The following documentation discusses the TX39 processor.

• Compiler issues for TX39 targets
• ABI summary for the Toshiba TX39 targets
• Assembler issues for TX39 targets
• Linker issues for TX39 targets
• Debugger issues for TX39 targets
• Simulator issues for TX39 targets

Tool description	Tool name
GCC compiler	mips-tx39-elf-gcc
C++ compiler	mips-tx39-elf-c++
GAS assembler	mips-tx39-elf-as
GNU linker	mips-tx39-elf-ld
Standalone simulator	mips-tx39-elf-run
Binary utilities	mips-tx39-elf-ar mips-tx39-elf-nm  mips-tx39-elf-objcopy  mips-tx39-elf-objdump  mips-tx39-elf-ranlib  mips-tx39-elf-size  mips-tx39-elf-strings  mips-tx39-elf-strip
GDB debugger	mips-tx39-elf-gdb

• GNUPro Instruction Set simulator
• Densan DVE-R3900/20 evaluation board

Host	Vendor
SunOS 4.1.4 (SPARC)	Sun
Solaris 2.4-2.5.1 (SPARC)	Sun
Windows NT (x86)	Microsoft
Windows 95 (x86)	Microsoft

-msingle-float
This option causes the compiler to generate output containing floating point instructions for single precision floating point operations, and library calls for double precision floating point operations.

-EB

-EL
Any MIPS configuration of the compiler can select big-endian or little-endian output at run time. Use -EB to select big-endian output and -EL for little-endian. If neither -EL nor -EB are defined, big-endian is the default.

__mips__
__R3000__
Each of these is always defined.
__MIPSEL__
__MIPSEB__
One of these will be defined depending on the endian compilation mode.
__mips_single_float
__mips_soft_float
One of these will be defined depending on the floating point compilation mode.

Type	Size (bytes)	Alignment (bytes)
char :	1 byte	1 byte
short :	2 bytes	2 bytes
int :	4 bytes	4 bytes
unsigned :	4 bytes	4 bytes
long :	4 bytes	4 bytes
long long :	8 bytes	8 bytes
float :	4 bytes	4 bytes
double :	8 bytes	8 bytes
pointer :	4 bytes	4 bytes

• Alignment within aggregates (structs and unions) is as above, with padding added if needed
• Aggregates have alignment equal to that of their most aligned member
• Aggregates have sizes which are a multiple of their alignment

Parameter registers
general-purpose	r4 through r11
floating point	f12 through f19

 

Register usage
fixed 0 value	r0
volatile	r1 through r15, as well as r24 and r25
non-volatile	r16 through r23,as well as r30
kernel reserved	r26, r27
gp (SDA base)	r28
stack pointer	r29
frame pointer	r30 (if needed)
return address	r31

•  General-purpose and floating point parameter registers are allocated independently.
• Structures that are less than or equal to 32 bits are passed as values.
• Structures that are greater than 32 bits are passed as pointers.

The following documentation describes the TX39 stack frame.

• The stack grows downwards from high addresses to low addresses.
• A leaf function need not allocate a stack frame if it does not need one.
• A frame pointer need not be allocated.
• The stack pointer shall always be aligned to 8 byte boundaries.

Consider the parameters in a function call as ordered from left (first parameter) to right. In the following algorithm, FR contains the number of the next available floating-point register (or register pair for modes in which floating-point registers hold only 32 bits). GR contains the number of the next available general-purpose register. STARG is the address of the next available stack parameter word.

INITIALIZE

SCAN

DOUBLE OR FLOAT

SIMPLE ARG

• One of the simple integer types which will fit into a general-purpose register
• A pointer to an object of any type
• A struct or union small enough to fit in a register
• A larger struct or union, which shall be treated as a pointer to the object or to a copy of the object (See below for when copies are made.)

LONG LONG in 32-bit mode

STACK

	ANSI mode	K&R mode
Normal param	Callee copies if needed	Caller copies
Varargs (...) param	Caller copies	Caller copies

Data type	Register
int :	r2
short :	r2
long :	r2
long long :	r2 through r3 (32-bit mode)
float :	f0
double :	f0 through f1 (32-bit mode)
struct/union :	(see following discussion)

Assembler issues for TX39 targets
 

Use -EB to select big-endian output, and -EL for little-endian output. The default is big-endian.

