Merge commit '57999ac9b9d32775efc271e0edfed27b784ee1c6' as 'parallel-rsp'

parallel-rsp/debug-toolchain/Makefile

@@ -0,0 +1,6 @@
+RSP_TEST = main
+
+PROGRAM = $(RSP_TEST)
+MIPS_OBJ = $(RSP_TEST).o
+
+include Makefile.mips

parallel-rsp/debug-toolchain/Makefile.mips

@@ -0,0 +1,40 @@
+####
+## Basic Makefile for RSP-MIPS
+##
+######
+
+TARGET_ELF = $(PROGRAM).elf
+TARGET_BIN = $(PROGRAM).bin
+TARGET_GLOBAL_BIN = $(PROGRAM).global.bin
+MIPS_LD_SCRIPT = rsp-mips.ld
+
+MIPS_OBJCOPY = mipsel-linux-gnu-objcopy
+MIPS_CC = mipsel-linux-gnu-gcc
+MIPS_AS = mipsel-linux-gnu-as
+MIPS_LD = mipsel-linux-gnu-ld
+CRT_OBJ = start.o rsp-mips.o
+
+all: $(TARGET_BIN) $(TARGET_GLOBAL_BIN)
+
+crt-obj: $(CRT_OBJ)
+
+$(TARGET_BIN): $(TARGET_ELF)
+	$(MIPS_OBJCOPY) -j .text $< $(TARGET_BIN) -O binary
+
+$(TARGET_GLOBAL_BIN): $(TARGET_ELF)
+	$(MIPS_OBJCOPY) -j .data $< $(TARGET_GLOBAL_BIN) -O binary
+
+$(TARGET_ELF): $(MIPS_OBJ) $(CRT_OBJ)
+	$(MIPS_LD) -T $(MIPS_LD_SCRIPT) -o $@ $(CRT_OBJ) $(MIPS_OBJ) -EB
+
+%.o: %.s
+	$(MIPS_AS) -o $@ $< -EB -mabi=eabi -march=mips1
+
+%.o: %.c rsp-mips.h
+	$(MIPS_CC) -c -o $@ $< -Os -EB -march=mips1 -mabi=eabi -mno-abicalls -std=gnu99 -nostdlib
+
+clean:
+	rm -f $(MIPS_OBJ) $(CRT_OBJ) $(TARGET_ELF) $(TARGET_HEX) $(TARGET_GLOBAL_HEX) $(TARGET_BIN) $(TARGET_GLOBAL_BIN)
+
+.PHONY: all clean tools
+

parallel-rsp/debug-toolchain/add.s

@@ -0,0 +1,13 @@
+.text
+.global main
+main:
+	li $t0, 40
+	li $t1, -20
+	add $t2, $t0, $t1
+	addu $t3, $t0, $t1
+
+	li $t0, -5
+	li $t1, -4
+	add $t2, $t0, $t1
+	addu $t3, $t0, $t1
+	break

parallel-rsp/debug-toolchain/addi.s

@@ -0,0 +1,9 @@
+.text
+.global main
+main:
+	li $a0, 17
+	addi $t0, $a0, -8
+	addi $t1, $a0, 8
+	addiu $t2, $a0, -8
+	addiu $t3, $a0, 8
+	break

parallel-rsp/debug-toolchain/and.s

@@ -0,0 +1,7 @@
+.text
+.global main
+main:
+	li $t0, 17
+	li $t1, 5
+	and $t2, $t0, $t1
+	break

parallel-rsp/debug-toolchain/andi.s

@@ -0,0 +1,6 @@
+.text
+.global main
+main:
+	li $a0, -1
+	andi $t0, $a0, 50000
+	break

parallel-rsp/debug-toolchain/beq-impossible-delay-slot-both-taken.s

@@ -0,0 +1,20 @@
+# This works on the Lightning JIT, but hangs the reference.
+
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	li $t1, 1
+	li $t2, 2
+	li $t3, 0
+	beq $t0, $t1, taken0 # This will be taken.
+	beq $t0, $t1, taken1 # This will also be taken.
+	li $t3, 3 # This will not be executed.
+
+taken0:
+	li $t6, 5 # This will be executed, the second branch then kicks in and runs taken1 path.
+	break
+taken1:
+	li $t4, 4 # This will be executed.
+	break

parallel-rsp/debug-toolchain/beq-impossible-delay-slot-first-taken.s

@@ -0,0 +1,18 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	li $t1, 1
+	li $t2, 2
+	li $t3, 0
+	beq $t0, $t1, taken0 # This will be taken.
+	beq $t0, $t2, taken1 # Will not be taken.
+	li $t3, 3
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/beq-impossible-delay-slot-second-taken.s

