<PROGRAMLISTING>
HAL_BREAKPOINT( label )
HAL_BREAKINST
HAL_BREAKINST_SIZE
</PROGRAMLISTING>

<PARA>
These macros provide support for breakpoints.
</PARA>

<PARA>
<FUNCTION>HAL_BREAKPOINT()</FUNCTION> executes a breakpoint
instruction. The label is defined at the breakpoint instruction so
that exception code can detect which breakpoint was executed.
</PARA>

<PARA>
<literal>HAL_BREAKINST</literal> contains the breakpoint instruction
code as an integer value. <literal>HAL_BREAKINST_SIZE</literal> is
the size of that breakpoint instruction in bytes. Together these
may be used to place a breakpoint in any code.
</PARA>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>GDB support</TITLE>

<PROGRAMLISTING>
HAL_THREAD_GET_SAVED_REGISTERS( sp, regs )
HAL_GET_GDB_REGISTERS( regval, regs )
HAL_SET_GDB_REGISTERS( regs, regval )
</PROGRAMLISTING>

<PARA>
These macros provide support for interfacing GDB to the HAL.
</PARA>

<PARA>
<FUNCTION>HAL_THREAD_GET_SAVED_REGISTERS()</FUNCTION> extracts a
pointer to a <STRUCTNAME>HAL_SavedRegisters</STRUCTNAME> structure
from a stack pointer value. The stack pointer passed in should be the
value saved by the thread context macros. The macro will assign a
pointer to the <STRUCTNAME>HAL_SavedRegisters</STRUCTNAME> structure
to the variable passed as the second argument.
</PARA>

<PARA>
<FUNCTION>HAL_GET_GDB_REGISTERS()</FUNCTION> translates a register
state as saved by the HAL and into a register dump in the format
expected by GDB. It takes a pointer to a
<STRUCTNAME>HAL_SavedRegisters</STRUCTNAME> structure in the
<parameter>regs</parameter> argument and a pointer to the memory to
contain the GDB register dump in the <parameter>regval</parameter>
argument.
</PARA>

<PARA>
<FUNCTION>HAL_SET_GDB_REGISTERS()</FUNCTION> translates a GDB format
register dump into a the format expected by the HAL.  It takes a
pointer to the memory containing the GDB register dump in the
<parameter>regval</parameter> argument and a pointer to a
<STRUCTNAME>HAL_SavedRegisters</STRUCTNAME> structure
in the <parameter>regs</parameter> argument.
</PARA>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Setjmp and longjmp support</TITLE>

<PROGRAMLISTING>
CYGARC_JMP_BUF_SIZE
hal_jmp_buf[CYGARC_JMP_BUF_SIZE]
hal_setjmp( hal_jmp_buf env )
hal_longjmp( hal_jmp_buf env, int val )
</PROGRAMLISTING>

<PARA>
These functions provide support for the C
<FUNCTION>setjmp()</FUNCTION> and <FUNCTION>longjmp()</FUNCTION>
functions.  Refer to the C library for further information.
</PARA>

</SECTION>

<!-- =================================================================== -->

<section>
<title>Stack Sizes</title>
<programlisting>
CYGNUM_HAL_STACK_SIZE_MINIMUM
CYGNUM_HAL_STACK_SIZE_TYPICAL
</programlisting>

<para>
The values of these macros define the minimum and typical sizes of
thread stacks.
</para>

<para>
<literal>CYGNUM_HAL_STACK_SIZE_MINIMUM</literal> defines the minimum
size of a thread stack. This is enough for the thread to function
correctly within eCos and allows it to take interrupts and context
switches. There should also be enough space for a simple thread entry
function to execute and call basic kernel operations on objects like
mutexes and semaphores. However there will not be enough room for much
more than this. When creating stacks for their own threads,
applications should determine the stack usage needed for application
purposes and then add
<literal>CYGNUM_HAL_STACK_SIZE_MINIMUM</literal>.
</para>

<para>
<literal>CYGNUM_HAL_STACK_SIZE_TYPICAL</literal> is a reasonable increment over
<literal>CYGNUM_HAL_STACK_SIZE_MINIMUM</literal>, usually about 1kB. This should be
adequate for most modest thread needs. Only threads that need to
define significant amounts of local data, or have very deep call trees
should need to use a larger stack size.
</para>

</section>


<!-- =================================================================== -->

<section>
<title>Address Translation</title>

<programlisting>
CYGARC_CACHED_ADDRESS(addr)
CYGARC_UNCACHED_ADDRESS(addr)
CYGARC_PHYSICAL_ADDRESS(addr)
</programlisting>

<para>
These macros provide address translation between different views of
memory. In many architectures a given memory location may be visible
at different addresses in both cached and uncached forms. It is also
possible that the MMU or some other address translation unit in the
CPU presents memory to the program at a different virtual address to
its physical address on the bus.
</para>

<para>
<function>CYGARC_CACHED_ADDRESS()</function> translates the given
address to its location in cached memory. This is typically where the
application will access the memory.
</para>

<para>
<function>CYGARC_UNCACHED_ADDRESS()</function> translates the given
address to its location in uncached memory. This is typically where
device drivers will access the memory to avoid cache problems. It may
additionally be necessary for the cache to be flushed before the
contents of this location is fully valid.
</para>

<para>
<function>CYGARC_PHYSICAL_ADDRESS()</function> translates the given
address to its location in the physical address space. This is
typically the address that needs to be passed to device hardware such
as a DMA engine, ethernet device or PCI bus bridge. The physical
address may not be directly accessible to the program, it may be
re-mapped by address translation.
</para>

</section>


<!-- =================================================================== -->

<section>
<title>Global Pointer</title>

<programlisting>
CYGARC_HAL_SAVE_GP()
CYGARC_HAL_RESTORE_GP()
</programlisting>

<para>
These macros insert code to save and restore any global data pointer
that the ABI uses. These are necessary when switching context between
two eCos instances - for example between an eCos application and
RedBoot.
</para>

</section>

</SECTION>

<!-- }}} -->
<!-- {{{ Interrupt Handling -->

<SECTION id="hal-interrupt-handling">
<TITLE>Interrupt Handling</TITLE>

<para>
These interfaces contain definitions related to interrupt
handling. They include definitions of exception and interrupt numbers,
interrupt enabling and masking, and realtime clock operations.
</para>

<PARA>
These definitions are normally found in
<FILENAME>cyg/hal/hal_intr.h</FILENAME>.  This file is supplied by the
architecture HAL.  Any variant or platform specific definitions will
be found in <filename>cyg/hal/var_intr.h</filename>,
<filename>cyg/hal/plf_intr.h</filename> or
<filename>cyg/hal/hal_platform_ints.h</filename> in the variant or platform
HAL, depending on the exact target. These files are include
automatically by this header, so need not be included explicitly.
</PARA>

<!-- =================================================================== -->

<SECTION>
<TITLE>Vector numbers</TITLE>

<PROGRAMLISTING>
CYGNUM_HAL_VECTOR_XXXX
CYGNUM_HAL_VSR_MIN
CYGNUM_HAL_VSR_MAX
CYGNUM_HAL_VSR_COUNT

CYGNUM_HAL_INTERRUPT_XXXX
CYGNUM_HAL_ISR_MIN
CYGNUM_HAL_ISR_MAX
CYGNUM_HAL_ISR_COUNT

CYGNUM_HAL_EXCEPTION_XXXX
CYGNUM_HAL_EXCEPTION_MIN
CYGNUM_HAL_EXCEPTION_MAX
CYGNUM_HAL_EXCEPTION_COUNT
</PROGRAMLISTING>

<PARA>
All possible VSR, interrupt and exception vectors are specified here,
together with maximum and minimum values for range checking. While the
VSR and exception numbers will be defined in this file, the interrupt
numbers will normally be defined in the variant or platform HAL file
that is included by this header. 
</PARA>

<PARA>
There are two ranges of numbers, those for the vector service
routines and those for the interrupt service routines. The relationship
between these two ranges is undefined, and no equivalence should
be assumed if vectors from the two ranges coincide.
</PARA>

<PARA>
The VSR vectors correspond to the set of exception vectors that can be
delivered by the CPU architecture, many of these will be internal
exception traps. The ISR vectors correspond to the set of external
interrupts that can be delivered and are usually determined by extra
decoding of the interrupt controller by the interrupt VSR.
</PARA>

<PARA>
Where a CPU supports synchronous exceptions, the range of such
exceptions allowed are defined by <literal>CYGNUM_HAL_EXCEPTION_MIN</literal> and
<literal>CYGNUM_HAL_EXCEPTION_MAX</literal>. The
<literal>CYGNUM_HAL_EXCEPTION_XXXX</literal> definitions are
standard names used by target independent code to test for the
presence of particular exceptions in the architecture. The actual
exception numbers will normally correspond to the VSR exception
range. In future other exceptions generated by the system software
(such as stack overflow) may be added.
</PARA>

