Subversion Repositories openrisc

@c Copyright 2001, 2002, 2003 Free Software Foundation, Inc.

@c This is part of the GAS manual.

@c For copying conditions, see the file as.texinfo.

@c MMIX description by Hans-Peter Nilsson, hp@bitrange.com

@ifset GENERIC

@page

@node MMIX-Dependent

@chapter MMIX Dependent Features

@end ifset

@ifclear GENERIC

@node Machine Dependencies

@chapter MMIX Dependent Features

@end ifclear

@cindex MMIX support

@menu

* MMIX-Opts::              Command-line Options

* MMIX-Expand::            Instruction expansion

* MMIX-Syntax::            Syntax

* MMIX-mmixal::            Differences to @code{mmixal} syntax and semantics

@end menu

@node MMIX-Opts

@section Command-line Options

@cindex options, MMIX

@cindex MMIX options

The MMIX version of @code{@value{AS}} has some machine-dependent options.

@cindex @samp{--fixed-special-register-names} command line option, MMIX

When @samp{--fixed-special-register-names} is specified, only the register

names specified in @ref{MMIX-Regs} are recognized in the instructions

@code{PUT} and @code{GET}.

@cindex @samp{--globalize-symbols} command line option, MMIX

You can use the @samp{--globalize-symbols} to make all symbols global.

This option is useful when splitting up a @code{mmixal} program into

several files.

@cindex @samp{--gnu-syntax} command line option, MMIX

The @samp{--gnu-syntax} turns off most syntax compatibility with

@code{mmixal}.  Its usability is currently doubtful.

@cindex @samp{--relax} command line option, MMIX

The @samp{--relax} option is not fully supported, but will eventually make

the object file prepared for linker relaxation.

@cindex @samp{--no-predefined-syms} command line option, MMIX

If you want to avoid inadvertently calling a predefined symbol and would

rather get an error, for example when using @code{@value{AS}} with a

compiler or other machine-generated code, specify

@samp{--no-predefined-syms}.  This turns off built-in predefined

definitions of all such symbols, including rounding-mode symbols, segment

symbols, @samp{BIT} symbols, and @code{TRAP} symbols used in @code{mmix}

``system calls''.  It also turns off predefined special-register names,

except when used in @code{PUT} and @code{GET} instructions.

@cindex @samp{--no-expand} command line option, MMIX

By default, some instructions are expanded to fit the size of the operand

or an external symbol (@pxref{MMIX-Expand}).  By passing

@samp{--no-expand}, no such expansion will be done, instead causing errors

at link time if the operand does not fit.

@cindex @samp{--no-merge-gregs} command line option, MMIX

The @code{mmixal} documentation (@pxref{mmixsite}) specifies that global

registers allocated with the @samp{GREG} directive (@pxref{MMIX-greg}) and

initialized to the same non-zero value, will refer to the same global

register.  This isn't strictly enforceable in @code{@value{AS}} since the

final addresses aren't known until link-time, but it will do an effort

unless the @samp{--no-merge-gregs} option is specified.  (Register merging

isn't yet implemented in @code{@value{LD}}.)

@cindex @samp{-x} command line option, MMIX

@code{@value{AS}} will warn every time it expands an instruction to fit an

operand unless the option @samp{-x} is specified.  It is believed that

this behaviour is more useful than just mimicking @code{mmixal}'s

behaviour, in which instructions are only expanded if the @samp{-x} option

is specified, and assembly fails otherwise, when an instruction needs to

be expanded.  It needs to be kept in mind that @code{mmixal} is both an

assembler and linker, while @code{@value{AS}} will expand instructions

that at link stage can be contracted.  (Though linker relaxation isn't yet

implemented in @code{@value{LD}}.)  The option @samp{-x} also imples

@samp{--linker-allocated-gregs}.