@@ -0,0 +1,18 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	li $t1, 1
+	li $t2, 2
+	li $t3, 0
+	beq $t0, $t2, taken0 # This will not be taken, should be ignored.
+	beq $t0, $t1, taken1 # This will be taken.
+	li $t3, 3
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/bgez.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, -1
+	bgez $t0, taken1 # This will not be taken.
+	li $t2, 10
+
+	li $t0, 0
+	bgez $t0, taken1 # This will be taken.
+	li $t2, 20
+
+	li $t0, 1
+	bgez $t0, taken1 # This will be taken.
+	li $t2, 30
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/bgezal.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, -1
+	bgezal $t0, taken1 # This will not be taken.
+	li $t2, 10
+
+	li $t0, 0
+	bgezal $t0, taken1 # This will be taken.
+	li $t2, 20
+
+	li $t0, 1
+	bgezal $t0, taken1 # This will be taken.
+	li $t2, 30
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/bgtz.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, -1
+	bgtz $t0, taken1 # This will not be taken.
+	li $t2, 10
+
+	li $t0, 0
+	bgtz $t0, taken1 # This will not be taken.
+	li $t2, 20
+
+	li $t0, 1
+	bgtz $t0, taken1 # This will be taken.
+	li $t2, 30
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/blez.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	blez $t0, taken1 # This will not be taken.
+	li $t2, 10
+
+	li $t0, 0
+	blez $t0, taken1 # This will be taken.
+	li $t2, 20
+
+	li $t0, 0
+	blez $t0, taken1 # This will be taken.
+	li $t2, 30
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/bltz.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	bltz $t0, taken1 # This will not be taken.
+	li $t2, 10
+
+	li $t0, 0
+	bltz $t0, taken1 # This will not be taken.
+	li $t2, 20
+
+	li $t0, -1
+	bltz $t0, taken1 # This will be taken.
+	li $t2, 30
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/bltzal.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	bltzal $t0, taken1 # This will not be taken.
+	li $t2, 10
+
+	li $t0, 0
+	bltzal $t0, taken1 # This will not be taken.
+	li $t2, 20
+
+	li $t0, -1
+	bltzal $t0, taken1 # This will be taken.
+	li $t2, 30
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/bne.s

@@ -0,0 +1,20 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	li $t1, 1
+	li $t2, 2
+	li $t3, 0
+	bne $t0, $t1, taken0 # This will not be taken.
+	nop
+
+	bne $t0, $t2, taken1 # Will be taken.
+	li $t3, 3
+
+taken0:
+	li $t4, 5
+	break
+taken1:
+	li $t4, 4
+	break

parallel-rsp/debug-toolchain/bug-shl-into-branch.s

@@ -0,0 +1,15 @@
+.text
+.global main
+main:
+	li $t1, ((0x50000000) << 5) & 0xffffffff
+	li $t0, 0x50000000
+	sll $t0, $t0, 5
+	bne $t0, $t1, error
+	nop
+
+	li $t2, 1
+	break
+
+error:
+	li $t2, 2
+	break

parallel-rsp/debug-toolchain/cop0.s

@@ -0,0 +1,14 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 100
+	mtc0 $t0, $1
+	bgez $t0, bar
+	mfc0 $t1, $1
+	li $t1, 200 # Should not execute.
+	break
+
+bar:
+	li $t2, 400
+	break

parallel-rsp/debug-toolchain/cop2-basic.s

@@ -0,0 +1,16 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 10
+	li $t2, 10
+	nop
+	.word (0x12 << 26) | (4 << 21) | (2 << 11) | (8 << 16) | (4 << 7)
+	.word (0x12 << 26) | (0 << 21) | (2 << 11) | (9 << 16) | (4 << 7)
+	#mtc2 $t0, $2, 4
+	#mfc2 $t1, $2, 4
+
+	ctc2 $t2, $2
+	cfc2 $t3, $2
+
+	break