<PARA>
<literal>CYGNUM_HAL_ISR_COUNT</literal>, <literal>CYGNUM_HAL_VSR_COUNT</literal> and
<literal>CYGNUM_HAL_EXCEPTION_COUNT</literal> define the number of
ISRs, VSRs and EXCEPTIONs respectively for the purposes of defining
arrays etc. There might be a translation from the supplied vector
numbers into array offsets. Hence
<literal>CYGNUM_HAL_XXX_COUNT</literal> may not simply be
<literal>CYGNUM_HAL_XXX_MAX</literal> - <literal>CYGNUM_HAL_XXX_MIN</literal> or <literal>CYGNUM_HAL_XXX_MAX</literal>&plus;1.
</PARA>

</SECTION>


<!-- =================================================================== -->

<SECTION>
<TITLE>Interrupt state control</TITLE>

<PROGRAMLISTING>
CYG_INTERRUPT_STATE
HAL_DISABLE_INTERRUPTS( old )
HAL_RESTORE_INTERRUPTS( old )
HAL_ENABLE_INTERRUPTS()
HAL_QUERY_INTERRUPTS( state )
</PROGRAMLISTING>

<PARA>
These macros provide control over the state of the CPUs interrupt mask
mechanism. They should normally manipulate a CPU status register to
enable and disable interrupt delivery. They should not access an
interrupt controller.
</PARA>


<para>
<literal>CYG_INTERRUPT_STATE</literal> is a data type that should be
used to store the interrupt state returned by
<function>HAL_DISABLE_INTERRUPTS()</function> and
<function>HAL_QUERY_INTERRUPTS()</function> and passed to
<function>HAL_RESTORE_INTERRUPTS()</function>.
</para>

<PARA>
<FUNCTION>HAL_DISABLE_INTERRUPTS()</FUNCTION> disables the delivery of
interrupts and stores the original state of the interrupt mask in the
variable passed in the <parameter>old</parameter> argument.
</PARA>

<PARA>
<FUNCTION>HAL_RESTORE_INTERRUPTS()</FUNCTION> restores the state of
the interrupt mask to that recorded in <parameter>old</parameter>.
</PARA>

<PARA>
<FUNCTION>HAL_ENABLE_INTERRUPTS()</FUNCTION> simply enables interrupts
regardless of the current state of the mask.
</PARA>

<PARA>
<FUNCTION>HAL_QUERY_INTERRUPTS()</FUNCTION> stores the state of the
interrupt mask in the variable passed in the <parameter>
state</parameter> argument. The state stored here should also be
capable of being passed to
<function>HAL_RESTORE_INTERRUPTS()</function> at a later point.
</PARA>

<PARA>
It is at the HAL implementer&rsquo;s discretion exactly
which interrupts are masked by this mechanism. Where a CPU has more
than one interrupt type that may be masked separately (e.g. the
ARM's IRQ and FIQ) only those that can raise DSRs need
to be masked here. A separate architecture specific mechanism may
then be used to control the other interrupt types.
</PARA>

</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>ISR and VSR management</TITLE>

<PROGRAMLISTING>
HAL_INTERRUPT_IN_USE( vector, state )
HAL_INTERRUPT_ATTACH( vector, isr, data, object )
HAL_INTERRUPT_DETACH( vector, isr )
HAL_VSR_SET( vector, vsr, poldvsr )
HAL_VSR_GET( vector, pvsr )
HAL_VSR_SET_TO_ECOS_HANDLER( vector, poldvsr )
</PROGRAMLISTING>

<PARA>
These macros manage the attachment of interrupt and vector service
routines to interrupt and exception vectors respectively.
</PARA>

<para>
<function>HAL_INTERRUPT_IN_USE()</function> tests the state of the
supplied interrupt vector and sets the value of the state parameter to
either 1 or 0 depending on whether there is already an ISR attached to
the vector. The HAL will only allow one ISR to be attached to each
vector, so it is a good idea to use this function before using
<function>HAL_INTERRUPT_ATTACH()</function>.
</para>

<PARA>
<FUNCTION>HAL_INTERRUPT_ATTACH()</FUNCTION> attaches
the ISR, data pointer and object pointer to the given
<parameter>vector</parameter>. When an interrupt occurs on this
vector the ISR is called using the C calling convention and the vector
number and data pointer are passed to it as the first and second
arguments respectively.
</PARA>

<PARA>
<FUNCTION>HAL_INTERRUPT_DETACH()</FUNCTION> detaches the ISR from the
vector.
</PARA>

<PARA>
<FUNCTION>HAL_VSR_SET()</FUNCTION> replaces the VSR attached to the
<parameter>vector</parameter> with the replacement supplied in
<parameter>vsr</parameter>. The old VSR is returned in the location
pointed to by <parameter>pvsr</parameter>.
</PARA>

<PARA>
<FUNCTION>HAL_VSR_GET()</FUNCTION> assigns
a copy of the VSR to the location pointed to by <parameter>pvsr</parameter>.
</PARA>

<para>
<function>HAL_VSR_SET_TO_ECOS_HANDLER()</function> ensures that the
VSR for a specific exception is pointing at the eCos exception VSR and
not one for RedBoot or some other ROM monitor. The default when
running under RedBoot is for exceptions to be handled by RedBoot and
passed to GDB. This macro diverts the exception to eCos so that it may
be handled by application code. The arguments are the VSR vector to be
replaces, and a location in which to store the old VSR pointer, so
that it may be replaced at a later point.
</para>

</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Interrupt controller management</TITLE>

<PROGRAMLISTING>
HAL_INTERRUPT_MASK( vector )
HAL_INTERRUPT_UNMASK( vector )
HAL_INTERRUPT_ACKNOWLEDGE( vector )
HAL_INTERRUPT_CONFIGURE( vector, level, up )
HAL_INTERRUPT_SET_LEVEL( vector, level )
</PROGRAMLISTING>

<PARA>
These macros exert control over any prioritized interrupt
controller that is present. If no priority controller exists, then
these macros should be empty.
</para>

<note>
  <para>
  These macros may not be reentrant, so care should be taken to
  prevent them being called while interrupts are enabled. This means
  that they can be safely used in initialization code before
  interrupts are enabled, and in ISRs. In DSRs, ASRs and thread code,
  however, interrupts must be disabled before these macros are
  called. Here is an example for use in a DSR where the interrupt
  source is unmasked after data processing:
  </para>
<PROGRAMLISTING>
 ...
 HAL_DISABLE_INTERRUPTS(old);
 HAL_INTERRUPT_UNMASK(CYGNUM_HAL_INTERRUPT_ETH);
 HAL_RESTORE_INTERRUPTS(old);
 ...
</PROGRAMLISTING>
</note>

<PARA>
<FUNCTION>HAL_INTERRUPT_MASK()</FUNCTION> causes the interrupt
associated with the given vector to be blocked.
</PARA>

<PARA>
<FUNCTION>HAL_INTERRUPT_UNMASK()</FUNCTION> causes the interrupt
associated with the given vector to be unblocked.
</PARA>

<PARA>
<FUNCTION>HAL_INTERRUPT_ACKNOWLEDGE()</FUNCTION> acknowledges the
current interrupt from the given vector. This is usually executed from
the ISR for this vector when it is prepared to allow further
interrupts.  Most interrupt controllers need some form of acknowledge
action before the next interrupt is allowed through. Executing this
macro may cause another interrupt to be delivered. Whether this
interrupts the current code depends on the state of the CPU interrupt
mask.
</PARA>

<PARA>
<FUNCTION>HAL_INTERRUPT_CONFIGURE()</FUNCTION> provides
control over how an interrupt signal is detected. The arguments
are:
</PARA>
<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>vector</TERM>
    <LISTITEM>
      <PARA>The interrupt vector to be configured.</PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>level</TERM>
    <LISTITEM>
      <PARA>
      Set to <varname>true</varname> if the interrupt is detected by
      level, and <varname>false</varname> if it is edge triggered.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>up</TERM>
    <LISTITEM>
      <PARA>
      If the interrupt is set to level detect, then if this is
      <VARNAME>true</VARNAME> it is detected by a high signal level,
      and if <VARNAME>false</VARNAME> by a low signal level. If the
      interrupt is set to edge triggered, then if this is
      <VARNAME>true</VARNAME> it is triggered by a rising edge and if
      <VARNAME>false</VARNAME> by a falling edge.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>

<PARA>
<FUNCTION>HAL_INTERRUPT_SET_LEVEL()</FUNCTION> provides control over
the hardware priority of the interrupt. The arguments are:
</PARA>

<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>vector</TERM>
    <LISTITEM>
      <PARA>The interrupt whose level is to be set.</PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>level</TERM>
    <LISTITEM>
      <PARA>
      The priority level to which the interrupt is to set. In some
      architectures the masking of an interrupt is achieved by
      changing its priority level. Hence this function,
      <FUNCTION>HAL_INTERRUPT_MASK()</FUNCTION> and
      <FUNCTION>HAL_INTERRUPT_UNMASK()</FUNCTION> may interfere with
      each other.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>

