Subversion Repositories openrisc

@c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,

@c 2000, 2003, 2004, 2006

@c Free Software Foundation, Inc.

@c This is part of the GAS manual.

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

@ifset GENERIC

@page

@node M32R-Dependent

@chapter M32R Dependent Features

@end ifset

@ifclear GENERIC

@node Machine Dependencies

@chapter M32R Dependent Features

@end ifclear

@cindex M32R support

@menu

* M32R-Opts::                   M32R Options

* M32R-Directives::             M32R Directives

* M32R-Warnings::               M32R Warnings

@end menu

@node M32R-Opts

@section M32R Options

@cindex options, M32R

@cindex M32R options

The Renease M32R version of @code{@value{AS}} has a few machine

dependent options:

@table @code

@item -m32rx

@cindex @samp{-m32rx} option, M32RX

@cindex architecture options, M32RX

@cindex M32R architecture options

@code{@value{AS}} can assemble code for several different members of the

Renesas M32R family.  Normally the default is to assemble code for

the M32R microprocessor.  This option may be used to change the default

to the M32RX microprocessor, which adds some more instructions to the

basic M32R instruction set, and some additional parameters to some of

the original instructions.

@item -m32r2

@cindex @samp{-m32rx} option, M32R2

@cindex architecture options, M32R2

@cindex M32R architecture options

This option changes the target processor to the the M32R2

microprocessor.

@item -m32r

@cindex @samp{-m32r} option, M32R

@cindex architecture options, M32R

@cindex M32R architecture options

This option can be used to restore the assembler's default behaviour of

assembling for the M32R microprocessor.  This can be useful if the

default has been changed by a previous command line option.

@item -little

@cindex @code{-little} option, M32R

This option tells the assembler to produce little-endian code and

data.  The default is dependent upon how the toolchain was

configured.

@item -EL

@cindex @code{-EL} option, M32R

This is a synonym for @emph{-little}.

@item -big

@cindex @code{-big} option, M32R

This option tells the assembler to produce big-endian code and

data.

@item -EB

@cindex @code{-EB} option, M32R

This is a synonum for @emph{-big}.

@item -KPIC

@cindex @code{-KPIC} option, M32R

@cindex PIC code generation for M32R

This option specifies that the output of the assembler should be

marked as position-independent code (PIC).

@item -parallel

@cindex @code{-parallel} option, M32RX

This option tells the assembler to attempts to combine two sequential

instructions into a single, parallel instruction, where it is legal to

do so.

@item -no-parallel

@cindex @code{-no-parallel} option, M32RX

This option disables a previously enabled @emph{-parallel} option.

@item -no-bitinst

@cindex @samp{-no-bitinst}, M32R2

This option disables the support for the extended bit-field

instructions provided by the M32R2.  If this support needs to be

re-enabled the @emph{-bitinst} switch can be used to restore it.

@item -O

@cindex @code{-O} option, M32RX

This option tells the assembler to attempt to optimize the

instructions that it produces.  This includes filling delay slots and

converting sequential instructions into parallel ones.  This option

implies @emph{-parallel}.

@item -warn-explicit-parallel-conflicts

@cindex @samp{-warn-explicit-parallel-conflicts} option, M32RX

Instructs @code{@value{AS}} to produce warning messages when

questionable parallel instructions are encountered.  This option is

enabled by default, but @code{@value{GCC}} disables it when it invokes

@code{@value{AS}} directly.  Questionable instructions are those whose

behaviour would be different if they were executed sequentially.  For

example the code fragment @samp{mv r1, r2 || mv r3, r1} produces a

different result from @samp{mv r1, r2 \n mv r3, r1} since the former

moves r1 into r3 and then r2 into r1, whereas the later moves r2 into r1

and r3.

@item -Wp

@cindex @samp{-Wp} option, M32RX

This is a shorter synonym for the @emph{-warn-explicit-parallel-conflicts}

option.

@item -no-warn-explicit-parallel-conflicts

@cindex @samp{-no-warn-explicit-parallel-conflicts} option, M32RX

Instructs @code{@value{AS}} not to produce warning messages when

questionable parallel instructions are encountered.

