binutils-gdb/gas/doc/c-nds32.texi

300 lines
7.6 KiB
Plaintext

@c Copyright (C) 2013-2014 Free Software Foundation, Inc.
@c This is part of the GAS manual.
@c For copying conditions, see the file as.texinfo.
@c man end
@ifset GENERIC
@page
@node NDS32-Dependent
@chapter NDS32 Dependent Features
@end ifset
@ifclear GENERIC
@node Machine Dependencies
@chapter NDS32 Dependent Features
@end ifclear
@cindex NDS32 processor
The NDS32 processors family includes high-performance and low-power 32-bit
processors for high-end to low-end. @sc{gnu} @code{@value{AS}} for NDS32
architectures supports NDS32 ISA version 3. For detail about NDS32
instruction set, please see the AndeStar ISA User Manual which is availible
at http://www.andestech.com/en/index/index.htm
@menu
* NDS32 Options:: Assembler options
* NDS32 Syntax:: High-level assembly macros
@end menu
@node NDS32 Options
@section NDS32 Options
@cindex NDS32 options
@cindex options for NDS32
The NDS32 configurations of @sc{gnu} @code{@value{AS}} support these
special options:
@c man begin OPTIONS
@table @code
@item -O1
Optimize for performance.
@item -Os
Optimize for space.
@item -EL
Produce little endian data output.
@item -EB
Produce little endian data output.
@item -mpic
Generate PIC.
@item -mno-fp-as-gp-relax
Suppress fp-as-gp relaxation for this file.
@item -mb2bb-relax
Back-to-back branch optimization.
@item -mno-all-relax
Suppress all relaxation for this file.
@item -march=<arch name>
Assemble for architecture <arch name> which could be v3, v3j, v3m, v3f,
v3s, v2, v2j, v2f, v2s.
@item -mbaseline=<baseline>
Assemble for baseline <baseline> which could be v2, v3, v3m.
@item -mfpu-freg=@var{FREG}
Specify a FPU configuration.
@table @code
@item 0 8 SP / 4 DP registers
@item 1 16 SP / 8 DP registers
@item 2 32 SP / 16 DP registers
@item 3 32 SP / 32 DP registers
@end table
@item -mabi=@var{abi}
Specify a abi version <abi> could be v1, v2, v2fp, v2fpp.
@item -m[no-]mac
Enable/Disable Multiply instructions support.
@item -m[no-]div
Enable/Disable Divide instructions support.
@item -m[no-]16bit-ext
Enable/Disable 16-bit extension
@item -m[no-]dx-regs
Enable/Disable d0/d1 registers
@item -m[no-]perf-ext
Enable/Disable Performance extension
@item -m[no-]perf2-ext
Enable/Disable Performance extension 2
@item -m[no-]string-ext
Enable/Disable String extension
@item -m[no-]reduced-regs
Enable/Disable Reduced Register configuration (GPR16) option
@item -m[no-]audio-isa-ext
Enable/Disable AUDIO ISA extension
@item -m[no-]fpu-sp-ext
Enable/Disable FPU SP extension
@item -m[no-]fpu-dp-ext
Enable/Disable FPU DP extension
@item -m[no-]fpu-fma
Enable/Disable FPU fused-multiply-add instructions
@item -mall-ext
Turn on all extensions and instructions support
@end table
@c man end
@node NDS32 Syntax
@section Syntax
@menu
* NDS32-Chars:: Special Characters
* NDS32-Regs:: Register Names
* NDS32-Ops:: Pseudo Instructions
@end menu
@node NDS32-Chars
@subsection Special Characters
Use @samp{#} at column 1 and @samp{!} anywhere in the line except inside
quotes.
Multiple instructions in a line are allowed though not recommended and
should be separated by @samp{;}.
Assembler is not case-sensitive in general except user defined label.
For example, @samp{jral F1} is different from @samp{jral f1} while it is
the same as @samp{JRAL F1}.
@node NDS32-Regs
@subsection Register Names
@table @code
@item General purpose registers (GPR)
There are 32 32-bit general purpose registers $r0 to $r31.
@item Accumulators d0 and d1
64-bit accumulators: $d0.hi, $d0.lo, $d1.hi, and $d1.lo.
@item Assembler reserved register $ta
Register $ta ($r15) is reserved for assembler using.
@item Operating system reserved registers $p0 and $p1
Registers $p0 ($r26) and $p1 ($r27) are used by operating system as scratch
registers.
@item Frame pointer $fp
Register $r28 is regarded as the frame pointer.
@item Global pointer
Register $r29 is regarded as the global pointer.