parallel-rsp/debug-toolchain/cop2-ls.s

@@ -0,0 +1,15 @@
+.data
+buffer:
+.hword 10, 20, 30, 40, 50, 60, 70, 80
+.hword 20, 30, 40, 50, 60, 70, 80, 90
+
+.text
+.set noreorder
+.global main
+main:
+	la $t0, buffer
+	# lqv 2, 0, 1(t0)
+	.word (0x32 << 26) | (4 << 11) | (2 << 16) | (0 << 7) | (1 << 0) | (8 << 21)
+	# sqv 2, 0, 2(t0)
+	.word (0x3a << 26) | (4 << 11) | (2 << 16) | (0 << 7) | (2 << 0) | (8 << 21)
+	break

parallel-rsp/debug-toolchain/cop2-vector.s

@@ -0,0 +1,18 @@
+.data
+buffer:
+.hword 10, 20, 30, 40, 50, 60, 70, 80
+.hword 20, 30, 40, 50, 60, 70, 80, 90
+
+.text
+.set noreorder
+.global main
+main:
+	la $t0, buffer
+	# lqv 2, 0, 0(t0)
+	.word (0x32 << 26) | (4 << 11) | (2 << 16) | (0 << 7) | (0 << 0) | (8 << 21)
+	# lqv 3, 0, 1(t0)
+	.word (0x32 << 26) | (4 << 11) | (3 << 16) | (0 << 7) | (1 << 0) | (8 << 21)
+
+	.word (0x25 << 25) | (4 << 6) | (2 << 11) | (3 << 16) | (17 << 0) | (0 << 21)
+	# vsub v4, v2, v3, 0
+	break

parallel-rsp/debug-toolchain/delay-slot-before-break.s

@@ -0,0 +1,15 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	blez $t0, foo
+	break
+
+	nop
+	nop
+	nop
+	nop
+
+foo:
+	break

parallel-rsp/debug-toolchain/delay-slot-before-new-block-illegal.s

@@ -0,0 +1,21 @@
+.text
+.set noreorder
+.global main
+.align 10
+main:
+	li $t0, 1
+	.rept 126
+		nop
+	.endr
+	bgtz $t0, foobar
+	# New block begins here,
+	j foobar2
+	li $t2, 2 # This should not execute
+
+foobar:
+	li $t3, 3 # This should execute, then we move to foobar2
+	break
+
+foobar2:
+	li $t4, 4 # This executes.
+	break

parallel-rsp/debug-toolchain/delay-slot-before-new-block-not-taken.s

@@ -0,0 +1,19 @@
+.text
+.set noreorder
+.global main
+.align 10
+main:
+	li $t0, 0
+	.rept 126
+		nop
+	.endr
+	bgtz $t0, foobar
+	# New block begins here,
+	# last instruction will be bgtz which needs to promote its delay slot
+	# into a new block.
+	# First instruction of any block have to handle potential latent delay slots like these.
+	li $t1, 1
+	li $t2, 2 # This should execute
+
+foobar:
+	break

parallel-rsp/debug-toolchain/delay-slot-before-new-block.s

@@ -0,0 +1,19 @@
+.text
+.set noreorder
+.global main
+.align 10
+main:
+	li $t0, 1
+	.rept 126
+		nop
+	.endr
+	bgtz $t0, foobar
+	# New block begins here,
+	# last instruction will be bgtz which needs to promote its delay slot
+	# into a new block.
+	# First instruction of any block have to handle potential latent delay slots like these.
+	li $t1, 1
+	li $t2, 2 # This should not execute
+
+foobar:
+	break

parallel-rsp/debug-toolchain/j.s

@@ -0,0 +1,20 @@
+.text
+.set noreorder
+.global main
+main:
+	j indelay
+	nop
+
+	j die # Test to make sure this becomes a conditional branch.
+
+indelay:
+	j notdie
+	li $t0, 1
+
+die:
+	li $t0, 0x10000
+
+notdie:
+	li $t1, 2
+	break
+

parallel-rsp/debug-toolchain/jal-into-indirect-delay-slot.s

@@ -0,0 +1,19 @@
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	bgtz $0, dummy
+	jal callsite
+	li $t5, 10
+
+	jr $s5
+callsite:
+	lw $t2, 184($0)
+	li $t1, 10
+	break
+
+dummy:
+	nop
+	break
+

parallel-rsp/debug-toolchain/jal.s

@@ -0,0 +1,21 @@
+.text
+.set noreorder
+.global main
+main:
+	jal indelay
+	nop
+
+	jal die # Test to make sure this becomes a conditional branch.
+
+indelay:
+	jal notdie
+	move $t2, $ra
+	li $t0, 1
+
+die:
+	li $t0, 0x10000
+
+notdie:
+	li $t1, 2
+	break
+

