path: root/gas/doc/c-tic54x.texi

@c Copyright 2000, 2002, 2003 Free Software Foundation, Inc.
@c This is part of the GAS manual.
@c For copying conditions, see the file as.texinfo.
@c TI TMS320C54X description by Timothy Wall, twall@cygnus.com
@ifset GENERIC
@page
@node TIC54X-Dependent
@chapter TIC54X Dependent Features
@end ifset
@ifclear GENERIC
@node Machine Dependencies
@chapter TIC54X Dependent Features
@end ifclear

@cindex TIC54X support
@menu
* TIC54X-Opts::              Command-line Options
* TIC54X-Block::             Blocking
* TIC54X-Env::               Environment Settings
* TIC54X-Constants::         Constants Syntax
* TIC54X-Subsyms::           String Substitution
* TIC54X-Locals::            Local Label Syntax
* TIC54X-Builtins::          Builtin Assembler Math Functions
* TIC54X-Ext::               Extended Addressing Support
* TIC54X-Directives::        Directives
* TIC54X-Macros::            Macro Features
* TIC54X-MMRegs::            Memory-mapped Registers
@end menu

@node TIC54X-Opts
@section Options

@cindex options, TIC54X
@cindex TIC54X options
The TMS320C54X version of @code{@value{AS}} has a few machine-dependent options.

@cindex @samp{-mfar-mode} option, far-mode
@cindex @samp{-mf} option, far-mode
You can use the @samp{-mfar-mode} option to enable extended addressing mode.
All addresses will be assumed to be > 16 bits, and the appropriate
relocation types will be used.  This option is equivalent to using the
@samp{.far_mode} directive in the assembly code.  If you do not use the
@samp{-mfar-mode} option, all references will be assumed to be 16 bits.  
This option may be abbreviated to @samp{-mf}.

@cindex @samp{-mcpu} option, cpu
You can use the @samp{-mcpu} option to specify a particular CPU.
This option is equivalent to using the @samp{.version} directive in the
assembly code.  For recognized CPU codes, see
@xref{TIC54X-Directives,,@code{.version}}.  The default CPU version is
@samp{542}.   

@cindex @samp{-merrors-to-file} option, stderr redirect
@cindex @samp{-me} option, stderr redirect
You can use the @samp{-merrors-to-file} option to redirect error output
to a file (this provided for those deficient environments which don't
provide adequate output redirection).  This option may be abbreviated to
@samp{-me}. 

@node TIC54X-Block
@section Blocking
A blocked section or memory block is guaranteed not to cross the blocking
boundary (usually a page, or 128 words) if it is smaller than the
blocking size, or to start on a page boundary if it is larger than the
blocking size.

@node TIC54X-Env
@section Environment Settings

@cindex environment settings, TIC54X
@cindex @samp{A_DIR} environment variable, TIC54X
@cindex @samp{C54XDSP_DIR} environment variable, TIC54X
@samp{C54XDSP_DIR} and @samp{A_DIR} are semicolon-separated
paths which are added to the list of directories normally searched for
source and include files.  @samp{C54XDSP_DIR} will override @samp{A_DIR}.

@node TIC54X-Constants
@section Constants Syntax

@cindex constants, TIC54X
The TIC54X version of @code{@value{AS}} allows the following additional
constant formats, using a suffix to indicate the radix:
@smallexample
@cindex binary constants, TIC54X

Binary                  @code{000000B, 011000b}
Octal                   @code{10Q, 224q}
Hexadecimal             @code{45h, 0FH}

@end smallexample

@node TIC54X-Subsyms
@section String Substitution
A subset of allowable symbols (which we'll call subsyms) may be assigned
arbitrary string values.  This is roughly equivalent to C preprocessor
#define macros.  When @code{@value{AS}} encounters one of these
symbols, the symbol is replaced in the input stream by its string value.
Subsym names @strong{must} begin with a letter. 

Subsyms may be defined using the @code{.asg} and @code{.eval} directives
(@xref{TIC54X-Directives,,@code{.asg}},
@xref{TIC54X-Directives,,@code{.eval}}. 

Expansion is recursive until a previously encountered symbol is seen, at
which point substitution stops.  

In this example, x is replaced with SYM2; SYM2 is replaced with SYM1, and SYM1
is replaced with x.  At this point, x has already been encountered
and the substitution stops.