@item -Wnp

@cindex @samp{-Wnp} option, M32RX

This is a shorter synonym for the @emph{-no-warn-explicit-parallel-conflicts}

option.

@item -ignore-parallel-conflicts

@cindex @samp{-ignore-parallel-conflicts} option, M32RX

This option tells the assembler's to stop checking parallel

instructions for constraint violations.  This ability is provided for

hardware vendors testing chip designs and should not be used under

normal circumstances.

@item -no-ignore-parallel-conflicts

@cindex @samp{-no-ignore-parallel-conflicts} option, M32RX

This option restores the assembler's default behaviour of checking

parallel instructions to detect constraint violations.

@item -Ip

@cindex @samp{-Ip} option, M32RX

This is a shorter synonym for the @emph{-ignore-parallel-conflicts}

option.

@item -nIp

@cindex @samp{-nIp} option, M32RX

This is a shorter synonym for the @emph{-no-ignore-parallel-conflicts}

option.

@item -warn-unmatched-high

@cindex @samp{-warn-unmatched-high} option, M32R

This option tells the assembler to produce a warning message if a

@code{.high} pseudo op is encountered without a matching @code{.low}

pseudo op.  The presence of such an unmatched pseudo op usually

indicates a programming error.

@item -no-warn-unmatched-high

@cindex @samp{-no-warn-unmatched-high} option, M32R

Disables a previously enabled @emph{-warn-unmatched-high} option.

@item -Wuh

@cindex @samp{-Wuh} option, M32RX

This is a shorter synonym for the @emph{-warn-unmatched-high} option.

@item -Wnuh

@cindex @samp{-Wnuh} option, M32RX

This is a shorter synonym for the @emph{-no-warn-unmatched-high} option.

@end table

@node M32R-Directives

@section M32R Directives

@cindex directives, M32R

@cindex M32R directives

The Renease M32R version of @code{@value{AS}} has a few architecture

specific directives:

@table @code

@cindex @code{low} directive, M32R

@item low @var{expression}

The @code{low} directive computes the value of its expression and

places the lower 16-bits of the result into the immediate-field of the

instruction.  For example:

@smallexample

   or3   r0, r0, #low(0x12345678) ; compute r0 = r0 | 0x5678

   add3, r0, r0, #low(fred)   ; compute r0 = r0 + low 16-bits of address of fred

@end smallexample

@item high @var{expression}

@cindex @code{high} directive, M32R

The @code{high} directive computes the value of its expression and

places the upper 16-bits of the result into the immediate-field of the

instruction.  For example:

@smallexample

   seth  r0, #high(0x12345678) ; compute r0 = 0x12340000

   seth, r0, #high(fred)       ; compute r0 = upper 16-bits of address of fred

@end smallexample

@item shigh @var{expression}

@cindex @code{shigh} directive, M32R

The @code{shigh} directive is very similar to the @code{high}

directive.  It also computes the value of its expression and places

the upper 16-bits of the result into the immediate-field of the

instruction.  The difference is that @code{shigh} also checks to see

if the lower 16-bits could be interpreted as a signed number, and if

so it assumes that a borrow will occur from the upper-16 bits.  To

compensate for this the @code{shigh} directive pre-biases the upper

16 bit value by adding one to it.  For example:

For example:

@smallexample

   seth  r0, #shigh(0x12345678) ; compute r0 = 0x12340000

   seth  r0, #shigh(0x00008000) ; compute r0 = 0x00010000

@end smallexample

In the second example the lower 16-bits are 0x8000.  If these are

treated as a signed value and sign extended to 32-bits then the value

becomes 0xffff8000.  If this value is then added to 0x00010000 then

the result is 0x00008000.

This behaviour is to allow for the different semantics of the

@code{or3} and @code{add3} instructions.  The @code{or3} instruction

treats its 16-bit immediate argument as unsigned whereas the

@code{add3} treats its 16-bit immediate as a signed value.  So for

example:

@smallexample

   seth  r0, #shigh(0x00008000)

   add3  r0, r0, #low(0x00008000)

@end smallexample

Produces the correct result in r0, whereas:

@smallexample

   seth  r0, #shigh(0x00008000)

   or3   r0, r0, #low(0x00008000)

@end smallexample

Stores 0xffff8000 into r0.