parallel-rsp/debug-toolchain/jalr.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	la $t5, notdie
+	la $ra, indelay
+	jalr $a0, $ra
+	move $ra, $a0
+
+	j die # Test to make sure this becomes a conditional branch.
+
+indelay:
+	jalr $a0, $t5
+	move $t0, $a0
+
+die:
+	li $t0, 0x10000
+
+notdie:
+	li $t1, 2
+	break
+

parallel-rsp/debug-toolchain/jr.s

@@ -0,0 +1,22 @@
+.text
+.set noreorder
+.global main
+main:
+	la $t5, notdie
+	la $ra, indelay
+	jr $ra
+	nop
+
+	j die # Test to make sure this becomes a conditional branch.
+
+indelay:
+	jr $t5
+	li $t0, 1
+
+die:
+	li $t0, 0x10000
+
+notdie:
+	li $t1, 2
+	break
+

parallel-rsp/debug-toolchain/lb.s

@@ -0,0 +1,11 @@
+.data
+.byte 0, 0, 0
+foobar:
+.byte 0x72, 0x91
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lb $t0, 1($t1)
+	break

parallel-rsp/debug-toolchain/lbu.s

@@ -0,0 +1,11 @@
+.data
+.byte 0, 0, 0
+foobar:
+.byte 0x72, 0x91
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lbu $t0, 1($t1)
+	break

parallel-rsp/debug-toolchain/lh-unaligned.s

@@ -0,0 +1,12 @@
+.data
+.byte 0, 0, 0, 0
+foobar:
+.byte 0
+.byte 0x92, 0x70
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lh $t0, 1($t1)
+	break

parallel-rsp/debug-toolchain/lh.s

@@ -0,0 +1,11 @@
+.data
+.byte 0, 0, 0, 0
+foobar:
+.byte 0x92, 0x70
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lh $t0, 0($t1)
+	break

parallel-rsp/debug-toolchain/lhu-unaligned.s

@@ -0,0 +1,12 @@
+.data
+.byte 0, 0, 0, 0
+foobar:
+.byte 0
+.byte 0x92, 0x70
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lhu $t0, 1($t1)
+	break

parallel-rsp/debug-toolchain/lhu.s

@@ -0,0 +1,11 @@
+.data
+.byte 0, 0, 0, 0
+foobar:
+.byte 0x92, 0x70
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lhu $t0, 0($t1)
+	break

parallel-rsp/debug-toolchain/lui.s

@@ -0,0 +1,5 @@
+.text
+.global main
+main:
+	lui $t0, 50000
+	break

parallel-rsp/debug-toolchain/lw-unaligned-in-branch-delay.s

@@ -0,0 +1,20 @@
+.data
+foo:
+.byte 0
+.byte 0xaa, 0xbb, 0xcc, 0xdd
+
+.text
+.set noreorder
+.global main
+main:
+	li $t0, 1
+	la $t1, foo
+	bgtz $t0, taken1
+	lw $t1, 1($t1)
+
+taken0:
+	li $t5, 100
+	break
+
+taken1:
+	break

parallel-rsp/debug-toolchain/lw-unaligned.s

@@ -0,0 +1,15 @@
+.data
+.byte 0, 0, 0, 0
+foobar:
+.word 12345678
+.word 30
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lw $t0, 0($t1)
+	lw $t1, 1($t1)
+	lw $t2, 2($t1)
+	lw $t3, 3($t1)
+	break

parallel-rsp/debug-toolchain/lw.s

@@ -0,0 +1,12 @@
+.data
+.byte 0, 0, 0, 0
+foobar:
+.word 12345678
+.word 30
+
+.text
+.global main
+main:
+	la $t1, foobar
+	lw $t0, 4($t1)
+	break

parallel-rsp/debug-toolchain/main.c

@@ -0,0 +1,21 @@
+#include "rsp-mips.h"
+
+u8 data[4] = { 0x10, 0x20, 0x30, 0x40 };
+
+__attribute__((noinline))
+int load_byte(int i)
+{
+	return data[i];
+}
+
+int count = 3;
+
+int main(void)
+{
+	int res = 0;
+	for (int i = 0; i < 3; i++)
+	{
+		res += load_byte(i);
+	}
+	rsp_debug_break(res, res, res, res);
+}