@cindex @samp{--no-pushj-stubs} command line option, MMIX

@cindex @samp{--no-stubs} command line option, MMIX

If instruction expansion is enabled, @code{@value{AS}} can expand a

@samp{PUSHJ} instruction into a series of instructions.  The shortest

expansion is to not expand it, but just mark the call as redirectable to a

stub, which @code{@value{LD}} creates at link-time, but only if the

original @samp{PUSHJ} instruction is found not to reach the target.  The

stub consists of the necessary instructions to form a jump to the target.

This happens if @code{@value{AS}} can assert that the @samp{PUSHJ}

instruction can reach such a stub.  The option @samp{--no-pushj-stubs}

disables this shorter expansion, and the longer series of instructions is

then created at assembly-time.  The option @samp{--no-stubs} is a synonym,

intended for compatibility with future releases, where generation of stubs

for other instructions may be implemented.

@cindex @samp{--linker-allocated-gregs} command line option, MMIX

Usually a two-operand-expression (@pxref{GREG-base}) without a matching

@samp{GREG} directive is treated as an error by @code{@value{AS}}.  When

the option @samp{--linker-allocated-gregs} is in effect, they are instead

passed through to the linker, which will allocate as many global registers

as is needed.

@node MMIX-Expand

@section Instruction expansion

@cindex instruction expansion, MMIX

When @code{@value{AS}} encounters an instruction with an operand that is

either not known or does not fit the operand size of the instruction,

@code{@value{AS}} (and @code{@value{LD}}) will expand the instruction into

a sequence of instructions semantically equivalent to the operand fitting

the instruction.  Expansion will take place for the following

instructions:

@table @asis

@item @samp{GETA}

Expands to a sequence of four instructions: @code{SETL}, @code{INCML},

@code{INCMH} and @code{INCH}.  The operand must be a multiple of four.

@item Conditional branches

A branch instruction is turned into a branch with the complemented

condition and prediction bit over five instructions; four instructions

setting @code{$255} to the operand value, which like with @code{GETA} must

be a multiple of four, and a final @code{GO $255,$255,0}.

@item @samp{PUSHJ}

Similar to expansion for conditional branches; four instructions set

@code{$255} to the operand value, followed by a @code{PUSHGO $255,$255,0}.

@item @samp{JMP}

Similar to conditional branches and @code{PUSHJ}.  The final instruction

is @code{GO $255,$255,0}.

@end table

The linker @code{@value{LD}} is expected to shrink these expansions for

code assembled with @samp{--relax} (though not currently implemented).

@node MMIX-Syntax

@section Syntax

The assembly syntax is supposed to be upward compatible with that

described in Sections 1.3 and 1.4 of @samp{The Art of Computer

Programming, Volume 1}.  Draft versions of those chapters as well as other

MMIX information is located at

@anchor{mmixsite}@url{http://www-cs-faculty.stanford.edu/~knuth/mmix-news.html}.

Most code examples from the mmixal package located there should work

unmodified when assembled and linked as single files, with a few

noteworthy exceptions (@pxref{MMIX-mmixal}).

Before an instruction is emitted, the current location is aligned to the

next four-byte boundary.  If a label is defined at the beginning of the

line, its value will be the aligned value.

In addition to the traditional hex-prefix @samp{0x}, a hexadecimal number

can also be specified by the prefix character @samp{#}.

After all operands to an MMIX instruction or directive have been

specified, the rest of the line is ignored, treated as a comment.

@menu

* MMIX-Chars::                  Special Characters

* MMIX-Symbols::                Symbols

* MMIX-Regs::                   Register Names

* MMIX-Pseudos::                Assembler Directives

@end menu

@node MMIX-Chars

@subsection Special Characters

@cindex line comment characters, MMIX

@cindex MMIX line comment characters

The characters @samp{*} and @samp{#} are line comment characters; each

start a comment at the beginning of a line, but only at the beginning of a

line.  A @samp{#} prefixes a hexadecimal number if found elsewhere on a

line.