@smallexample @code
 .asg   "x",SYM1 
 .asg   "SYM1",SYM2
 .asg   "SYM2",x
 add    x,a             ; final code assembled is "add  x, a"
@end smallexample

Macro parameters are converted to subsyms; a side effect of this is the normal
@code{@value{AS}} '\ARG' dereferencing syntax is unnecessary.  Subsyms
defined within a macro will have global scope, unless the @code{.var}
directive is used to identify the subsym as a local macro variable
@pxref{TIC54X-Directives,,@code{.var}}.

Substitution may be forced in situations where replacement might be
ambiguous by placing colons on either side of the subsym.  The following
code: 

@smallexample @code
 .eval  "10",x
LAB:X:  add     #x, a
@end smallexample

When assembled becomes:

@smallexample @code
LAB10  add     #10, a
@end smallexample

Smaller parts of the string assigned to a subsym may be accessed with
the following syntax:

@table @code
@item @code{:@var{symbol}(@var{char_index}):}
Evaluates to a single-character string, the character at @var{char_index}.
@item @code{:@var{symbol}(@var{start},@var{length}):}
Evaluates to a substring of @var{symbol} beginning at @var{start} with
length @var{length}.
@end table

@node TIC54X-Locals
@section Local Labels
Local labels may be defined in two ways:

@itemize @bullet
@item
$N, where N is a decimal number between 0 and 9
@item
LABEL?, where LABEL is any legal symbol name.  
@end itemize

Local labels thus defined may be redefined or automatically generated.
The scope of a local label is based on when it may be undefined or reset.
This happens when one of the following situations is encountered:

@itemize @bullet
@item
.newblock directive @pxref{TIC54X-Directives,,@code{.newblock}}
@item
The current section is changed (.sect, .text, or .data)
@item
Entering or leaving an included file
@item
The macro scope where the label was defined is exited
@end itemize

@node TIC54X-Builtins
@section Math Builtins

@cindex math builtins, TIC54X
@cindex TIC54X builtin math functions
@cindex builtin math functions, TIC54X

The following built-in functions may be used to generate a
floating-point value.  All return a floating-point value except
@samp{$cvi}, @samp{$int}, and @samp{$sgn}, which return an integer
value. 

@table @code
@cindex @code{$acos} math builtin, TIC54X
@item @code{$acos(@var{expr})}
Returns the floating point arccosine of @var{expr}.

@cindex @code{$asin} math builtin, TIC54X
@item @code{$asin(@var{expr})}
Returns the floating point arcsine of @var{expr}.

@cindex @code{$atan} math builtin, TIC54X
@item @code{$atan(@var{expr})}
Returns the floating point arctangent of @var{expr}.

@cindex @code{$atan2} math builtin, TIC54X
@item @code{$atan2(@var{expr1},@var{expr2})}
Returns the floating point arctangent of @var{expr1} / @var{expr2}.

@cindex @code{$ceil} math builtin, TIC54X
@item @code{$ceil(@var{expr})}
Returns the smallest integer not less than @var{expr} as floating point.

@cindex @code{$cosh} math builtin, TIC54X
@item @code{$cosh(@var{expr})}
Returns the floating point hyperbolic cosine of @var{expr}.

@cindex @code{$cos} math builtin, TIC54X
@item @code{$cos(@var{expr})}
Returns the floating point cosine of @var{expr}.

@cindex @code{$cvf} math builtin, TIC54X
@item @code{$cvf(@var{expr})}
Returns the integer value @var{expr} converted to floating-point.

@cindex @code{$cvi} math builtin, TIC54X
@item @code{$cvi(@var{expr})}
Returns the floating point value @var{expr} converted to integer.

@cindex @code{$exp} math builtin, TIC54X
@item @code{$exp(@var{expr})}
Returns the floating point value e ^ @var{expr}.

@cindex @code{$fabs} math builtin, TIC54X
@item @code{$fabs(@var{expr})}
Returns the floating point absolute value of @var{expr}.

@cindex @code{$floor} math builtin, TIC54X
@item @code{$floor(@var{expr})}
Returns the largest integer that is not greater than @var{expr} as
floating point.

@cindex @code{$fmod} math builtin, TIC54X
@item @code{$fmod(@var{expr1},@var{expr2})}
Returns the floating point remainder of @var{expr1} / @var{expr2}.