Note - the @code{shigh} directive does not know where in the assembly

source code the lower 16-bits of the value are going set, so it cannot

check to make sure that an @code{or3} instruction is being used rather

than an @code{add3} instruction.  It is up to the programmer to make

sure that correct directives are used.

@cindex @code{.m32r} directive, M32R

@item .m32r

The directive performs a similar thing as the @emph{-m32r} command

line option.  It tells the assembler to only accept M32R instructions

from now on.  An instructions from later M32R architectures are

refused.

@cindex @code{.m32rx} directive, M32RX

@item .m32rx

The directive performs a similar thing as the @emph{-m32rx} command

line option.  It tells the assembler to start accepting the extra

instructions in the M32RX ISA as well as the ordinary M32R ISA.

@cindex @code{.m32r2} directive, M32R2

@item .m32r2

The directive performs a similar thing as the @emph{-m32r2} command

line option.  It tells the assembler to start accepting the extra

instructions in the M32R2 ISA as well as the ordinary M32R ISA.

@cindex @code{.little} directive, M32RX

@item .little

The directive performs a similar thing as the @emph{-little} command

line option.  It tells the assembler to start producing little-endian

code and data.  This option should be used with care as producing

mixed-endian binary files is fraught with danger.

@cindex @code{.big} directive, M32RX

@item .big

The directive performs a similar thing as the @emph{-big} command

line option.  It tells the assembler to start producing big-endian

code and data.  This option should be used with care as producing

mixed-endian binary files is fraught with danger.

@end table

@node M32R-Warnings

@section M32R Warnings

@cindex warnings, M32R

@cindex M32R warnings

There are several warning and error messages that can be produced by

@code{@value{AS}} which are specific to the M32R:

@table @code

@item output of 1st instruction is the same as an input to 2nd instruction - is this intentional ?

This message is only produced if warnings for explicit parallel

conflicts have been enabled.  It indicates that the assembler has

encountered a parallel instruction in which the destination register of

the left hand instruction is used as an input register in the right hand

instruction.  For example in this code fragment

@samp{mv r1, r2 || neg r3, r1} register r1 is the destination of the

move instruction and the input to the neg instruction.

@item output of 2nd instruction is the same as an input to 1st instruction - is this intentional ?

This message is only produced if warnings for explicit parallel

conflicts have been enabled.  It indicates that the assembler has

encountered a parallel instruction in which the destination register of

the right hand instruction is used as an input register in the left hand

instruction.  For example in this code fragment

@samp{mv r1, r2 || neg r2, r3} register r2 is the destination of the

neg instruction and the input to the move instruction.

@item instruction @samp{...} is for the M32RX only

This message is produced when the assembler encounters an instruction

which is only supported by the M32Rx processor, and the @samp{-m32rx}

command line flag has not been specified to allow assembly of such

instructions.

@item unknown instruction @samp{...}

This message is produced when the assembler encounters an instruction

which it does not recognize.

@item only the NOP instruction can be issued in parallel on the m32r

This message is produced when the assembler encounters a parallel

instruction which does not involve a NOP instruction and the

@samp{-m32rx} command line flag has not been specified.  Only the M32Rx

processor is able to execute two instructions in parallel.

@item instruction @samp{...} cannot be executed in parallel.

This message is produced when the assembler encounters a parallel

instruction which is made up of one or two instructions which cannot be

executed in parallel.

@item Instructions share the same execution pipeline

This message is produced when the assembler encounters a parallel

instruction whoes components both use the same execution pipeline.

@item Instructions write to the same destination register.

This message is produced when the assembler encounters a parallel

instruction where both components attempt to modify the same register.

For example these code fragments will produce this message:

@samp{mv r1, r2 || neg r1, r3}

@samp{jl r0 || mv r14, r1}

@samp{st r2, @@-r1 || mv r1, r3}

@samp{mv r1, r2 || ld r0, @@r1+}

@samp{cmp r1, r2 || addx r3, r4} (Both write to the condition bit)

@end table