 

Symbol	Register
$at	$1
$kt0	$26
$kt1	$27
$gp	$28
$sp	$29
$fp	$30

Assembler directives for TX39 targets
 

.abicalls	.dcb.b	.fail	.irepc	.psize
.abort	.dcb.d	.file	.irp	.purgem
.aent	.dcb.l	.fill	.irpc	.quad
.align	.dcb.s	.float	.lcomm	.rdata
.appfile	.dcb.w	.fmask	.lflags	.rep
.appline	.dcb.x	.format	.linkonce	.rept
.ascii	.debug	.frame	.list	.rva
.asciiz	.double	.global	.livereg	.sbttl
.asciz	.ds	.globl	.llen	.sdata
.balign	.ds.b	.gpword	.loc	.set
.balignl	.ds.d	.half	.long	.short
.balignw	.ds.l	.hword	.lsym	.single
.bgnb	.ds.p	.if	.macro	.skip
.bss	.ds.s	.ifc	.mask	.space
.byte	.ds.w	.ifdef	.mexit	.spc
.comm	.ds.x	.ifeq	.mri	.stabd
.common	.dword	.ifeqs	.name	.stabn
.common.s	.eject	.ifge	.noformat	.stabs
.cpadd	.else	.ifgt	.nolist	.string
.cpload	.elsec	.ifle	.nopage	.struct
.cprestore	.end	.iflt	.octa	.text
.data	.endb	.ifnc	.offset	.title
.dc	.endc	.ifndef	.option	.ttl
.dc.b	.endif	.ifne	.org	.verstamp
.dc.d	.ent	.ifnes	.p2align	.word
.dc.l	.equ	.ifnotdef	.p2alignl	.xcom
.dc.s	.equiv	.include	.p2alignw	.xdef
.dc.w	.err	.insn	.page	.xref
.dc.x	.exitm	.int	.plen	.xstabs
.dcb	.extern	.irep	.print	.zero

MIPS synthetic instructions supported for the TX39
 

Instruction	Expansion
abs R R	bgez    $a0,0x5c move    $t1,$a0  neg     $t1,$a0
div R R R	div     $zero,$t1,$a0 bnez    $a0,0x120  nop  break   0x7  li      $at,-1  bne     $a0,$at,0x138  lui     $at,0x8000  bne     $t1,$at,0x138  nop  break   0x6  mflo    $a3
div R R	div     $zero,$a3,$t1 bnez    $t1,0x154  nop  break   0x7  li      $at,-1  bne     $t1,$at,0x16c  lui     $at,0x8000  bne     $a3,$at,0x16c  nop  break   0x6  mflo    $a3
div R R I	li      $at,<I> div     $zero,$a3,$at  mflo    $a1
divu R R R	divu    $zero,$a0,$a2 bnez    $a2,0x1cc  nop  break   0x7  mflo    $t1
divu R R	divu    $zero,$t1,$a0 bnez    $a0,0x1e8  nop  break   0x7  mflo    $t1
divu R R I	li      $at,<I> divu    $zero,$v1,$at  mflo    $t0
jal E	jalr    <E>
lwc0 R I	ll      <R> <I>
mul R R R	multu   $t0,$v1 mflo    $a2
mul R R I	li      $at,<I> mult    $t0,$at  mflo    $a2
mulo R R R	mult    $t0,$v1 mflo    $a2  sra     $a2,$a2,0x1f  mfhi    $at  beq     $a2,$at,0x440  nop  break   0x6  mflo    $a2
mulou R R R	multu   $t0,$v1 mfhi    $at  mflo    $a2  beqz    $at,0x464  nop  break   0x6
nor R R I	ori     $t1,$a0,0xe nor     $t1,$t1,$zero
not R R	nor     $t1,$a0,$zero
or R R I	ori     $a2,$t0,<I>
rem R R R	div     $zero,$t1,$a0 bnez    $a0,0x4ec  nop  break   0x7  li      $at,-1  bne     $a0,$at,0x504  lui     $at,0x8000  bne     $t1,$at,0x504  nop  break   0x6  mfhi    $a3
rem R R I	li      $at,11 div     $zero,$t1,$at  mfhi    $a3
remu R R R	divu    $zero,$v1,$a1 bnez    $a1,0x564  nop  break   0x7  mfhi    $t0
remu R R I	li      $at,<I> divu    $zero,$a0,$at  mfhi    $t1
rol R R R	negu    $at,$a3 srlv    $at,$a1,$at  sllv    $v1,$a1,$a3  or      $v1,$v1,$at
rol R R I	sll     $at,$v1,<I> srl     $t0,$v1,0x11  or      $t0,$t0,$at
ror R R R	negu    $at,$a1 sllv    $at,$v1,$at  srlv    $t0,$v1,$a1  or      $t0,$t0,$at
ror R R I	srl     $at,$t0,<I> sll     $a2,$t0,0x11  or      $a2,$a2,$at
sd R E	sw      $a1,<E> sw      $a2,<E+4>
seq R R R	xor     $t1,$a0,$a2 sltiu   $t1,$t1,1
seq R R I	xori    $a3,$t1,<I> sltiu   $a3,$a3,1
sge R R R	slt     $a3,$t1,$a0 xori    $a3,$a3,0x1
sge R R I	slti    $a1,$a3,<I> xori    $a1,$a1,0x1
sgeu R R R	sltu    $a1,$a3,$t1 xori    $a1,$a1,0x1
sgeu R R I	sltiu   $a0,$a2,<I> xori    $a0,$a0,0x1
sgt R R R	slt     $a1,$t1,$a3
sgt R R I	li      $at,<I> slt     $v1,$at,$a1
sgtu R R R	sltu    $v1,$a3,$a1
sgtu R R I	li      $at,<I> sltu    $t1,$at,$a0
sle R R R	slt     $v1,$a3,$a1 xori    $v1,$v1,0x1
sle R R I	li      $at,<I> slt     $t0,$at,$v1  xori    $t0,$t0,0x1
sleu R R R	sltu    $t0,$a1,$v1 xori    $t0,$t0,0x1
sleu R R I	li      $at,<I> sltu    $a3,$at,$t1  xori    $a3,$a3,0x1
sll R R R	sllv    $t0,$v1,$a1
slt R R I	slti    $a3,$t1,<I>
sltu R R I	sltiu   $a3,$t1,<I>
sne R R R	xor     $a1,$a3,$t1 sltu    $a1,$zero,$a1
sne R R I	xori    $v1,$a1,<I> sltu    $v1,$zero,$v1
sra R R R	srav    $a0,$a2,$t0
srl R R R	srlv    $v1,$a1,$a3
sub R R I	addi    $a0,$a2,-<I>
subu R R I	addiu   $t1,$a0,-<I>
swc0 R E	sc      $a1,<E>
ulh R E	lb      $a0,<E> lbu     $at,<E+1>  sll     $a0,$a0,0x8  or      $a0,$a0,$at
ulhu R E	lbu     $a0,<E> lbu     $at,<E+1>  sll     $a0,$a0,0x8  or      $a0,$a0,$at
ulw R E	lwl     $a2,<E> lwr     $a2,<E+K>
ush R E	sb      $a2,<E+K> srl     $at,$a2,0x8  sb      $at,<E>
usw R E	swl     $a2,<E> swr     $a2,<E+K>
xor R R I	xori    $a2,$t0,0x10