@cindex @code{$int} math builtin, TIC54X
@item @code{$int(@var{expr})}
Returns 1 if @var{expr} evaluates to an integer, zero otherwise.

@cindex @code{$ldexp} math builtin, TIC54X
@item @code{$ldexp(@var{expr1},@var{expr2})}
Returns the floating point value @var{expr1} * 2 ^ @var{expr2}.

@cindex @code{$log10} math builtin, TIC54X
@item @code{$log10(@var{expr})}
Returns the base 10 logarithm of @var{expr}.

@cindex @code{$log} math builtin, TIC54X
@item @code{$log(@var{expr})}
Returns the natural logarithm of @var{expr}.

@cindex @code{$max} math builtin, TIC54X
@item @code{$max(@var{expr1},@var{expr2})}
Returns the floating point maximum of @var{expr1} and @var{expr2}.

@cindex @code{$min} math builtin, TIC54X
@item @code{$min(@var{expr1},@var{expr2})}
Returns the floating point minimum of @var{expr1} and @var{expr2}.

@cindex @code{$pow} math builtin, TIC54X
@item @code{$pow(@var{expr1},@var{expr2})}
Returns the floating point value @var{expr1} ^ @var{expr2}.

@cindex @code{$round} math builtin, TIC54X
@item @code{$round(@var{expr})}
Returns the nearest integer to @var{expr} as a floating point number.

@cindex @code{$sgn} math builtin, TIC54X
@item @code{$sgn(@var{expr})}
Returns -1, 0, or 1 based on the sign of @var{expr}.

@cindex @code{$sin} math builtin, TIC54X
@item @code{$sin(@var{expr})}
Returns the floating point sine of @var{expr}.

@cindex @code{$sinh} math builtin, TIC54X
@item @code{$sinh(@var{expr})}
Returns the floating point hyperbolic sine of @var{expr}.

@cindex @code{$sqrt} math builtin, TIC54X
@item @code{$sqrt(@var{expr})}
Returns the floating point square root of @var{expr}.

@cindex @code{$tan} math builtin, TIC54X
@item @code{$tan(@var{expr})}
Returns the floating point tangent of @var{expr}.

@cindex @code{$tanh} math builtin, TIC54X
@item @code{$tanh(@var{expr})}
Returns the floating point hyperbolic tangent of @var{expr}.

@cindex @code{$trunc} math builtin, TIC54X
@item @code{$trunc(@var{expr})}
Returns the integer value of @var{expr} truncated towards zero as
floating point.

@end table

@node TIC54X-Ext
@section Extended Addressing
The @code{LDX} pseudo-op is provided for loading the extended addressing bits
of a label or address.  For example, if an address @code{_label} resides
in extended program memory, the value of @code{_label} may be loaded as
follows:
@smallexample @code
 ldx     #_label,16,a    ; loads extended bits of _label
 or      #_label,a       ; loads lower 16 bits of _label
 bacc    a               ; full address is in accumulator A
@end smallexample

@node TIC54X-Directives
@section Directives

@cindex machine directives, TIC54X
@cindex TIC54X machine directives

@table @code

@cindex @code{align} directive, TIC54X
@cindex @code{even} directive, TIC54X
@item .align [@var{size}]
@itemx .even
Align the section program counter on the next boundary, based on
@var{size}.  @var{size} may be any power of 2.  @code{.even} is
equivalent to @code{.align} with a @var{size} of 2.
@table @code
@item 1
Align SPC to word boundary
@item 2
Align SPC to longword boundary (same as .even)
@item 128       
Align SPC to page boundary
@end table

@cindex @code{asg} directive, TIC54X
@item .asg @var{string}, @var{name}
Assign @var{name} the string @var{string}.  String replacement is
performed on @var{string} before assignment.

@cindex @code{eval} directive, TIC54X
@itemx .eval @var{string}, @var{name}
Evaluate the contents of string @var{string} and assign the result as a
string to the subsym @var{name}.  String replacement is performed on
@var{string} before assignment. 

@cindex @code{bss} directive, TIC54X
@item .bss @var{symbol}, @var{size} [, [@var{blocking_flag}] [,@var{alignment_flag}]]
Reserve space for @var{symbol} in the .bss section.  @var{size} is in
words.  If present, @var{blocking_flag} indicates the allocated space
should be aligned on a page boundary if it would otherwise cross a page
boundary.  If present, @var{alignment_flag} causes the assembler to
allocate @var{size} on a long word boundary.