Two other characters, @samp{%} and @samp{!}, each start a comment anywhere

on the line.  Thus you can't use the @samp{modulus} and @samp{not}

operators in expressions normally associated with these two characters.

A @samp{;} is a line separator, treated as a new-line, so separate

instructions can be specified on a single line.

@node MMIX-Symbols

@subsection Symbols

The character @samp{:} is permitted in identifiers.  There are two

exceptions to it being treated as any other symbol character: if a symbol

begins with @samp{:}, it means that the symbol is in the global namespace

and that the current prefix should not be prepended to that symbol

(@pxref{MMIX-prefix}).  The @samp{:} is then not considered part of the

symbol.  For a symbol in the label position (first on a line), a @samp{:}

at the end of a symbol is silently stripped off.  A label is permitted,

but not required, to be followed by a @samp{:}, as with many other

assembly formats.

The character @samp{@@} in an expression, is a synonym for @samp{.}, the

current location.

In addition to the common forward and backward local symbol formats

(@pxref{Symbol Names}), they can be specified with upper-case @samp{B} and

@samp{F}, as in @samp{8B} and @samp{9F}.  A local label defined for the

current position is written with a @samp{H} appended to the number:

@smallexample

3H LDB $0,$1,2

@end smallexample

This and traditional local-label formats cannot be mixed: a label must be

defined and referred to using the same format.

There's a minor caveat: just as for the ordinary local symbols, the local

symbols are translated into ordinary symbols using control characters are

to hide the ordinal number of the symbol.  Unfortunately, these symbols

are not translated back in error messages.  Thus you may see confusing

error messages when local symbols are used.  Control characters

@samp{\003} (control-C) and @samp{\004} (control-D) are used for the

MMIX-specific local-symbol syntax.

The symbol @samp{Main} is handled specially; it is always global.

By defining the symbols @samp{__.MMIX.start..text} and

@samp{__.MMIX.start..data}, the address of respectively the @samp{.text}

and @samp{.data} segments of the final program can be defined, though when

linking more than one object file, the code or data in the object file

containing the symbol is not guaranteed to be start at that position; just

the final executable.  @xref{MMIX-loc}.

@node MMIX-Regs

@subsection Register names

@cindex register names, MMIX

@cindex MMIX register names

Local and global registers are specified as @samp{$0} to @samp{$255}.

The recognized special register names are @samp{rJ}, @samp{rA}, @samp{rB},

@samp{rC}, @samp{rD}, @samp{rE}, @samp{rF}, @samp{rG}, @samp{rH},

@samp{rI}, @samp{rK}, @samp{rL}, @samp{rM}, @samp{rN}, @samp{rO},

@samp{rP}, @samp{rQ}, @samp{rR}, @samp{rS}, @samp{rT}, @samp{rU},

@samp{rV}, @samp{rW}, @samp{rX}, @samp{rY}, @samp{rZ}, @samp{rBB},

@samp{rTT}, @samp{rWW}, @samp{rXX}, @samp{rYY} and @samp{rZZ}.  A leading

@samp{:} is optional for special register names.

Local and global symbols can be equated to register names and used in

place of ordinary registers.

Similarly for special registers, local and global symbols can be used.

Also, symbols equated from numbers and constant expressions are allowed in

place of a special register, except when either of the options

@code{--no-predefined-syms} and @code{--fixed-special-register-names} are

specified.  Then only the special register names above are allowed for the

instructions having a special register operand; @code{GET} and @code{PUT}.

@node MMIX-Pseudos

@subsection Assembler Directives

@cindex assembler directives, MMIX

@cindex pseudo-ops, MMIX

@cindex MMIX assembler directives

@cindex MMIX pseudo-ops

@table @code

@item LOC

@cindex assembler directive LOC, MMIX

@cindex pseudo-op LOC, MMIX

@cindex MMIX assembler directive LOC

@cindex MMIX pseudo-op LOC

@anchor{MMIX-loc}

The @code{LOC} directive sets the current location to the value of the

operand field, which may include changing sections.  If the operand is a

constant, the section is set to either @code{.data} if the value is

@code{0x2000000000000000} or larger, else it is set to @code{.text}.