Linker issues for TX39 targets
 

 

 

/* Linker script for Densan DVE-R3900/20A board */  ENTRY(_start)   OUTPUT_ARCH("mips:3000")   OUTPUT_FORMAT("elf32-bigmips", "elf32-bigmips", "elf32-littlemips")   GROUP(-lc -ldve -lgcc)   SEARCH_DIR(.)   __DYNAMIC = 0;   /*    * Allocate the stack to be at the top of memory, since the    * stack grows down    */   PROVIDE (__stack = 0);   /* PROVIDE (__global = 0); */   /*    * Initalize some symbols to be zero so we can reference them    * in the crt0 without core dumping. These functions are all    * optional, but we do this so we can have our crt0 always use    * them if they exist.    * This is so BSPs work better when using the crt0 installed    * with gcc.    * We have to initalize them twice, so we multiple object file    * formats, as some prepend an underscore.    */   PROVIDE (hardware_init_hook = 0);   PROVIDE (software_init_hook = 0);     SECTIONS   {    . = 0xA0040000;    .text : {      _ftext = . ;      *(.init)      eprol = .;      *(.text)      *(.mips16.fn.*)      *(.mips16.call.*)      PROVIDE (__runtime_reloc_start = .);      *(.rel.sdata)      PROVIDE (__runtime_reloc_stop = .);      *(.fini)      etext = .;      _etext = .;   }   . = .;   .rdata : {      *(.rdata)   }    _fdata = ALIGN(16);    .data : {      *(.data)      CONSTRUCTORS   }    . = ALIGN(8);   _gp = . + 0x8000;   __global = _gp;   .lit8 : {     *(.lit8)   }   .lit4 : {     *(.lit4)   }   .sdata : {     *(.sdata)   }   . = ALIGN(4);   edata = .;   _edata = .;   _fbss = .;   .sbss : {     *(.sbss)     *(.scommon)   }   .bss : {     _bss_start = . ;     *(.bss)     *(COMMON)   }   end = .;   _end = .;     /* DWARF debug sections.     Symbols in the DWARF debugging sections are relative to     the beginning of the section so we begin them at 0. */   /* DWARF 1 */   .debug 0 : { *(.debug) }   .line 0 : { *(.line) }     /* GNU DWARF 1 extensions */   .debug_srcinfo 0 : { *(.debug_srcinfo) }   .debug_sfnames 0 : { *(.debug_sfnames) }   /* DWARF 1.1 and DWARF 2 */   .debug_aranges 0 : { *(.debug_aranges) }   .debug_pubnames 0 : { *(.debug_pubnames) }   /* DWARF 2 */   .debug_info 0 : { *(.debug_info) }   .debug_abbrev 0 : { *(.debug_abbrev) }   .debug_line 0 : { *(.debug_line) }   .debug_frame 0 : { *(.debug_frame) }   .debug_str 0 : { *(.debug_str) }   .debug_loc 0 : { *(.debug_loc) }   .debug_macinfo 0 : { *(.debug_macinfo) }    /* SGI/MIPS DWARF 2 extensions */    .debug_weaknames 0 : { *(.debug_weaknames) }    .debug_funcnames 0 : { *(.debug_funcnames) }    .debug_typenames 0 : { *(.debug_typenames) }    .debug_varnames 0 : { *(.debug_varnames) }  }