@item Link pointer
Register $r30 is regarded as the link pointer.
@item Stack pointer
Register $r31 is regarded as the stack pointer.
@end table
@node NDS32-Ops
@subsection Pseudo Instructions
@table @code
@item li rt5,imm32
load 32-bit integer into register rt5. @samp{sethi rt5,hi20(imm32)} and then
@samp{ori rt5,reg,lo12(imm32)}.
@item la rt5,var
Load 32-bit address of var into register rt5. @samp{sethi rt5,hi20(var)} and
then @samp{ori reg,rt5,lo12(var)}
@item l.[bhw] rt5,var
Load value of var into register rt5. @samp{sethi $ta,hi20(var)} and then
@samp{l[bhw]i rt5,[$ta+lo12(var)]}
@item l.[bh]s rt5,var
Load value of var into register rt5. @samp{sethi $ta,hi20(var)} and then
@samp{l[bh]si rt5,[$ta+lo12(var)]}
@item l.[bhw]p rt5,var,inc
Load value of var into register rt5 and increment $ta by amount inc.
@samp{la $ta,var} and then @samp{l[bhw]i.bi rt5,[$ta],inc}
@item l.[bhw]pc rt5,inc
Continue loading value of var into register rt5 and increment $ta by amount inc.
@samp{l[bhw]i.bi rt5,[$ta],inc.}
@item l.[bh]sp rt5,var,inc
Load value of var into register rt5 and increment $ta by amount inc.
@samp{la $ta,var} and then @samp{l[bh]si.bi rt5,[$ta],inc}
@item l.[bh]spc rt5,inc
Continue loading value of var into register rt5 and increment $ta by amount inc.
@samp{l[bh]si.bi rt5,[$ta],inc.}
@item s.[bhw] rt5,var
Store register rt5 to var.
@samp{sethi $ta,hi20(var)} and then @samp{s[bhw]i rt5,[$ta+lo12(var)]}
@item s.[bhw]p rt5,var,inc
Store register rt5 to var and increment $ta by amount inc.
@samp{la $ta,var} and then @samp{s[bhw]i.bi rt5,[$ta],inc}
@item s.[bhw]pc rt5,inc
Continue storing register rt5 to var and increment $ta by amount inc.
@samp{s[bhw]i.bi rt5,[$ta],inc.}
@item not rt5,ra5
Alias of @samp{nor rt5,ra5,ra5}.
@item neg rt5,ra5
Alias of @samp{subri rt5,ra5,0}.
@item br rb5
Depending on how it is assembled, it is translated into @samp{r5 rb5}
or @samp{jr rb5}.
@item b label
Branch to label depending on how it is assembled, it is translated into
@samp{j8 label}, @samp{j label}, or "@samp{la $ta,label} @samp{br $ta}".
@item bral rb5
Alias of jral br5 depending on how it is assembled, it is translated
into @samp{jral5 rb5} or @samp{jral rb5}.
@item bal fname
Alias of jal fname depending on how it is assembled, it is translated into
@samp{jal fname} or "@samp{la $ta,fname} @samp{bral $ta}".
@item call fname
Call function fname same as @samp{jal fname}.
@item move rt5,ra5
For 16-bit, this is @samp{mov55 rt5,ra5}.
For no 16-bit, this is @samp{ori rt5,ra5,0}.
@item move rt5,var
This is the same as @samp{l.w rt5,var}.
@item move rt5,imm32
This is the same as @samp{li rt5,imm32}.
@item pushm ra5,rb5
Push contents of registers from ra5 to rb5 into stack.
@item push ra5
Push content of register ra5 into stack. (same @samp{pushm ra5,ra5}).
@item push.d var
Push value of double-word variable var into stack.
@item push.w var
Push value of word variable var into stack.
@item push.h var
Push value of half-word variable var into stack.
@item push.b var
Push value of byte variable var into stack.
@item pusha var
Push 32-bit address of variable var into stack.
@item pushi imm32
Push 32-bit immediate value into stack.
@item popm ra5,rb5
Pop top of stack values into registers ra5 to rb5.
@item pop rt5
Pop top of stack value into register. (same as @samp{popm rt5,rt5}.)
@item pop.d var,ra5
Pop value of double-word variable var from stack using register ra5
as 2nd scratch register. (1st is $ta)
@item pop.w var,ra5
Pop value of word variable var from stack using register ra5.
@item pop.h var,ra5
Pop value of half-word variable var from stack using register ra5.
@item pop.b var,ra5
Pop value of byte variable var from stack using register ra5.
@end table