</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Clock control</TITLE>

<PROGRAMLISTING>
HAL_CLOCK_INITIALIZE( period )
HAL_CLOCK_RESET( vector, period )
HAL_CLOCK_READ( pvalue )
</PROGRAMLISTING>

<PARA>
These macros provide control over a clock or timer device that may be
used by the kernel to provide time-out, delay and scheduling
services. The clock is assumed to be implemented by some form of
counter that is incremented or decremented by some external source and
which raises an interrupt when it reaches a predetermined value.
</PARA>

<PARA>
<FUNCTION>HAL_CLOCK_INITIALIZE()</FUNCTION> initializes the timer
device to interrupt at the given period. The period is essentially the
value used to initialize the timer counter and must be calculated from
the timer frequency and the desired interrupt rate. The timer device
should generate an interrupt every <varname>period</varname> cycles.
</PARA>

<PARA>
<FUNCTION>HAL_CLOCK_RESET()</FUNCTION> re-initializes the timer to
provoke the next interrupt. This macro is only really necessary when
the timer device needs to be reset in some way after each interrupt.
</PARA>

<PARA>
<FUNCTION>HAL_CLOCK_READ()</FUNCTION> reads the current value of the
timer counter and puts the value in the location pointed to by
<parameter>pvalue</parameter>. The value stored will always be the
number of timer cycles since the last interrupt, and hence ranges
between zero and the initial period value. If this is a count-down
cyclic timer, some arithmetic may be necessary to generate this value.
</PARA>

</SECTION>

<!-- =================================================================== -->

<section>
<title>Microsecond Delay</title>

<programlisting width=72>
HAL_DELAY_US(us)
</programlisting>

<para>
This is an optional definition. If defined the macro implements a busy
loop delay for the given number of microseconds. This is usually
implemented by waiting for the required number of hardware timer ticks
to pass. 
</para>

<para>
This operation should normally be used when a very short delay is
needed when controlling hardware, programming FLASH devices and similar
situations where a wait/timeout loop would otherwise be used. Since it
may disable interrupts, and is implemented by busy waiting, it should
not be used in code that is sensitive to interrupt or context switch
latencies.
</para>

</section>

</SECTION>

<!-- }}} -->
<!-- {{{ Input and Output -->

<SECTION id="hal-input-and-output">
<TITLE>HAL I/O</TITLE>

<PARA>
This section contains definitions for supporting access
to device control registers in an architecture neutral
fashion.
</PARA>

<para>
These definitions are normally found in the header file
<FILENAME>cyg/hal/hal_io.h</FILENAME>.  This file itself contains
macros that are generic to the architecture. If there are variant or
platform specific IO access macros then these will be found in
<filename>cyg/hal/var_io.h</filename> and
<filename>cyg/hal/plf_io.h</filename> in the variant or platform HALs
respectively. These files are include automatically by this header, so
need not be included explicitly.
</para>

<para>
This header (or more likely <filename>cyg/hal/plf_io.h</filename>) also
defines the PCI access macros. For more information on these see <xref
linkend="pci-library-reference">.
</para>

<!-- =================================================================== -->

<SECTION>
<TITLE>Register address</TITLE>

<PROGRAMLISTING>
HAL_IO_REGISTER
</PROGRAMLISTING>

<PARA>
This type is used to store the address of an I/O register. It will
normally be a memory address, an integer port address or an offset
into an I/O space. More complex architectures may need to code an
address space plus offset pair into a single word, or may represent it
as a structure.
</PARA>

<PARA>
Values of variables and constants of this type will usually be
supplied by configuration mechanisms or in target specific headers.
</PARA>

</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Register read</TITLE>

<PROGRAMLISTING>
HAL_READ_XXX( register, value )
HAL_READ_XXX_VECTOR( register, buffer, count, stride )
</PROGRAMLISTING>

<PARA>
These macros support the reading of I/O registers in various
sizes. The <replaceable>XXX</replaceable> component of the name may be
<literal>UINT8</literal>, <literal>UINT16</literal>,
<literal>UINT32</literal>.
</PARA>

<PARA>
<FUNCTION>HAL_READ_XXX()</FUNCTION> reads the appropriately sized
value from the register and stores it in the variable passed as the
second argument.
</PARA>

<PARA>
<FUNCTION>HAL_READ_XXX_VECTOR()</FUNCTION> reads
<parameter>count</parameter> values of the appropriate size into
<parameter>buffer</parameter>. The <parameter>stride</parameter>
controls how the pointer advances through the register space. A stride
of zero will read the same register repeatedly, and a stride of one
will read adjacent registers of the given size. Greater strides will
step by larger amounts, to allow for sparsely mapped registers for
example.</PARA>
</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Register write</TITLE>

<PROGRAMLISTING>
HAL_WRITE_XXX( register, value )
HAL_WRITE_XXX_VECTOR( register, buffer,count, stride )
</PROGRAMLISTING>

<PARA>
These macros support the writing of I/O registers in various
sizes. The <replaceable>XXX</replaceable> component of the name may be
<LITERAL>UINT8</LITERAL>, <LITERAL>UINT16</LITERAL>,
<LITERAL>UINT32</LITERAL>.
</PARA>

<PARA>
<FUNCTION>HAL_WRITE_XXX()</FUNCTION> writes
the appropriately sized value from the variable passed as the second argument
stored it in the register.</PARA>
<PARA><FUNCTION>HAL_WRITE_XXX_VECTOR()</FUNCTION> writes
<parameter>count</parameter> values of the appropriate size from <parameter>
buffer</parameter>. The <parameter>stride</parameter> controls
how the pointer advances through the register space. A stride of
zero will write the same register repeatedly, and a stride of one
will write adjacent registers of the given size. Greater strides
will step by larger amounts, to allow for sparsely mapped registers
for example.</PARA>
</SECTION>
</SECTION>

<!-- }}} -->
<!-- {{{ Cache Control -->

<SECTION id="hal-cache-control">
<TITLE>Cache Control</TITLE>

<PARA>This section contains definitions for supporting control
of the caches on the CPU.
</PARA>

<para>
These definitions are usually found in the header file
<FILENAME>cyg/hal/hal_cache.h</FILENAME>.  This file may be defined in
the architecture, variant or platform HAL, depending on where the
caches are implemented for the target. Often there will be a generic
implementation of the cache control macros in the architecture HAL
with the ability to override or undefine them in the variant or
platform HAL. Even when the implementation of the cache macros is in
the architecture HAL, the cache dimensions will be defined in the
variant or platform HAL. As with other files, the variant or platform
specific definitions are usually found in
<filename>cyg/hal/var_cache.h</filename> and
<filename>cyg/hal/plf_cache.h</filename> respectively.  These files
are include automatically by this header, so need not be included
explicitly.
</para>

<PARA>
There are versions of the macros defined here for both the Data and
Instruction caches. these are distinguished by the use of either
<literal>DCACHE</literal> or <literal>ICACHE</literal> in the macro
names. Some architectures have a unified cache, where both data and
instruction share the same cache. In these cases the control macros
use <literal>UCACHE</literal> and the <literal>DCACHE</literal> and
<literal>ICACHE</literal> macros will just be calls to the
<literal>UCACHE</literal> version. In the following descriptions,
<literal>XCACHE</literal> is used to stand for any of these. Where
there are issues specific to a particular cache, this will be
explained in the text.
</PARA>

<PARA>
There might be target specific restrictions on the use of some of the
macros which it is the user's responsibility to comply with. Such
restrictions are documented in the header file with the macro
definition.
</PARA>

<PARA>
Note that destructive cache macros should be used with caution.
Preceding a cache invalidation with a cache synchronization is not
safe in itself since an interrupt may happen after the synchronization
but before the invalidation. This might cause the state of dirty data
lines created during the interrupt to be lost.
</PARA>

<PARA>
Depending on the architecture's capabilities, it may be possible to
temporarily disable the cache while doing the synchronization and
invalidation which solves the problem (no new data would be cached
during an interrupt). Otherwise it is necessary to disable interrupts
while manipulating the cache which may take a long time.
</PARA>

<PARA>
Some platform HALs now support a pair of cache state query
macros: <function>HAL_ICACHE_IS_ENABLED( x )</function> and
<function>HAL_DCACHE_IS_ENABLED( x )</function> which set the argument
to true if the instruction or data cache is enabled,
respectively. Like most cache control macros, these are optional,
because the capabilities of different targets and boards can vary
considerably. Code which uses them, if it is to be considered
portable, should test for their existence first by means of
<literal>#ifdef</literal>.  Be sure to include
<filename>&lt;cyg/hal/hal_cache.h&gt;</filename> in order to do this
test and (maybe) use the macros.
</PARA>