 

• Simulator

GDBs built-in software simulation of the TX39 processor allows the debugging of programs compiled for the TX39 without requiring any access to actual hardware. To activate this mode in GDB type target sim . Then load the code into the simulator by typing load and debug it in the normal fashion.
• Remote target board by serial connection

To connect to the target board in GDB, using the command
target remote <devicename> where <devicename>  will be a serial device such as /dev/ttya (Unix) or com2 (Windows NT). Then load the code onto the target board by using the command, load. After being downloaded, the program can be executed. The target command would be something like the following example's input after the (gdb) prompt.

• Remote target board by ethernet connection :

Connecting to the ethernet port with GDB is very similar to connecting to a serial port. The target command would be something like the following example's input after the (gdb) prompt.

 

% tx39-elf-run --help Usage: run [options] program [program args]  Options:   -l FILE, --log FILELog file   -n MODEL, --name MODELSelect arch model   -p[on|off], --profile [on|off]  Enable profiling   -t[on|off], --trace [on|off]  Enable tracing   -z FILE, --tracefile FILEWrite trace to file   -y FREQ, --frequency FREQProfile frequency   -y SIZE, --samples SIZE Profile sample size

 

% tx39-elf-run --trace hello

Hello, world!

3 + 4 = 7

The following output shows the first 10 lines of the file produced by the previous command.

2 a0040004 ; width 4 ; load instruction 2 a0040008 ; width 4 ; load instruction  2 a004000c ; width 4 ; load instruction  2 a0040010 ; width 4 ; load instruction  2 a0040014 ; width 4 ; load instruction  2 a0040018 ; width 4 ; load instruction  2 a004001c ; width 4 ; load instruction  2 a0040020 ; width 4 ; load instruction  2 a0040024 ; width 4 ; load instruction  2 a0040028 ; width 4 ; load instruction

% mips-tx39-elf-run --trace --tracefile=trace.out hello

Placing trace information into file "trace.out"

Hello, world!

3 + 4 = 7

--profile [on|off]
This option creates a file called gmon.out that contains profiling information. This file can be used as input to gprof, the GNU profiler.

% mips-tx39-elf-run --trace hello

Hello, world!

3 + 4 = 7

--frequency <frequency>
By default, the simulator samples the running program every 256 instructions. This option allows you to change this profiling frequency to some other number. Smaller numbers increasing the accuracy of the profile, but make the simulator run slightly slower. Also, because the counters used in the profile are only 16 bits, a high sampling frequency may cause the counters to overflow.

% mips-tx39-elf-run --profile --frequency 128 hello

Hello, world!

3 + 4 = 7

--samples <size>
By default, the simulator uses a profiling sample size of 131072 (128K). This option allows you to change the sample <size>. Increasing the sample size will make the profile more accurate, but will also increase the size of the profile output file (gmon.out).

The simulator rounds the sample size up the next power of two.

% mips-tx39-elf-run --profile --samples 20000 hello

Hello, world!

3 + 4 = 7

---

Top|Contents|Index|Previous|Next <

© 2005 Red Hat, Inc. All rights reserved. About Red Hat  :  Legal statement  :  Privacy statement  :  Contact