Within a section, the current location may only be changed to

monotonically higher addresses.  A LOC expression must be a previously

defined symbol or a ``pure'' constant.

An example, which sets the label @var{prev} to the current location, and

updates the current location to eight bytes forward:

@smallexample

prev LOC @@+8

@end smallexample

When a LOC has a constant as its operand, a symbol

@code{__.MMIX.start..text} or @code{__.MMIX.start..data} is defined

depending on the address as mentioned above.  Each such symbol is

interpreted as special by the linker, locating the section at that

address.  Note that if multiple files are linked, the first object file

with that section will be mapped to that address (not necessarily the file

with the LOC definition).

@item LOCAL

@cindex assembler directive LOCAL, MMIX

@cindex pseudo-op LOCAL, MMIX

@cindex MMIX assembler directive LOCAL

@cindex MMIX pseudo-op LOCAL

@anchor{MMIX-local}

Example:

@smallexample

 LOCAL external_symbol

 LOCAL 42

 .local asymbol

@end smallexample

This directive-operation generates a link-time assertion that the operand

does not correspond to a global register.  The operand is an expression

that at link-time resolves to a register symbol or a number.  A number is

treated as the register having that number.  There is one restriction on

the use of this directive: the pseudo-directive must be placed in a

section with contents, code or data.

@item IS

@cindex assembler directive IS, MMIX

@cindex pseudo-op IS, MMIX

@cindex MMIX assembler directive IS

@cindex MMIX pseudo-op IS

@anchor{MMIX-is}

The @code{IS} directive:

@smallexample

asymbol IS an_expression

@end smallexample

sets the symbol @samp{asymbol} to @samp{an_expression}.  A symbol may not

be set more than once using this directive.  Local labels may be set using

this directive, for example:

@smallexample

5H IS @@+4

@end smallexample

@item GREG

@cindex assembler directive GREG, MMIX

@cindex pseudo-op GREG, MMIX

@cindex MMIX assembler directive GREG

@cindex MMIX pseudo-op GREG

@anchor{MMIX-greg}

This directive reserves a global register, gives it an initial value and

optionally gives it a symbolic name.  Some examples:

@smallexample

areg GREG

breg GREG data_value

     GREG data_buffer

     .greg creg, another_data_value

@end smallexample

The symbolic register name can be used in place of a (non-special)

register.  If a value isn't provided, it defaults to zero.  Unless the

option @samp{--no-merge-gregs} is specified, non-zero registers allocated

with this directive may be eliminated by @code{@value{AS}}; another

register with the same value used in its place.

Any of the instructions

@samp{CSWAP},

@samp{GO},

@samp{LDA},

@samp{LDBU},

@samp{LDB},

@samp{LDHT},

@samp{LDOU},

@samp{LDO},

@samp{LDSF},

@samp{LDTU},

@samp{LDT},

@samp{LDUNC},

@samp{LDVTS},

@samp{LDWU},

@samp{LDW},

@samp{PREGO},

@samp{PRELD},

@samp{PREST},

@samp{PUSHGO},

@samp{STBU},

@samp{STB},

@samp{STCO},

@samp{STHT},

@samp{STOU},

@samp{STSF},

@samp{STTU},

@samp{STT},

@samp{STUNC},

@samp{SYNCD},

@samp{SYNCID},

can have a value nearby @anchor{GREG-base}an initial value in place of its

second and third operands.  Here, ``nearby'' is defined as within the

range 0@dots{}255 from the initial value of such an allocated register.