<!-- =================================================================== -->

<SECTION>
<TITLE>Cache Dimensions</TITLE>

<PROGRAMLISTING>
HAL_XCACHE_SIZE
HAL_XCACHE_LINE_SIZE
HAL_XCACHE_WAYS
HAL_XCACHE_SETS
</PROGRAMLISTING>
<PARA>
These macros define the size and dimensions of the Instruction
and Data caches.
</PARA>

<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>HAL_XCACHE_SIZE 	</TERM>
    <LISTITEM>
      <PARA>Defines the total size of the cache in bytes.</PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_XCACHE_LINE_SIZE 	</TERM>
    <LISTITEM>
      <PARA>Defines the cache line size in bytes.</PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_XCACHE_WAYS 	</TERM>
    <LISTITEM>
      <PARA>
      Defines the number of ways in each set and defines its level
      of associativity. This would be 1 for a direct mapped
      cache, 2 for a 2-way cache, 4 for 4-way and so on.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_XCACHE_SETS 	</TERM>
    <LISTITEM>
      <PARA>
      Defines the number of sets in the cache, and is calculated from
      the previous values.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>

</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Global Cache Control</TITLE>

<PROGRAMLISTING>
HAL_XCACHE_ENABLE()
HAL_XCACHE_DISABLE()
HAL_XCACHE_INVALIDATE_ALL()
HAL_XCACHE_SYNC()
HAL_XCACHE_BURST_SIZE( size )
HAL_DCACHE_WRITE_MODE( mode )
HAL_XCACHE_LOCK( base, size )
HAL_XCACHE_UNLOCK( base, size )
HAL_XCACHE_UNLOCK_ALL()
</PROGRAMLISTING>

<PARA>
These macros affect the state of the entire cache, or a large part of
it.
</PARA>

<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>HAL_XCACHE_ENABLE() and HAL_XCACHE_DISABLE()</TERM>
    <LISTITEM>
      <PARA>Enable and disable the cache.</PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_XCACHE_INVALIDATE_ALL()</TERM>
    <LISTITEM>
      <PARA>
      Causes the entire contents of the cache to be invalidated.
      Depending on the hardware, this may require the cache to be disabled
      during the invalidation process. If so, the implementation must
      use <function>HAL_XCACHE_IS_ENABLED()</function> to save and
      restore the previous state.
      </PARA>
      <note>
        <para>
	If this macro is called after
	<function>HAL_XCACHE_SYNC()</function> with the intention of clearing
	the cache (invalidating the cache after writing dirty data back to
	memory), you must prevent interrupts from happening between the two
	calls:
	</para>
<PROGRAMLISTING>
 ...
 HAL_DISABLE_INTERRUPTS(old);
 HAL_XCACHE_SYNC();
 HAL_XCACHE_INVALIDATE_ALL();
 HAL_RESTORE_INTERRUPTS(old);
 ...
</PROGRAMLISTING>
        <para>
	Since the operation may take a very long time, real-time
	responsiveness could be affected, so only do this when it is
	absolutely required and you know the delay will not interfere
	with the operation of drivers or the application.
	</para>
      </note>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_XCACHE_SYNC()</TERM>
    <LISTITEM>
      <PARA>
      Causes the contents of the cache to be brought into synchronization
      with the contents of memory. In some implementations this may be
      equivalent to <function>HAL_XCACHE_INVALIDATE_ALL()</function>.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_XCACHE_BURST_SIZE()</TERM>
    <LISTITEM>
      <PARA>
      Allows the size of cache to/from memory bursts to
      be controlled. This macro will only be defined if this functionality
      is available.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_DCACHE_WRITE_MODE()</TERM>
    <LISTITEM>
      <PARA>
      Controls the way in which data cache lines are written back to
      memory. There will be definitions for the possible
      modes. Typical definitions are
      <literal>HAL_DCACHE_WRITEBACK_MODE</literal> and
      <literal>HAL_DCACHE_WRITETHRU_MODE</literal>. This macro will
      only be defined if this functionality is available.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_XCACHE_LOCK()</TERM>
    <LISTITEM>
      <PARA>
      Causes data to be locked into the cache. The base and size
      arguments define the memory region that will be locked into the
      cache. It is architecture dependent whether more than one locked
      region is allowed at any one time, and whether this operation
      causes the cache to cease acting as a cache for addresses
      outside the region during the duration of the lock. This macro
      will only be defined if this functionality is available.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_XCACHE_UNLOCK()</TERM>
    <LISTITEM>
      <PARA>
      Cancels the locking of the memory region given. This should
      normally correspond to a region supplied in a matching lock
      call.  This macro will only be defined if this functionality is
      available.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_XCACHE_UNLOCK_ALL()</TERM>
    <LISTITEM>
      <PARA>
      Cancels all existing locked memory regions. This may be required
      as part of the cache initialization on some architectures. This
      macro will only be defined if this functionality is available.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>

</SECTION>

<!-- =================================================================== -->

<SECTION>
<TITLE>Cache Line Control</TITLE>

<PROGRAMLISTING>
HAL_DCACHE_ALLOCATE( base , size )
HAL_DCACHE_FLUSH( base , size )
HAL_XCACHE_INVALIDATE( base , size )
HAL_DCACHE_STORE( base , size )
HAL_DCACHE_READ_HINT( base , size )
HAL_DCACHE_WRITE_HINT( base , size )
HAL_DCACHE_ZERO( base , size )
</PROGRAMLISTING>

<PARA>
All of these macros apply a cache operation to all cache lines that
match the memory address region defined by the base and size
arguments. These macros will only be defined if the described
functionality is available. Also, it is not guaranteed that the cache
function will only be applied to just the described regions, in some
architectures it may be applied to the whole cache.
</PARA>

<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>HAL_DCACHE_ALLOCATE()</TERM>
    <LISTITEM>
      <PARA>
      Allocates lines in the cache for the given region without
      reading their contents from memory, hence the contents of the lines
      is undefined. This is useful for preallocating lines which are to
      be completely overwritten, for example in a block copy
      operation.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_DCACHE_FLUSH()</TERM>
    <LISTITEM>
      <PARA>
      Invalidates all cache lines in the region after writing any
      dirty lines to memory.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_XCACHE_INVALIDATE() </TERM>
    <LISTITEM>
      <PARA>
      Invalidates all cache lines in the region. Any dirty lines
      are invalidated without being written to memory.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_DCACHE_STORE() </TERM>
    <LISTITEM>
      <PARA>
      Writes all dirty lines in the region to memory, but does not
      invalidate any lines.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_DCACHE_READ_HINT() </TERM>
    <LISTITEM>
      <PARA>
      Hints to the cache that the region is going to be read from
      in the near future. This may cause the region to be speculatively
      read into the cache.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>HAL_DCACHE_WRITE_HINT() </TERM>
    <LISTITEM>
      <PARA>
      Hints to the cache that the region is going to be written
      to in the near future. This may have the identical behavior to
      HAL_DCACHE_READ_HINT().
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>HAL_DCACHE_ZERO()</TERM>
    <LISTITEM>
      <PARA>
      Allocates and zeroes lines in the cache for the given
      region without reading memory. This is useful if a large area of
      memory is to be cleared.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>

</SECTION>
</SECTION>

<!-- }}} -->
<!-- {{{ Linker Scripts -->

<SECTION id="hal-linker-scripts">
<TITLE><!-- <xref> -->Linker Scripts</TITLE>

<para>
When an eCos application is linked it must be done under the control
of a linker script. This script defines the memory areas, addresses
and sized, into which the code and data are to be put, and allocates
the various sections generated by the compiler to these.
</para>

<para>
The linker script actually used is in
<filename>lib/target.ld</filename> in the install directory. This is
actually manufactured out of two other files: a base linker script and
an <literal>.ldi</literal> file that was generated by the memory
layout tool.
</para>

<para>
The base linker script is usually supplied either by the architecture
HAL or the variant HAL. It consists of a set of linker script
fragments, in the form of C preprocessor macros, that define the major
output sections to be generated by the link operation. The
<literal>.ldi</literal> file, which is <literal>#include'ed</literal>
by the base linker script, uses these macro definitions to assign the
output sections to the required memory areas and link addresses.
</para>

<para>
The <literal>.ldi</literal> file is supplied by the platform HAL, and
contains knowledge of the memory layout of the target platform. These
files generally conform to a standard naming convention, each file
being of the form:
</para>
<para>
<filename>pkgconf/mlt_&lt;architecture&gt;_&lt;variant&gt;_&lt;platform&gt;_&lt;startup&gt;.ldi</filename>
</para>
<para>
where <literal>&lt;architecture&gt;</literal>,
<literal>&lt;variant&gt;</literal> and
<literal>&lt;platform&gt;</literal> are the respective HAL package
names and <literal>&lt;startup&gt;</literal> is the startup type which
is usually one of <literal>ROM</literal>, <literal>RAM</literal> or
<literal>ROMRAM</literal>.
</para>