@cindex @code{byte} directive, TIC54X
@cindex @code{ubyte} directive, TIC54X
@cindex @code{char} directive, TIC54X
@cindex @code{uchar} directive, TIC54X
@item .byte @var{value} [,...,@var{value_n}]
@itemx .ubyte @var{value} [,...,@var{value_n}]
@itemx .char @var{value} [,...,@var{value_n}]
@itemx .uchar @var{value} [,...,@var{value_n}]
Place one or more bytes into consecutive words of the current section.
The upper 8 bits of each word is zero-filled.  If a label is used, it
points to the word allocated for the first byte encountered.

@cindex @code{clink} directive, TIC54X
@item .clink ["@var{section_name}"]
Set STYP_CLINK flag for this section, which indicates to the linker that
if no symbols from this section are referenced, the section should not
be included in the link.  If @var{section_name} is omitted, the current
section is used.

@cindex @code{c_mode} directive, TIC54X
@item .c_mode
TBD.

@cindex @code{copy} directive, TIC54X
@item .copy "@var{filename}" | @var{filename}
@itemx .include "@var{filename}" | @var{filename}
Read source statements from @var{filename}.  The normal include search
path is used.  Normally .copy will cause statements from the included
file to be printed in the assembly listing and .include will not, but
this distinction is not currently implemented.

@cindex @code{data} directive, TIC54X
@item .data
Begin assembling code into the .data section.

@cindex @code{double} directive, TIC54X
@cindex @code{ldouble} directive, TIC54X
@cindex @code{float} directive, TIC54X
@cindex @code{xfloat} directive, TIC54X
@item .double @var{value} [,...,@var{value_n}]
@itemx .ldouble @var{value} [,...,@var{value_n}]
@itemx .float @var{value} [,...,@var{value_n}]
@itemx .xfloat @var{value} [,...,@var{value_n}]
Place an IEEE single-precision floating-point representation of one or
more floating-point values into the current section.  All but
@code{.xfloat} align the result on a longword boundary.  Values are
stored most-significant word first. 

@cindex @code{drlist} directive, TIC54X
@cindex @code{drnolist} directive, TIC54X
@item .drlist
@itemx .drnolist
Control printing of directives to the listing file.  Ignored.

@cindex @code{emsg} directive, TIC54X
@cindex @code{mmsg} directive, TIC54X
@cindex @code{wmsg} directive, TIC54X
@item .emsg @var{string}
@itemx .mmsg @var{string}
@itemx .wmsg @var{string}
Emit a user-defined error, message, or warning, respectively.

@cindex @code{far_mode} directive, TIC54X
@item .far_mode
Use extended addressing when assembling statements.  This should appear
only once per file, and is equivalent to the -mfar-mode option @pxref{TIC54X-Opts,,@code{-mfar-mode}}.

@cindex @code{fclist} directive, TIC54X
@cindex @code{fcnolist} directive, TIC54X
@item .fclist
@itemx .fcnolist
Control printing of false conditional blocks to the listing file.

@cindex @code{field} directive, TIC54X
@item .field @var{value} [,@var{size}]
Initialize a bitfield of @var{size} bits in the current section.  If
@var{value} is relocatable, then @var{size} must be 16.  @var{size}
defaults to 16 bits.  If @var{value} does not fit into @var{size} bits,
the value will be truncated.  Successive @code{.field} directives will
pack starting at the current word, filling the most significant bits
first, and aligning to the start of the next word if the field size does
not fit into the space remaining in the current word.  A @code{.align}
directive with an operand of 1 will force the next @code{.field}
directive to begin packing into a new word.  If a label is used, it
points to the word that contains the specified field.

@cindex @code{global} directive, TIC54X
@cindex @code{def} directive, TIC54X
@cindex @code{ref} directive, TIC54X
@item .global @var{symbol} [,...,@var{symbol_n}]
@itemx .def @var{symbol} [,...,@var{symbol_n}]
@itemx .ref @var{symbol} [,...,@var{symbol_n}]
@code{.def} nominally identifies a symbol defined in the current file
and available to other files.  @code{.ref} identifies a symbol used in
the current file but defined elsewhere.  Both map to the standard
@code{.global} directive.