@smallexample

buffer1 BYTE 0,0,0,0,0

buffer2 BYTE 0,0,0,0,0

 @dots{}

 GREG buffer1

 LDOU $42,buffer2

@end smallexample

In the example above, the @samp{Y} field of the @code{LDOUI} instruction

(LDOU with a constant Z) will be replaced with the global register

allocated for @samp{buffer1}, and the @samp{Z} field will have the value

5, the offset from @samp{buffer1} to @samp{buffer2}.  The result is

equivalent to this code:

@smallexample

buffer1 BYTE 0,0,0,0,0

buffer2 BYTE 0,0,0,0,0

 @dots{}

tmpreg GREG buffer1

 LDOU $42,tmpreg,(buffer2-buffer1)

@end smallexample

Global registers allocated with this directive are allocated in order

higher-to-lower within a file.  Other than that, the exact order of

register allocation and elimination is undefined.  For example, the order

is undefined when more than one file with such directives are linked

together.  With the options @samp{-x} and @samp{--linker-allocated-gregs},

@samp{GREG} directives for two-operand cases like the one mentioned above

can be omitted.  Sufficient global registers will then be allocated by the

linker.

@item BYTE

@cindex assembler directive BYTE, MMIX

@cindex pseudo-op BYTE, MMIX

@cindex MMIX assembler directive BYTE

@cindex MMIX pseudo-op BYTE

@anchor{MMIX-byte}

The @samp{BYTE} directive takes a series of operands separated by a comma.

If an operand is a string (@pxref{Strings}), each character of that string

is emitted as a byte.  Other operands must be constant expressions without

forward references, in the range 0@dots{}255.  If you need operands having

expressions with forward references, use @samp{.byte} (@pxref{Byte}).  An

operand can be omitted, defaulting to a zero value.

@item WYDE

@itemx TETRA

@itemx OCTA

@cindex assembler directive WYDE, MMIX

@cindex pseudo-op WYDE, MMIX

@cindex MMIX assembler directive WYDE

@cindex MMIX pseudo-op WYDE

@cindex assembler directive TETRA, MMIX

@cindex pseudo-op TETRA, MMIX

@cindex MMIX assembler directive TETRA

@cindex MMIX pseudo-op TETRA

@cindex assembler directive OCTA, MMIX

@cindex pseudo-op OCTA, MMIX

@cindex MMIX assembler directive OCTA

@cindex MMIX pseudo-op OCTA

@anchor{MMIX-constants}

The directives @samp{WYDE}, @samp{TETRA} and @samp{OCTA} emit constants of

two, four and eight bytes size respectively.  Before anything else happens

for the directive, the current location is aligned to the respective

constant-size boundary.  If a label is defined at the beginning of the

line, its value will be that after the alignment.  A single operand can be

omitted, defaulting to a zero value emitted for the directive.  Operands

can be expressed as strings (@pxref{Strings}), in which case each

character in the string is emitted as a separate constant of the size

indicated by the directive.

@item PREFIX

@cindex assembler directive PREFIX, MMIX

@cindex pseudo-op PREFIX, MMIX

@cindex MMIX assembler directive PREFIX

@cindex MMIX pseudo-op PREFIX

@anchor{MMIX-prefix}

The @samp{PREFIX} directive sets a symbol name prefix to be prepended to

all symbols (except local symbols, @pxref{MMIX-Symbols}), that are not

prefixed with @samp{:}, until the next @samp{PREFIX} directive.  Such

prefixes accumulate.  For example,

@smallexample

 PREFIX a

 PREFIX b

c IS 0

@end smallexample

defines a symbol @samp{abc} with the value 0.