<para>
In addition to the <literal>.ldi</literal> file, there is also a
congruously name <literal>.h</literal> file. This may be used by the
application to access information defined in the
<literal>.ldi</literal> file. Specifically it contains the memory
layout defined there, together with any additional section names
defined by the user. Examples of the latter are heap areas or PCI bus
memory access windows.
</para>

<para>
The <literal>.ldi</literal> is manufactured by the <application>Memory
Layout Tool</application> (MLT). The <application>MLT</application> saves the memory
configuration into a file named
</para>
<para>
<filename>include/pkgconf/mlt_&lt;architecture&gt;_&lt;variant&gt;_&lt;platform&gt;_&lt;startup&gt;.mlt</filename>
</para>
<para>
in the platform HAL. This file is used by the
<application>MLT</application> to manufacture both the
<literal>.ldi</literal> and <literal>.h</literal> files. Users should
beware that direct edits the either of these files may be overwritten
if the <application>MLT</application> is run and regenerates them from the
<literal>.mlt</literal> file.
</para>

<para>
The names of the <literal>.ldi</literal> and <literal>.h</literal>
files are defined by macro definitions in
<FILENAME>pkgconf/system.h</FILENAME>. These are
<literal>CYGHWR_MEMORY_LAYOUT_LDI</literal> and
<literal>CYGHWR_MEMORY_LAYOUT_H</literal> respectively. While there
will be little need for the application to refer to the
<literal>.ldi</literal> file directly, it may include the
<literal>.h</literal> file as follows:
</para>

<programlisting>
#include CYGHWR_MEMORY_LAYOUT_H
</programlisting>

</SECTION>

<!-- }}} -->
<!-- {{{ Diagnostic Support -->

<SECTION id="hal-diagnostic-support">
<TITLE>Diagnostic Support</TITLE>

<para>
The HAL provides support for low level diagnostic IO. This is
particularly useful during early development as an aid to bringing up
a new platform. Usually this diagnostic channel is a UART or some
other serial IO device, but it may equally be a a memory
buffer, a simulator supported output channel, a ROM emulator virtual
UART, and LCD panel, a memory mapped video buffer or any other output
device.
</para>

<PARA>
<FUNCTION>HAL_DIAG_INIT()</FUNCTION> performs any initialization
required on the device being used to generate diagnostic output. This
may include, for a UART, setting baud rate, and stop, parity and
character bits. For other devices it may include initializing a
controller or establishing contact with a remote device.
</PARA>

<PARA>
<FUNCTION>HAL_DIAG_WRITE_CHAR(c)</FUNCTION> writes
the character supplied to the diagnostic output device.
</PARA>

<PARA>
<FUNCTION>HAL_DIAG_READ_CHAR(c)</FUNCTION> reads a character from the
diagnostic device into the supplied variable. This is not supported
for all diagnostic devices.
</PARA>

<para>
These macros are defined in the header file
<filename>cyg/hal/hal_diag.h</filename>. This file is usually supplied
by the variant or platform HAL, depending on where the IO device being
used is located. For example for on-chip UARTs it would be in the
variant HAL, but for a board-level LCD panel it would be in the
platform HAL.
</para>

</section>

<!-- }}} -->
<!-- {{{ SMP Support -->

<section id="hal-smp-support">
<TITLE>SMP Support</TITLE>

<para>
eCos contains support for limited Symmetric Multi-Processing
(SMP). This is only available on selected architectures and platforms.
</para>

<section>
<title>Target Hardware Limitations</title>

<para>
To allow a reasonable implementation of SMP, and to reduce the
disruption to the existing source base, a number of assumptions have
been made about the features of the target hardware.
</para>

<itemizedlist>
<listitem>
<para>
Modest multiprocessing. The typical number of CPUs supported is two
to four, with an upper limit around eight. While there are no
inherent limits in the code, hardware and algorithmic limitations
will probably become significant beyond this point.
</para>
</listitem>

<listitem>
<para>
SMP synchronization support. The hardware must supply a mechanism to
allow software on two CPUs to synchronize. This is normally provided
as part of the instruction set in the form of test-and-set,
compare-and-swap or load-link/store-conditional instructions. An
alternative approach is the provision of hardware semaphore
registers which can be used to serialize implementations of these
operations. Whatever hardware facilities are available, they are
used in eCos to implement spinlocks.
</para>
</listitem>

<listitem>
<para>
Coherent caches. It is assumed that no extra effort will be required
to access shared memory from any processor. This means that either
there are no caches, they are shared by all processors, or are
maintained in a coherent state by the hardware. It would be too
disruptive to the eCos sources if every memory access had to be
bracketed by cache load/flush operations. Any hardware that requires
this is not supported.
</para>
</listitem>

<listitem>
<para>
Uniform addressing. It is assumed that all memory that is
shared between CPUs is addressed at the same location from all
CPUs. Like non-coherent caches, dealing with CPU-specific address
translation is considered too disruptive to the eCos source
base. This does not, however, preclude systems with non-uniform
access costs for different CPUs.
</para>
</listitem>

<listitem>
<para>
Uniform device addressing. As with access to memory, it is assumed
that all devices are equally accessible to all CPUs. Since device
access is often made from thread contexts, it is not possible to
restrict access to device control registers to certain CPUs, since
there is currently no support for binding or migrating threads to CPUs.
</para>
</listitem>

<listitem>
<para>
Interrupt routing. The target hardware must have an interrupt
controller that can route interrupts to specific CPUs. It is
acceptable for all interrupts to be delivered to just one CPU, or
for some interrupts to be bound to specific CPUs, or for some
interrupts to be local to each CPU. At present dynamic routing,
where a different CPU may be chosen each time an interrupt is
delivered, is not supported. ECos cannot support hardware where all
interrupts are delivered to all CPUs simultaneously with the
expectation that software will resolve any conflicts.
</para>
</listitem>

<listitem>
<para>
Inter-CPU interrupts. A mechanism to allow one CPU to interrupt
another is needed. This is necessary so that events on one CPU can
cause rescheduling on other CPUs.
</para>
</listitem>

<listitem>
<para>
CPU Identifiers. Code running on a CPU must be able to determine
which CPU it is running on. The CPU Id is usually provided either in
a CPU status register, or in a register associated with the
inter-CPU interrupt delivery subsystem. ECos expects CPU Ids to be
small positive integers, although alternative representations, such
as bitmaps, can be converted relatively easily. Complex mechanisms
for getting the CPU Id cannot be supported. Getting the CPU Id must
be a cheap operation, since it is done often, and in performance
critical places such as interrupt handlers and the scheduler.
</para>
</listitem>
</itemizedlist>

</section>  

<section>
<title>HAL Support</title>

<para>
SMP support in any platform depends on the HAL supplying the
appropriate operations. All HAL SMP support is defined in the
<filename>cyg/hal/hal_smp.h</filename> header. Variant and platform
specific definitions will be in <filename>cyg/hal/var_smp.h</filename>
and <filename>cyg/hal/plf_smp.h</filename> respectively. These files
are include automatically by this header, so need not be included
explicitly.
</para>

<para>
SMP support falls into a number of functional groups.
</para>

<section>
<title>CPU Control</title>

<para>
This group consists of descriptive and control macros for managing the
CPUs in an SMP system.
</para>

<variablelist>
<varlistentry>
<term><literal>HAL_SMP_CPU_TYPE</literal></term>
<listitem>
<para>
A type that can contain a CPU id. A CPU id is
usually a small integer that is used to index
arrays of variables that are managed on an
per-CPU basis.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_CPU_MAX</literal></term>
<listitem>
<para>
The maximum number of CPUs that can be
supported. This is used to provide the size of
any arrays that have an element per CPU.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_CPU_COUNT()</literal></term>
<listitem>
<para>
Returns the number of CPUs currently
operational. This may differ from
HAL_SMP_CPU_MAX depending on the runtime
environment.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_CPU_THIS()</literal></term>
<listitem>
<para>
Returns the CPU id of the current CPU.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_CPU_NONE</literal></term>
<listitem>
<para>
A value that does not match any real CPU
id. This is uses where a CPU type variable
must be set to a null value.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_CPU_START( cpu )</literal></term>
<listitem>
<para>
Starts the given CPU executing at a defined
HAL entry point. After performing any HAL
level initialization, the CPU calls up into
the kernel at <function>cyg_kernel_cpu_startup()</function>.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_CPU_RESCHEDULE_INTERRUPT( cpu, wait )</literal></term>
<listitem>
<para>
Sends the CPU a reschedule interrupt, and if
<parameter>wait</parameter> is non-zero, waits for an
acknowledgment. The interrupted CPU should call
<function>cyg_scheduler_set_need_reschedule()</function> in its DSR to
cause the reschedule to occur.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_CPU_TIMESLICE_INTERRUPT( cpu, wait )</literal></term>
<listitem>
<para>
Sends the CPU a timeslice interrupt, and if
<parameter>wait</parameter> is non-zero, waits for an
acknowledgment. The interrupted CPU should call
<function>cyg_scheduler_timeslice_cpu()</function> to cause the
timeslice event to be processed.
</para>
</listitem>
</varlistentry>
</variablelist>
</section>