@cindex @code{half} directive, TIC54X
@cindex @code{uhalf} directive, TIC54X
@cindex @code{short} directive, TIC54X
@cindex @code{ushort} directive, TIC54X
@cindex @code{int} directive, TIC54X
@cindex @code{uint} directive, TIC54X
@cindex @code{word} directive, TIC54X
@cindex @code{uword} directive, TIC54X
@item .half @var{value} [,...,@var{value_n}]
@itemx .uhalf @var{value} [,...,@var{value_n}]
@itemx .short @var{value} [,...,@var{value_n}]
@itemx .ushort @var{value} [,...,@var{value_n}]
@itemx .int @var{value} [,...,@var{value_n}]
@itemx .uint @var{value} [,...,@var{value_n}]
@itemx .word @var{value} [,...,@var{value_n}]
@itemx .uword @var{value} [,...,@var{value_n}]
Place one or more values into consecutive words of the current section.
If a label is used, it points to the word allocated for the first value
encountered. 

@cindex @code{label} directive, TIC54X
@item .label @var{symbol}
Define a special @var{symbol} to refer to the load time address of the
current section program counter.

@cindex @code{length} directive, TIC54X
@cindex @code{width} directive, TIC54X
@item .length
@itemx .width
Set the page length and width of the output listing file.  Ignored.

@cindex @code{list} directive, TIC54X
@cindex @code{nolist} directive, TIC54X
@item .list
@itemx .nolist
Control whether the source listing is printed.  Ignored.

@cindex @code{long} directive, TIC54X
@cindex @code{ulong} directive, TIC54X
@cindex @code{xlong} directive, TIC54X
@item .long @var{value} [,...,@var{value_n}]
@itemx .ulong @var{value} [,...,@var{value_n}]
@itemx .xlong @var{value} [,...,@var{value_n}]
Place one or more 32-bit values into consecutive words in the current
section.  The most significant word is stored first.  @code{.long} and
@code{.ulong} align the result on a longword boundary; @code{xlong} does
not. 

@cindex @code{loop} directive, TIC54X
@cindex @code{break} directive, TIC54X
@cindex @code{endloop} directive, TIC54X
@item .loop [@var{count}]
@itemx .break [@var{condition}]
@itemx .endloop
Repeatedly assemble a block of code.  @code{.loop} begins the block, and
@code{.endloop} marks its termination.  @var{count} defaults to 1024,
and indicates the number of times the block should be repeated.
@code{.break} terminates the loop so that assembly begins after the
@code{.endloop} directive.  The optional @var{condition} will cause the
loop to terminate only if it evaluates to zero.

@cindex @code{macro} directive, TIC54X
@cindex @code{endm} directive, TIC54X
@item @var{macro_name} .macro [@var{param1}][,...@var{param_n}]
@itemx [.mexit]
@itemx .endm
See the section on macros for more explanation (@xref{TIC54X-Macros}.

@cindex @code{mlib} directive, TIC54X
@item .mlib "@var{filename}" | @var{filename}
Load the macro library @var{filename}.  @var{filename} must be an
archived library (BFD ar-compatible) of text files, expected to contain
only macro definitions.   The standard include search path is used.

@cindex @code{mlist} directive, TIC54X
@cindex @code{mnolist} directive, TIC54X
@item .mlist
@item .mnolist
Control whether to include macro and loop block expansions in the
listing output.  Ignored.

@cindex @code{mmregs} directive, TIC54X
@item .mmregs
Define global symbolic names for the 'c54x registers.  Supposedly
equivalent to executing @code{.set} directives for each register with
its memory-mapped value, but in reality is provided only for
compatibility and does nothing.

@cindex @code{newblock} directive, TIC54X
@item .newblock
This directive resets any TIC54X local labels currently defined.  Normal
@code{@value{AS}} local labels are unaffected.

@cindex @code{option} directive, TIC54X
@item .option @var{option_list}
Set listing options.  Ignored.

@cindex @code{sblock} directive, TIC54X
@item .sblock "@var{section_name}" | @var{section_name} [,"@var{name_n}" | @var{name_n}]
Designate @var{section_name} for blocking.  Blocking guarantees that a
section will start on a page boundary (128 words) if it would otherwise
cross a page boundary.  Only initialized sections may be designated with
this directive.  See also @xref{TIC54X-Block}.

@cindex @code{sect} directive, TIC54X
@item .sect "@var{section_name}"
Define a named initialized section and make it the current section.