parallel-rsp/debug-toolchain/nor.s

@@ -0,0 +1,7 @@
+.text
+.global main
+main:
+	li $t0, 17
+	li $t1, 5
+	nor $t2, $t0, $t1
+	break

parallel-rsp/debug-toolchain/or.s

@@ -0,0 +1,7 @@
+.text
+.global main
+main:
+	li $t0, 8
+	li $t1, 50
+	or $t2, $t0, $t1
+	break

parallel-rsp/debug-toolchain/ori.s

@@ -0,0 +1,6 @@
+.text
+.global main
+main:
+	li $a0, 500
+	ori $t0, $a0, 50000
+	break

parallel-rsp/debug-toolchain/rsp-mips.h

@@ -0,0 +1,17 @@
+#ifndef __RSP_MIPS_H
+#define __RSP_MIPS_H
+
+typedef signed char s8;
+typedef unsigned char u8;
+typedef signed short s16;
+typedef unsigned short u16;
+typedef signed int s32;
+typedef unsigned int u32;
+typedef signed long long s64;
+typedef unsigned long long u64;
+typedef u32 size_t;
+
+void rsp_break(void);
+void rsp_debug_break(u32 a, u32 b, u32 c, u32 d);
+
+#endif

parallel-rsp/debug-toolchain/rsp-mips.ld

@@ -0,0 +1,9 @@
+SECTIONS {
+   ENTRY(rsp_mips_start)
+   . = 0x00400000;
+   .text : { start.o (.text); * (.text*); }
+
+   . = 0x00700000;
+   .data : { * (.data); * (.rodata*); * (.sbss); * (.scommon); }
+   .bss : { * (.bss); }
+}

parallel-rsp/debug-toolchain/rsp-mips.s

@@ -0,0 +1,20 @@
+.text
+.section .text
+
+.global rsp_break
+.ent rsp_break 
+.type rsp_break, @function
+rsp_break:
+	break
+	jr $ra
+.end rsp_break
+.size rsp_break, .-rsp_break
+
+.global rsp_debug_break
+.ent rsp_debug_break 
+.type rsp_debug_break, @function
+rsp_debug_break:
+	break
+	jr $ra
+.end rsp_debug_break
+.size rsp_debug_break, .-rsp_debug_break

parallel-rsp/debug-toolchain/sb.s

@@ -0,0 +1,21 @@
+.data
+foobar:
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+
+.text
+.global main
+main:
+	li $t0, 0xaabbccdd
+	la $t1, foobar
+	sb $t0, 0($t1)
+	sb $t0, 1($t1)
+	sb $t0, 3($t1)
+	break
+

parallel-rsp/debug-toolchain/sh-unaligned.s

@@ -0,0 +1,20 @@
+.data
+foobar:
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+
+.text
+.global main
+main:
+	li $t0, 0xaabbccdd
+	la $t1, foobar
+	sh $t0, 0($t1)
+	sh $t0, 3($t1)
+	break
+

parallel-rsp/debug-toolchain/sh.s

@@ -0,0 +1,21 @@
+.data
+foobar:
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+
+.text
+.global main
+main:
+	li $t0, 0xaabbccdd
+	la $t1, foobar
+	sh $t0, 0($t1)
+	sh $t0, 2($t1)
+	sh $t0, 6($t1)
+	break
+

parallel-rsp/debug-toolchain/sll.s

@@ -0,0 +1,6 @@
+.text
+.global main
+main:
+	li $t0, 40
+	sll $t0, $t0, 20
+	break

parallel-rsp/debug-toolchain/sllv.s

@@ -0,0 +1,7 @@
+.text
+.global main
+main:
+	li $t0, 40
+	li $t1, 20
+	sllv $t0, $t0, $t1
+	break

parallel-rsp/debug-toolchain/slt.s

@@ -0,0 +1,12 @@
+.text
+.global main
+main:
+	li $a0, 17
+	li $a1, 5
+	li $a2, -17
+	li $a3, -5
+	slt $t0, $a0, $a1
+	slt $t1, $a3, $a0
+	slt $t2, $a2, $a3
+	slt $t3, $a2, $a1
+	break

parallel-rsp/debug-toolchain/slti.s

@@ -0,0 +1,10 @@
+.text
+.global main
+main:
+	li $a0, 100000
+	li $a1, -100000
+	slti $t0, $a0, -40
+	slti $t1, $a0, 40
+	slti $t2, $a1, -40
+	slti $t3, $a1, 40
+	break