<section>
<title>Test-and-set Support</title>

<para>
Test-and-set is the foundation of the SMP synchronization
mechanisms.
</para>

<variablelist>
<varlistentry>
<term><literal>HAL_TAS_TYPE</literal></term>
<listitem>
<para>
The type for all test-and-set variables. The
test-and-set macros only support operations on
a single bit (usually the least significant
bit) of this location. This allows for maximum
flexibility in the implementation.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_TAS_SET( tas, oldb )</literal></term>
<listitem>
<para>
Performs a test and set operation on the
location <parameter>tas</parameter>. <parameter>oldb</parameter> will contain <literal>true</literal> if
the location was already set, and <literal>false</literal> if
it was clear.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_TAS_CLEAR( tas, oldb )</literal></term>
<listitem>
<para>
Performs a test and clear operation on the
location <parameter>tas</parameter>. <parameter>oldb</parameter> will contain <literal>true</literal> if
the location was already set, and <literal>false</literal> if
it was clear.
</para>
</listitem>
</varlistentry>
</variablelist>
</section>
<section>

<title>Spinlocks</title>

<para>
Spinlocks provide inter-CPU locking. Normally they will be implemented
on top of the test-and-set mechanism above, but may also be
implemented by other means if, for example, the hardware has more
direct support for spinlocks.
</para>

<variablelist>
<varlistentry>
<term><literal>HAL_SPINLOCK_TYPE</literal></term>
<listitem>
<para>
The type for all spinlock variables.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SPINLOCK_INIT_CLEAR</literal></term>
<listitem>
<para>
A value that may be assigned to a spinlock
variable to initialize it to clear.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SPINLOCK_INIT_SET</literal></term>
<listitem>
<para>
A value that may be assigned to a spinlock
variable to initialize it to set.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SPINLOCK_SPIN( lock )</literal></term>
<listitem>
<para>
The caller spins in a busy loop waiting for
the lock to become clear. It then sets it and
continues. This is all handled atomically, so
that there are no race conditions between CPUs.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SPINLOCK_CLEAR( lock )</literal></term>
<listitem>
<para>
The caller clears the lock. One of any waiting
spinners will then be able to proceed.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SPINLOCK_TRY( lock, val )</literal></term>
<listitem>
<para>
Attempts to set the lock. The value put in
<parameter>val</parameter> will be <literal>true</literal> if the lock was
claimed successfully, and <literal>false</literal> if it was
not.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SPINLOCK_TEST( lock, val )</literal></term>
<listitem>
<para>
Tests the current value of the lock. The value
put in <parameter>val</parameter> will be <literal>true</literal> if the lock is
claimed and <literal>false</literal> of it is clear.
</para>
</listitem>
</varlistentry>
</variablelist>
</section>
<section>

<title>Scheduler Lock</title>

<para>
The scheduler lock is the main protection for all kernel data
structures. By default the kernel implements the scheduler lock itself
using a spinlock. However, if spinlocks cannot be supported by the
hardware, or there is a more efficient implementation available, the
HAL may provide macros to implement the scheduler lock.
</para>

<variablelist>
<varlistentry>
<term><literal>HAL_SMP_SCHEDLOCK_DATA_TYPE</literal></term>
<listitem>
<para>
A data type, possibly a structure, that
contains any data items needed by the
scheduler lock implementation. A variable of
this type will be instantiated as a static
member of the Cyg_Scheduler_SchedLock class
and passed to all the following macros.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_SCHEDLOCK_INIT( lock, data )</literal></term>
<listitem>
<para>
Initialize the scheduler lock. The <parameter>lock</parameter>
argument is the scheduler lock counter and the
<parameter>data</parameter> argument is a variable of
HAL_SMP_SCHEDLOCK_DATA_TYPE type.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_SCHEDLOCK_INC( lock, data )</literal></term>
<listitem>
<para>
Increment the scheduler lock. The first
increment of the lock from zero to one for any
CPU may cause it to wait until the lock is
zeroed by another CPU. Subsequent increments
should be less expensive since this CPU
already holds the lock.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_SCHEDLOCK_ZERO( lock, data )</literal></term>
<listitem>
<para>
Zero the scheduler lock. This operation will
also clear the lock so that other CPUs may
claim it.
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_SMP_SCHEDLOCK_SET( lock, data, new )</literal></term>
<listitem>
<para>
Set the lock to a different value, in
<parameter>new</parameter>. This is only called when the lock is
already known to be owned by the current CPU. It is never called to
zero the lock, or to increment it from zero.
</para>
</listitem>
</varlistentry>

</variablelist>
</section>
<section>

<title>Interrupt Routing</title>

<para>
The routing of interrupts to different CPUs is supported by two new
interfaces in hal_intr.h.
</para>

<para>
Once an interrupt has been routed to a new CPU, the existing vector
masking and configuration operations should take account of the CPU
routing. For example, if the operation is not invoked on the
destination CPU itself, then the HAL may need to arrange to transfer
the operation to the destination CPU for correct application.
</para>

<variablelist>
<varlistentry>
<term><literal>HAL_INTERRUPT_SET_CPU( vector, cpu )</literal></term>
<listitem>
<para>
Route the interrupt for the given <parameter>vector</parameter> to
the given <parameter>cpu</parameter>. 
</para>
</listitem>
</varlistentry>

<varlistentry>
<term><literal>HAL_INTERRUPT_GET_CPU( vector, cpu )</literal></term>
<listitem>
<para>
Set <parameter>cpu</parameter> to the id of the CPU to which this
vector is routed.
</para>
</listitem>
</varlistentry>
</variablelist>

</section>

</section>


</section>

<!-- }}} -->

</CHAPTER>

<!-- }}} -->
<!-- {{{ Exception Handling -->

<CHAPTER id="hal-exception-handling">
<TITLE>Exception Handling</TITLE>

<!-- {{{ Intro -->

<para>
Most of the HAL consists of simple macros or functions that are
called via the interfaces described in the previous section. These
just perform whatever operation is required by accessing the hardware
and then return. The exception to this is the handling of exceptions:
either synchronous hardware traps or asynchronous device
interrupts. Here control is passed first to the HAL, which then passed
it on to eCos or the application. After eCos has finished with it,
control is then passed back to the HAL for it to tidy up the CPU state
and resume processing from the point at which the exception occurred.
</para>

<PARA>
The HAL exceptions handling code is usually found in the file
<FILENAME>vectors.S</FILENAME> in the architecture HAL.  Since the
reset entry point is usually implemented as one of these it also deals
with system startup.
</PARA>

<PARA>
The exact implementation of this code is under the control of the HAL
implementer. So long as it interacts correctly with the interfaces
defined previously it may take any form.  However, all current
implementation follow the same pattern, and there should be a very
good reason to break with this. The rest of this section describes
these operate.
</PARA>

<para>
Exception handling normally deals with the following broad areas of
functionality:
</para>

<ITEMIZEDLIST>
  <LISTITEM>
    <PARA>Startup and initialization.</PARA>
  </LISTITEM>

  <LISTITEM>
    <PARA>Hardware exception delivery.</PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>Default handling of synchronous exceptions.</PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>Default handling of asynchronous interrupts.</PARA>
  </LISTITEM>
</ITEMIZEDLIST>

<!-- }}} -->
<!-- {{{ HAL Startup -->

<SECTION id="hal-startup">
<TITLE><!-- <index></index> --><!-- <xref> -->HAL Startup</TITLE>

<PARA>
Execution normally begins at the reset vector with
the machine in a minimal startup state. From here the HAL needs to get
the machine running, set up the execution environment for the
application, and finally invoke its entry point.
</PARA>

<PARA>
The following is a list of the jobs that need to be done in
approximately the order in which they should be accomplished. Many
of these will not be needed in some configurations.
</PARA>

<ITEMIZEDLIST>
  <listitem>
  <para>
  Initialize the hardware. This may involve initializing several
  subsystems in both the architecture, variant and platform
  HALs. These include:
  </para>
    <itemizedlist>
      <LISTITEM>
	<PARA>
	Initialize various CPU status registers. Most importantly, the CPU
	interrupt mask should be set to disable interrupts.
	</PARA>
      </LISTITEM>

      <LISTITEM>
	<PARA>
	Initialize the MMU, if it is used. On many platforms it is
	only possible to control the cacheability of address ranges
	via the MMU. Also, it may be necessary to remap RAM and device
	registers to locations other than their defaults. However, for
	simplicity, the mapping should be kept as close to one-to-one
	physical-to-virtual as possible.
	</PARA>
      </LISTITEM>

      <LISTITEM>
	<PARA>
	Set up the memory controller to access RAM, ROM and I/O devices
	correctly. Until this is done it may not be possible to access
	RAM. If this is a ROMRAM startup then the program code can
	now be copied to its RAM address and control transferred to it.
	</PARA>
      </LISTITEM>