@cindex @code{set} directive, TIC54X
@cindex @code{equ} directive, TIC54X
@item @var{symbol} .set "@var{value}"
@itemx @var{symbol} .equ "@var{value}"
Equate a constant @var{value} to a @var{symbol}, which is placed in the
symbol table.  @var{symbol} may not be previously defined.  

@cindex @code{space} directive, TIC54X
@cindex @code{bes} directive, TIC54X
@item .space @var{size_in_bits}
@itemx .bes @var{size_in_bits}
Reserve the given number of bits in the current section and zero-fill
them.  If a label is used with @code{.space}, it points to the
@strong{first} word reserved.  With @code{.bes}, the label points to the
@strong{last} word reserved.

@cindex @code{sslist} directive, TIC54X
@cindex @code{ssnolist} directive, TIC54X
@item .sslist
@itemx .ssnolist
Controls the inclusion of subsym replacement in the listing output.  Ignored.

@cindex @code{string} directive, TIC54X
@cindex @code{pstring} directive, TIC54X
@item .string "@var{string}" [,...,"@var{string_n}"]
@itemx .pstring "@var{string}" [,...,"@var{string_n}"]
Place 8-bit characters from @var{string} into the current section.
@code{.string} zero-fills the upper 8 bits of each word, while
@code{.pstring} puts two characters into each word, filling the
most-significant bits first.  Unused space is zero-filled.  If a label
is used, it points to the first word initialized.

@cindex @code{struct} directive, TIC54X
@cindex @code{tag} directive, TIC54X
@cindex @code{endstruct} directive, TIC54X
@item [@var{stag}] .struct [@var{offset}]
@itemx [@var{name_1}] element [@var{count_1}]
@itemx [@var{name_2}] element [@var{count_2}]
@itemx [@var{tname}] .tag @var{stagx} [@var{tcount}]
@itemx ...
@itemx [@var{name_n}] element [@var{count_n}]
@itemx [@var{ssize}] .endstruct
@itemx @var{label} .tag [@var{stag}]
Assign symbolic offsets to the elements of a structure.  @var{stag}
defines a symbol to use to reference the structure.  @var{offset}
indicates a starting value to use for the first element encountered;
otherwise it defaults to zero.  Each element can have a named offset,
@var{name}, which is a symbol assigned the value of the element's offset
into the structure.  If @var{stag} is missing, these become global
symbols.  @var{count} adjusts the offset that many times, as if
@code{element} were an array.  @code{element} may be one of
@code{.byte}, @code{.word}, @code{.long}, @code{.float}, or any
equivalent of those, and the structure offset is adjusted accordingly.
@code{.field} and @code{.string} are also allowed; the size of
@code{.field} is one bit, and @code{.string} is considered to be one
word in size.  Only element descriptors, structure/union tags,
@code{.align} and conditional assembly directives are allowed within
@code{.struct}/@code{.endstruct}.  @code{.align} aligns member offsets
to word boundaries only.  @var{ssize}, if provided, will always be
assigned the size of the structure.

The @code{.tag} directive, in addition to being used to define a
structure/union element within a structure, may be used to apply a
structure to a symbol.  Once applied to @var{label}, the individual
structure elements may be applied to @var{label} to produce the desired
offsets using @var{label} as the structure base. 

@cindex @code{tab} directive, TIC54X
@item .tab
Set the tab size in the output listing.  Ignored.

@cindex @code{union} directive, TIC54X
@cindex @code{tag} directive, TIC54X
@cindex @code{endunion} directive, TIC54X
@item [@var{utag}] .union 
@itemx [@var{name_1}] element [@var{count_1}]
@itemx [@var{name_2}] element [@var{count_2}]
@itemx [@var{tname}] .tag @var{utagx}[,@var{tcount}]
@itemx ...
@itemx [@var{name_n}] element [@var{count_n}]
@itemx [@var{usize}] .endstruct
@itemx @var{label} .tag [@var{utag}]
Similar to @code{.struct}, but the offset after each element is reset to
zero, and the @var{usize} is set to the maximum of all defined elements.
Starting offset for the union is always zero.