parallel-rsp/debug-toolchain/sltiu.s

@@ -0,0 +1,10 @@
+.text
+.global main
+main:
+	li $a0, 100000
+	li $a1, -100000
+	sltiu $t0, $a0, -40
+	sltiu $t1, $a0, 40
+	sltiu $t2, $a1, -40
+	sltiu $t3, $a1, 40
+	break

parallel-rsp/debug-toolchain/sltu.s

@@ -0,0 +1,12 @@
+.text
+.global main
+main:
+	li $a0, 17
+	li $a1, 5
+	li $a2, -17
+	li $a3, -5
+	sltu $t0, $a0, $a1
+	sltu $t1, $a3, $a0
+	sltu $t2, $a2, $a3
+	sltu $t3, $a2, $a1
+	break

parallel-rsp/debug-toolchain/sra.s

@@ -0,0 +1,6 @@
+.text
+.global main
+main:
+	li $t0, -15254
+	sra $t0, $t0, 20
+	break

parallel-rsp/debug-toolchain/srav.s

@@ -0,0 +1,7 @@
+.text
+.global main
+main:
+	li $t0, -15254
+	li $t1, 20
+	srav $t0, $t0, $t1
+	break

parallel-rsp/debug-toolchain/srl.s

@@ -0,0 +1,6 @@
+.text
+.global main
+main:
+	li $t0, -15254
+	srl $t0, $t0, 20
+	break

parallel-rsp/debug-toolchain/srlv.s

@@ -0,0 +1,7 @@
+.text
+.global main
+main:
+	li $t0, -15254
+	li $t1, 20
+	srlv $t0, $t0, $t1
+	break

parallel-rsp/debug-toolchain/start.s

@@ -0,0 +1,28 @@
+###
+##
+# Entry point and setup for our RSP
+##
+
+.equ RAM_SIZE, 4096
+
+.text
+.align 2
+.section .text
+.global rsp_mips_start
+.extern main
+.ent rsp_mips_start
+.type rsp_mips_start, @function
+
+rsp_mips_start:
+   nop
+redo:
+   li $a0, 0 # argc = 0, it will never be used anyways on this stuff. :D
+   li $a1, 0 # argv = 0
+   li $sp, (0x00700000 + RAM_SIZE - 4) # Set up stack.
+
+   jal main
+   j redo
+
+.end rsp_mips_start
+.size rsp_mips_start, .-rsp_mips_start
+

parallel-rsp/debug-toolchain/sub.s

@@ -0,0 +1,14 @@
+.text
+.global main
+main:
+	li $t0, 40
+	li $t1, -20
+	sub $t2, $t0, $t1
+	subu $t3, $t0, $t1
+
+	li $t0, -5
+	li $t1, -4
+	sub $t2, $t0, $t1
+	subu $t3, $t0, $t1
+
+	break

parallel-rsp/debug-toolchain/sw-unaligned.s

@@ -0,0 +1,22 @@
+.data
+foobar:
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+
+.text
+.global main
+main:
+	li $t0, 0xaabbccdd
+	la $t1, foobar
+	sw $t0, 0($t1)
+	sw $t0, 5($t1)
+	sw $t0, 10($t1)
+	sw $t0, 15($t1)
+	break
+

parallel-rsp/debug-toolchain/sw.s

@@ -0,0 +1,21 @@
+.data
+foobar:
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+.word 0
+
+.text
+.global main
+main:
+	li $t0, 0xaabbccdd
+	la $t1, foobar
+	sw $t0, 0($t1)
+	sw $t0, 4($t1)
+	sw $t0, 12($t1)
+	break
+

parallel-rsp/debug-toolchain/unconditional-delay-slot-before-break.s

@@ -0,0 +1,14 @@
+.text
+.set noreorder
+.global main
+main:
+	jal foo
+	break
+
+	nop
+	nop
+	nop
+	nop
+
+foo:
+	break

parallel-rsp/debug-toolchain/xor.s

@@ -0,0 +1,7 @@
+.text
+.global main
+main:
+	li $t0, 17
+	li $t1, 5
+	xor $t2, $t0, $t1
+	break

parallel-rsp/debug-toolchain/xori.s

@@ -0,0 +1,6 @@
+.text
+.global main
+main:
+	li $a0, 500
+	xori $t0, $a0, 50000
+	break