      <LISTITEM>
	<PARA>
	Set up any bus bridges and support chips. Often access to
	device registers needs to go through various bus bridges and
	other intermediary devices. In many systems these are combined
	with the memory controller, so it makes sense to set these up
	together. This is particularly important if early diagnostic
	output needs to go through one of these devices.
	</PARA>
      </LISTITEM>

      <LISTITEM>
	<PARA>
	Set up diagnostic mechanisms. If the platform includes an LED or
	LCD output device, it often makes sense to output progress
	indications on this during startup. This helps with diagnosing
	hardware and software errors.
	</PARA>
      </LISTITEM>

      <LISTITEM>
	<PARA>
	Initialize floating point and other extensions such as SIMD
	and multimedia engines. It is usually necessary to enable
	these and maybe initialize control and exception registers for
	these extensions.
	</PARA>
      </LISTITEM>


      <LISTITEM>
	<PARA>
	Initialize interrupt controller. At the very least, it should
	be configured to mask all interrupts. It may also be necessary
	to set up the mapping from the interrupt controller's vector
	number space to the CPU's exception number space. Similar
	mappings may need to be set up between primary and secondary
	interrupt controllers.
	</PARA>
      </LISTITEM>
      
      <LISTITEM>
	<PARA>
	Disable and initialize the caches. The caches should not
	normally be enabled at this point, but it may be necessary to
	clear or initialize them so that they can be enabled
	later. Some architectures require that the caches be
	explicitly reinitialized after a power-on reset.
	</PARA>
      </LISTITEM>


      <LISTITEM>
	<PARA>
	Initialize the timer, clock etc. While the timer used for RTC
	interrupts will be initialized later, it may be necessary to
	set up the clocks that drive it here.
	</PARA>
      </LISTITEM>
      
    </itemizedlist>
    <para>
    The exact order in which these initializations is done is
    architecture or variant specific. It is also often not necessary
    to do anything at all for some of these options. These fragments
    of code should concentrate on getting the target up and running so
    that C function calls can be made and code can be run. More
    complex initializations that cannot be done in assembly code may
    be postponed until calls to
    <function>hal_variant_init()</function> or
    <function>hal_platform_init()</function> are made.
    </para>

    <para>
    Not all of these initializations need to be done for all startup
    types. In particular, RAM startups can reasonably assume that the
    ROM monitor or loader has already done most of this work.
    </para>
    
  </listitem>
  
  <LISTITEM>
    <PARA>
    Set up the stack pointer, this allows subsequent initialization
    code to make proper procedure calls. Usually the interrupt stack
    is used for this purpose since it is available, large enough, and
    will be reused for other purposes later.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Initialize any global pointer register needed for access to
    globally defined variables. This allows subsequent initialization
    code to access global variables.
    </PARA>
  </LISTITEM>

  <LISTITEM>
    <PARA>
    If the system is starting from ROM, copy the ROM template of the 
    <filename>.data</filename> section out to its correct position in
    RAM. (<xref linkend="hal-linker-scripts">).
    </PARA>
  </LISTITEM>

  <LISTITEM>
    <PARA>
    Zero the <filename>.bss</filename> section.
    </PARA>
  </LISTITEM>
      
  <LISTITEM>
    <PARA>
    Create a suitable C call stack frame. This may involve making
    stack space for call frames, and arguments, and initializing the
    back pointers to halt a GDB backtrace operation.
    </PARA>
  </LISTITEM>

  <LISTITEM>
    <PARA>
    Call <function>hal_variant_init()</function> and
    <function>hal_platform_init()</function>. These will perform any
    additional initialization needed by the variant and platform. This
    typically includes further initialization of the interrupt
    controller, PCI bus bridges, basic IO devices and enabling the
    caches.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Call <FUNCTION>cyg_hal_invoke_constructors()</FUNCTION> to run any
    static constructors.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Call <FUNCTION>cyg_start()</FUNCTION>. If
    <FUNCTION>cyg_start()</FUNCTION> returns, drop into an infinite
    loop.
    </PARA>
  </LISTITEM>
</ITEMIZEDLIST>

</SECTION>

<!-- }}} -->
<!-- {{{ Vectors and VSRs -->

<SECTION id="hal-vectors-and-vsrs"><!-- <index></index> -->
<TITLE>Vectors and VSRs</TITLE>

<PARA>
The CPU delivers all <!-- <index></index> --> exceptions, whether
synchronous faults or asynchronous interrupts, to a set of hardware
defined vectors. Depending on the architecture, these may be
implemented in a number of different ways. Examples of existing
mechanisms are:
</PARA>

<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>PowerPC</TERM>
    <LISTITEM>
      <PARA>
      Exceptions are vectored to locations 256 bytes apart starting at
      either zero or <literal>0xFFF00000</literal>. There are 16 such
      vectors defined by the basic architecture and extra vectors may
      be defined by specific variants. One of the base vectors is for
      all external interrupts, and another is for the architecture
      defined timer.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>MIPS</TERM>
    <LISTITEM>
      <PARA>
      Most exceptions and all interrupts are vectored to a single
      address at either <literal>0x80000000</literal> or
      <literal>0xBFC00180</literal>. Software is responsible for
      reading the exception code from the CPU <literal>cause</literal>
      register to discover its true source. Some TLB and debug
      exceptions are delivered to different vector addresses, but
      these are not used currently by eCos. One of the exception codes
      in the <literal>cause</literal> register indicates an external
      interrupt. Additional bits in the <literal>cause</literal>
      register provide a first-level decode for the interrupt source,
      one of which represents an architecture defined timer.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>IA32</TERM>
    <LISTITEM>
      <PARA>
      Exceptions are delivered via an Interrupt Descriptor Table (IDT)
      which is essentially an indirection table indexed by exception
      number. The IDT may be placed anywhere in memory. In PC hardware
      the standard interrupt controller can be programmed to deliver
      the external interrupts to a block of 16 vectors at any offset
      in the IDT. There is no hardware supplied mechanism for
      determining the vector taken, other than from the address jumped
      to.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>ARM</TERM>
    <LISTITEM>
      <PARA>
      All exceptions, including the FIQ and IRQ interrupts, are
      vectored to locations four bytes apart starting at zero. There
      is only room for one instruction here, which must immediately
      jump out to handling code higher in memory. Interrupt sources
      have to be decoded entirely from the interrupt controller.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>


<para>
With such a wide variety of hardware approaches, it is not possible to
provide a generic mechanism for the substitution of exception vectors
directly. Therefore, eCos translates all of these mechanisms in to a
common approach that can be used by portable code on all platforms.
</para>

<para>
The mechanism implemented is to attach to each hardware vector a short
piece of trampoline code that makes an indirect jump via a table to
the actual handler for the exception. This handler is called the
Vector Service Routine (VSR) and the table is called the VSR table.
</para>

<para>
The trampoline code performs the absolute minimum processing necessary
to identify the exception source, and jump to the VSR. The VSR is then
responsible for saving the CPU state and taking the necessary actions
to handle the exception or interrupt. The entry conditions for the VSR
are as close to the raw hardware exception entry state as possible -
although on some platforms the trampoline will have had to move or
reorganize some registers to do its job.
</para>

<para>
To make this more concrete, consider how the trampoline code operates
in each of the architectures described above:
</para>


<VARIABLELIST>
  <VARLISTENTRY>
    <TERM>PowerPC</TERM>
    <LISTITEM>
      <PARA>
      A separate trampoline is contained in each of the vector
      locations. This code saves a few work registers away to the
      special purposes registers available, loads the exception number
      into a register and then uses that to index the VSR table and
      jump to the VSR. The VSR is entered with some registers move to
      the SPRs, and one of the data register containing the number of
      the vector taken.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
  
  <VARLISTENTRY>
    <TERM>MIPS</TERM>
    <LISTITEM>
      <PARA>
      A single trampoline routine attached to the common vector reads
      the exception code out of the <literal>cause</literal> register
      and uses that value to index the VSR table and jump to the VSR.
      The trampoline uses the two registers defined in the ABI for
      kernel use to do this, one of these will contain the exception
      vector number for the VSR.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>

  <VARLISTENTRY>
    <TERM>IA32</TERM>
    <LISTITEM>
      <PARA>
      There is a separate 3 or 4 instruction trampoline pointed to by
      each active IDT table entry. The trampoline for exceptions that
      also have an error code pop it from the stack and put it into a
      memory location. Trampolines for non-error-code exceptions just
      zero the memory location. Then all trampolines push an
      interrupt/exception number onto the stack, and take an indirect
      jump through a precalculated offset in the VSR table. This is
      all done without saving any registers, using memory-only
      operations. The VSR is entered with the vector number pushed
      onto the stack on top of the standard hardware saved state.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
    