@item BSPEC

@itemx ESPEC

@cindex assembler directive BSPEC, MMIX

@cindex pseudo-op BSPEC, MMIX

@cindex MMIX assembler directive BSPEC

@cindex MMIX pseudo-op BSPEC

@cindex assembler directive ESPEC, MMIX

@cindex pseudo-op ESPEC, MMIX

@cindex MMIX assembler directive ESPEC

@cindex MMIX pseudo-op ESPEC

@anchor{MMIX-spec}

A pair of @samp{BSPEC} and @samp{ESPEC} directives delimit a section of

special contents (without specified semantics).  Example:

@smallexample

 BSPEC 42

 TETRA 1,2,3

 ESPEC

@end smallexample

The single operand to @samp{BSPEC} must be number in the range

0@dots{}255.  The @samp{BSPEC} number 80 is used by the GNU binutils

implementation.

@end table

@node MMIX-mmixal

@section Differences to @code{mmixal}

@cindex mmixal differences

@cindex differences, mmixal

The binutils @code{@value{AS}} and @code{@value{LD}} combination has a few

differences in function compared to @code{mmixal} (@pxref{mmixsite}).

The replacement of a symbol with a GREG-allocated register

(@pxref{GREG-base}) is not handled the exactly same way in

@code{@value{AS}} as in @code{mmixal}.  This is apparent in the

@code{mmixal} example file @code{inout.mms}, where different registers

with different offsets, eventually yielding the same address, are used in

the first instruction.  This type of difference should however not affect

the function of any program unless it has specific assumptions about the

allocated register number.

Line numbers (in the @samp{mmo} object format) are currently not

supported.

Expression operator precedence is not that of mmixal: operator precedence

is that of the C programming language.  It's recommended to use

parentheses to explicitly specify wanted operator precedence whenever more

than one type of operators are used.

The serialize unary operator @code{&}, the fractional division operator

@samp{//}, the logical not operator @code{!} and the modulus operator

@samp{%} are not available.

Symbols are not global by default, unless the option

@samp{--globalize-symbols} is passed.  Use the @samp{.global} directive to

globalize symbols (@pxref{Global}).

Operand syntax is a bit stricter with @code{@value{AS}} than

@code{mmixal}.  For example, you can't say @code{addu 1,2,3}, instead you

must write @code{addu $1,$2,3}.

You can't LOC to a lower address than those already visited

(i.e., ``backwards'').

A LOC directive must come before any emitted code.

Predefined symbols are visible as file-local symbols after use.  (In the

ELF file, that is---the linked mmo file has no notion of a file-local

symbol.)

Some mapping of constant expressions to sections in LOC expressions is

attempted, but that functionality is easily confused and should be avoided

unless compatibility with @code{mmixal} is required.  A LOC expression to

@samp{0x2000000000000000} or higher, maps to the @samp{.data} section and

lower addresses map to the @samp{.text} section (@pxref{MMIX-loc}).

The code and data areas are each contiguous.  Sparse programs with

far-away LOC directives will take up the same amount of space as a

contiguous program with zeros filled in the gaps between the LOC

directives.  If you need sparse programs, you might try and get the wanted

effect with a linker script and splitting up the code parts into sections

(@pxref{Section}).  Assembly code for this, to be compatible with

@code{mmixal}, would look something like:

@smallexample

 .if 0

 LOC away_expression

 .else

 .section away,"ax"

 .fi

@end smallexample

@code{@value{AS}} will not execute the LOC directive and @code{mmixal}

ignores the lines with @code{.}.  This construct can be used generally to

help compatibility.

Symbols can't be defined twice--not even to the same value.

Instruction mnemonics are recognized case-insensitive, though the

@samp{IS} and @samp{GREG} pseudo-operations must be specified in

upper-case characters.

There's no unicode support.

The following is a list of programs in @samp{mmix.tar.gz}, available at

@url{http://www-cs-faculty.stanford.edu/~knuth/mmix-news.html}, last

checked with the version dated 2001-08-25 (md5sum

c393470cfc86fac040487d22d2bf0172) that assemble with @code{mmixal} but do

not assemble with @code{@value{AS}}:

@table @code

@item silly.mms

LOC to a previous address.

@item sim.mms

Redefines symbol @samp{Done}.

@item test.mms

Uses the serial operator @samp{&}.

@end table