@cindex @code{usect} directive, TIC54X
@item [@var{symbol}] .usect "@var{section_name}", @var{size}, [,[@var{blocking_flag}] [,@var{alignment_flag}]]
Reserve space for variables in a named, uninitialized section (similar to
.bss).  @code{.usect} allows definitions sections independent of .bss.
@var{symbol} points to the first location reserved by this allocation.
The symbol may be used as a variable name.  @var{size} is the allocated
size in words.  @var{blocking_flag} indicates whether to block this
section on a page boundary (128 words) (@pxref{TIC54X-Block}).
@var{alignment flag} indicates whether the section should be
longword-aligned.  

@cindex @code{var} directive, TIC54X
@item .var @var{sym}[,..., @var{sym_n}]
Define a subsym to be a local variable within a macro.  See
@xref{TIC54X-Macros}. 

@cindex @code{version} directive, TIC54X
@item .version @var{version}
Set which processor to build instructions for.  Though the following
values are accepted, the op is ignored.
@table @code
@item 541
@itemx 542
@itemx 543
@itemx 545
@itemx 545LP
@itemx 546LP
@itemx 548
@itemx 549
@end table
@end table

@node TIC54X-Macros
@section Macros

@cindex TIC54X-specific macros
@cindex macros, TIC54X
Macros do not require explicit dereferencing of arguments (i.e., \ARG).  

During macro expansion, the macro parameters are converted to subsyms.
If the number of arguments passed the macro invocation exceeds the
number of parameters defined, the last parameter is assigned the string
equivalent of all remaining arguments.  If fewer arguments are given
than parameters, the missing parameters are assigned empty strings.  To
include a comma in an argument, you must enclose the argument in quotes.

@cindex subsym builtins, TIC54X
@cindex TIC54X subsym builtins
@cindex builtin subsym functions, TIC54X
The following built-in subsym functions allow examination of the string
value of subsyms (or ordinary strings).  The arguments are strings
unless otherwise indicated (subsyms passed as args will be replaced by
the strings they represent).
@table @code
@cindex @code{$symlen} subsym builtin, TIC54X
@item @code{$symlen(@var{str})}
Returns the length of @var{str}.

@cindex @code{$symcmp} subsym builtin, TIC54X
@item @code{$symcmp(@var{str1},@var{str2})}
Returns 0 if @var{str1} == @var{str2}, non-zero otherwise.  

@cindex @code{$firstch} subsym builtin, TIC54X
@item @code{$firstch(@var{str},@var{ch})}
Returns index of the first occurrence of character constant @var{ch} in
@var{str}. 

@cindex @code{$lastch} subsym builtin, TIC54X
@item @code{$lastch(@var{str},@var{ch})}
Returns index of the last occurrence of character constant @var{ch} in
@var{str}. 

@cindex @code{$isdefed} subsym builtin, TIC54X
@item @code{$isdefed(@var{symbol})}
Returns zero if the symbol @var{symbol} is not in the symbol table,
non-zero otherwise.

@cindex @code{$ismember} subsym builtin, TIC54X
@item @code{$ismember(@var{symbol},@var{list})}
Assign the first member of comma-separated string @var{list} to
@var{symbol}; @var{list} is reassigned the remainder of the list.  Returns
zero if @var{list} is a null string.  Both arguments must be subsyms.

@cindex @code{$iscons} subsym builtin, TIC54X
@item @code{$iscons(@var{expr})}
Returns 1 if string @var{expr} is binary, 2 if octal, 3 if hexadecimal,
4 if a character, 5 if decimal, and zero if not an integer.

@cindex @code{$isname} subsym builtin, TIC54X
@item @code{$isname(@var{name})}
Returns 1 if @var{name} is a valid symbol name, zero otherwise.

@cindex @code{$isreg} subsym builtin, TIC54X
@item @code{$isreg(@var{reg})}
Returns 1 if @var{reg} is a valid predefined register name (AR0-AR7 only).

@cindex @code{$structsz} subsym builtin, TIC54X
@item @code{$structsz(@var{stag})}
Returns the size of the structure or union represented by @var{stag}.

@cindex @code{$structacc} subsym builtin, TIC54X
@item @code{$structacc(@var{stag})}
Returns the reference point of the structure or union represented by
@var{stag}.   Always returns zero.

@end table

@node TIC54X-MMRegs
@section Memory-mapped Registers

@cindex TIC54X memory-mapped registers
@cindex registers, TIC54X memory-mapped
@cindex memory-mapped registers, TIC54X
The following symbols are recognized as memory-mapped registers:

@table @code
@end table