  <VARLISTENTRY>
    <TERM>ARM</TERM>
    <LISTITEM>
      <PARA>
      The trampoline consists solely of the single instruction at the
      exception entry point. This is an indirect jump via a location
      32 bytes higher in memory. These locations, from
      <literal>0x20</literal> up, form the VSR table. Since each VSR
      is entered in a different CPU mode
      (<literal>SVC,UNDEF,ABORT,IRQ or FIQ</literal>) there has to be a
      different VSR for each exception that knows how to save the CPU
      state correctly.
      </PARA>
    </LISTITEM>
  </VARLISTENTRY>
</VARIABLELIST>

</section>

<!-- }}} -->
<!-- {{{ Synchronous Exception Handling -->

<SECTION id="hal-default-synchronous-exception-handling">
<TITLE><!-- <index></index> -->Default Synchronous Exception Handling</TITLE>

<PARA>
Most synchronous exception VSR table entries will point to a default
exception VSR which is responsible for handling all exceptions in a
generic manner. The default VSR simply saves the CPU state, makes any
adjustments to the CPU state that is necessary, and calls
<function>cyg_hal_exception_handler()</function>.
</PARA>

<PARA>
<function>cyg_hal_exception_handler()</function> needs to pass the
exception on to some handling code.  There are two basic destinations:
enter GDB or pass the exception up to eCos. Exactly which
destination is taken depends on the configuration. When the GDB stubs are
included then the exception is passed to them, otherwise it is passed
to eCos.
</PARA>

<para>
If an eCos application has been loaded by RedBoot then the VSR table
entries will all point into RedBoot's exception VSR, and will
therefore enter GDB if an exception occurs. If the eCos application
wants to handle an exception itself, it needs to replace the the VSR
table entry with one pointing to its own VSR. It can do this with the
<function>HAL_VSR_SET_TO_ECOS_HANDLER()</function> macro.
</para>

</SECTION>

<!-- }}} -->
<!-- {{{ Interrupt Handling -->

<SECTION id="hal-default-interrupt-handling">
<TITLE><!-- <index></index> -->Default Interrupt Handling</TITLE>

<PARA>
Most asynchronous external interrupt vectors will point to a default
interrupt VSR which decodes the actual interrupt being delivered from
the interrupt controller and invokes the appropriate ISR.
</PARA>

<PARA>
The default interrupt VSR has a number of responsibilities if it is
going to interact with the Kernel cleanly and allow interrupts to
cause thread preemption.
</PARA>

<PARA>
To support this VSR an ISR vector table is needed. For each valid
vector three pointers need to be stored: the ISR, its data pointer and
an opaque (to the HAL) interrupt object pointer needed by the
kernel. It is implementation defined whether these are stored in a
single table of triples, or in three separate tables.
</PARA>

<PARA>
The VSR follows the following approximate plan:
</PARA>

<ORDEREDLIST>
  <LISTITEM>
    <PARA>
    Save the CPU state. In non-debug configurations, it may be
    possible to get away with saving less than the entire machine
    state. The option
    <literal>CYGDBG_HAL_COMMON_INTERRUPTS_SAVE_MINIMUM_CONTEXT</literal>
    is supported in some targets to do this.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Increment the kernel scheduler lock. This is a static member of
    the Cyg_Scheduler class, however it has also been aliased to
    <literal>cyg_scheduler_sched_lock</literal> so that it can be
    accessed from assembly code.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    (Optional) Switch to an interrupt stack if not already running on
    it. This allows nested interrupts to be delivered without needing
    every thread to have a stack large enough to take the maximum
    possible nesting. It is implementation defined how to detect
    whether this is a nested interrupt but there are two basic
    techniques. The first is to inspect the stack pointer and switch
    only if it is not currently within the interrupt stack range; the
    second is to maintain a counter of the interrupt nesting level and
    switch only if it is zero. The option
    <literal>CYGIMP_HAL_COMMON_INTERRUPTS_USE_INTERRUPT_STACK</literal>
    controls whether this happens.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Decode the actual external interrupt being delivered from
    the interrupt controller. This will yield the ISR vector
    number. The code to do this usually needs to come from the
    variant or platform HAL, so is usually present in the form of a
    macro or procedure callout.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    (Optional) Re-enable interrupts to permit nesting. At this point
    we can potentially allow higher priority interrupts to occur. It
    depends on the interrupt architecture of the CPU and platform
    whether more interrupts will occur at this point, or whether they
    will only be delivered after the current interrupt has been
    acknowledged (by a call to
    <function>HAL_INTERRUPT_ACKNOWLEDGE()</function> in the ISR).
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Using the ISR vector number as an index, retrieve the
    ISR pointer and its data pointer from the ISR vector table.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Construct a C call stack frame. This may involve making stack
    space for call frames, and arguments, and initializing the back
    pointers to halt a GDB backtrace operation.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Call the ISR, passing the vector number and data pointer.  The
    vector number and a pointer to the saved state should be preserved
    across this call, preferably by storing them in registers that are
    defined to be callee-saved by the calling conventions.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    If this is an un-nested interrupt and a separate interrupt
    stack is being used, switch back to the interrupted thread's
    own stack.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Use the saved ISR vector number to get the interrupt object
    pointer from the ISR vector table.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    Call <FUNCTION>interrupt_end()</FUNCTION> passing it the return
    value from the ISR, the interrupt object pointer and a pointer to
    the saved CPU state. This function is implemented by the Kernel
    and is responsible for finishing off the interrupt
    handling. Specifically, it may post a DSR depending on the ISR
    return value, and will decrement the scheduler lock. If the lock
    is zeroed by this operation then any posted DSRs may be called and
    may in turn result in a thread context switch.
    </PARA>
  </LISTITEM>
  
  <LISTITEM>
    <PARA>
    The return from <FUNCTION>interrupt_end()</FUNCTION> may occur
    some time after the call. Many other threads may have executed in
    the meantime. So here all we may do is restore the machine state
    and resume execution of the interrupted thread. Depending on the
    architecture, it may be necessary to disable interrupts again for
    part of this.
    </PARA>
  </LISTITEM>
  
</ORDEREDLIST>

<PARA>
The detailed order of these steps may vary slightly depending on the
architecture, in particular where interrupts are enabled and disabled.
</PARA>

</SECTION>

<!-- }}} -->

</CHAPTER>

<!-- }}} -->
<!-- {{{ Porting Guide -->

&hal-common-porting-sgml;

<!-- }}} -->
<!-- {{{ Future Developments -->

<CHAPTER id="hal-future-developments">
<TITLE><!-- <index></index> --><!-- <xref> -->Future developments</TITLE>

<PARA>
The HAL is not complete, and will evolve and increase over
time. Among the intended developments are:
</PARA>
<ITEMIZEDLIST>
<LISTITEM>
<PARA>Common macros for interpreting the contents of a saved
machine context. These would allow portable code, such as debug
stubs, to extract such values as the program counter and stack pointer
from a state without having to interpret a <STRUCTNAME>HAL_SavedRegisters</STRUCTNAME> structure
directly.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>Debugging support. Macros to set and clear hardware and
software breakpoints. Access to other areas of machine state may
also be supported.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>Static initialization support. The current HAL provides a
dynamic interface to things like thread context initialization and ISR
attachment. We also need to be able to define the system entirely
statically so that it is ready to go on restart, without needing to
run code. This will require extra macros to define these
initializations.  Such support may have a consequential effect on the
current HAL specification.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>CPU state control. Many CPUs have both kernel and user
states. Although it is not intended to run any code in user state
for the foreseeable future, it is possible that this may happen
eventually. If this is the case, then some minor changes may be needed
to the current HAL API to accommodate this. These should mostly
be extensions, but minor changes in semantics may also be required.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>Physical memory management. Many embedded systems have
multiple memory areas with varying properties such as base address,
size, speed, bus width, cacheability and persistence. An API is
needed to support the discovery of this information about the machine's
physical memory map.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>Memory management control. Some embedded processors have
a memory management unit. In some cases this must be enabled to
allow the cache to be controlled, particularly if different regions
of memory must have different caching properties. For some purposes,
in some systems, it will be useful to manipulate the MMU settings
dynamically.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>Power management. Macros to access and control any power
management mechanisms available on the CPU implementation. These
would provide a substrate for a more general power management system
that also involved device drivers and other hardware components.</PARA>
</LISTITEM>
<LISTITEM>
<PARA>Generic serial line macros. Most serial line devices operate
in the same way, the only real differences being exactly which bits
in which registers perform the standard functions. It should be
possible to develop a set of HAL macros that provide basic serial
line services such as baud rate setting, enabling interrupts, polling
for transmit or receive ready, transmitting and receiving data etc.
Given these it should be possible to create a generic serial line
device driver that will allow rapid bootstrapping on any new platform.
It may be possible to extend this mechanism to other device types.</PARA>
</LISTITEM>
</ITEMIZEDLIST>
</CHAPTER>

<!-- }}} -->

</part>