Reflect changes to ISA

Conflicts:

	isa/Makefile

16 files changed, 169 insertions, 820 deletions

diff --git a/benchmarks/Makefile b/benchmarks/Makefile
index 220a033..5d33f86 100644
--- a/benchmarks/Makefile
+++ b/benchmarks/Makefile
@@ -50,7 +50,7 @@ HOST_OPTS = -std=gnu99 -DPREALLOCATE=0 -DHOST_DEBUG=1
 HOST_COMP = gcc $(HOST_OPTS)
 
 RISCV_GCC = riscv-gcc
-RISCV_GCC_OPTS = -std=gnu99 -DSET_STATS -O2 -nostdlib -nostartfiles -ffast-math
+RISCV_GCC_OPTS = -Wa,-march=RVIMAFDXhwacha -std=gnu99 -O2 -nostdlib -nostartfiles -ffast-math
 RISCV_LINK = riscv-gcc -T $(bmarkdir)/common/test.ld
 RISCV_LINK_MT = riscv-gcc -T $(bmarkdir)/common/test-mt.ld
 RISCV_LINK_OPTS = -lc
diff --git a/benchmarks/common/crt-mt.S b/benchmarks/common/crt-mt.S
index 283b3bf..90cd755 100644
--- a/benchmarks/common/crt-mt.S
+++ b/benchmarks/common/crt-mt.S
@@ -1,3 +1,5 @@
+#include "pcr.h"
+
   .data
   .globl _heapend
   .globl environ
@@ -43,57 +45,53 @@ _start:
   li  x31,0
 
   # enable fp
-  mfpcr  x1,cr0
-  ori  x1,x1,0x2
-  mtpcr  x1,cr0
+  setpcr status, SR_EF
 
   # enable vec
-  mfpcr  x1,cr0
-  ori  x1,x1,0x4
-  mtpcr  x1,cr0
+  setpcr t0, status, SR_EV
 
   ## if that didn't stick, we don't have an FPU, so don't initialize it
-  mfpcr x1,cr0
-  andi x1,x1,0x2
-  beqz x1,1f
+  and t0, t0, SR_EF
+  beqz t0, 1f
 
-  mtfsr   x0
-  mxtf.s  f0, x0
-  mxtf.s  f1, x0
-  mxtf.s  f2, x0
-  mxtf.s  f3, x0
-  mxtf.s  f4, x0
-  mxtf.s  f5, x0
-  mxtf.s  f6, x0
-  mxtf.s  f7, x0
-  mxtf.s  f8, x0
-  mxtf.s  f9, x0
-  mxtf.s  f10,x0
-  mxtf.s  f11,x0
-  mxtf.s  f12,x0
-  mxtf.s  f13,x0
-  mxtf.s  f14,x0
-  mxtf.s  f15,x0
-  mxtf.s  f16,x0
-  mxtf.s  f17,x0
-  mxtf.s  f18,x0
-  mxtf.s  f19,x0
-  mxtf.s  f20,x0
-  mxtf.s  f21,x0
-  mxtf.s  f22,x0
-  mxtf.s  f23,x0
-  mxtf.s  f24,x0
-  mxtf.s  f25,x0
-  mxtf.s  f26,x0
-  mxtf.s  f27,x0
-  mxtf.s  f28,x0
-  mxtf.s  f29,x0
-  mxtf.s  f30,x0
-  mxtf.s  f31,x0
+  fssr    x0
+  fmv.s.x f0, x0
+  fmv.s.x f1, x0
+  fmv.s.x f2, x0
+  fmv.s.x f3, x0
+  fmv.s.x f4, x0
+  fmv.s.x f5, x0
+  fmv.s.x f6, x0
+  fmv.s.x f7, x0
+  fmv.s.x f8, x0
+  fmv.s.x f9, x0
+  fmv.s.x f10,x0
+  fmv.s.x f11,x0
+  fmv.s.x f12,x0
+  fmv.s.x f13,x0
+  fmv.s.x f14,x0
+  fmv.s.x f15,x0
+  fmv.s.x f16,x0
+  fmv.s.x f17,x0
+  fmv.s.x f18,x0
+  fmv.s.x f19,x0
+  fmv.s.x f20,x0
+  fmv.s.x f21,x0
+  fmv.s.x f22,x0
+  fmv.s.x f23,x0
+  fmv.s.x f24,x0
+  fmv.s.x f25,x0
+  fmv.s.x f26,x0
+  fmv.s.x f27,x0
+  fmv.s.x f28,x0
+  fmv.s.x f29,x0
+  fmv.s.x f30,x0
+  fmv.s.x f31,x0
 1:
 
+
   # get core id and number of cores
-  mfpcr a0,cr10
+  mfpcr a0,hartid
   lw a1, 4(zero)
 
   slli a2, a0, 13
diff --git a/benchmarks/common/crt.S b/benchmarks/common/crt.S
index d153210..ac5d9ba 100644
--- a/benchmarks/common/crt.S
+++ b/benchmarks/common/crt.S
@@ -1,3 +1,5 @@
+#include "pcr.h"
+
   .data
   .globl _heapend
   .globl environ
@@ -43,57 +45,52 @@ _start:
   li  x31,0
 
   # enable fp
-  mfpcr  x1,cr0
-  ori  x1,x1,0x2
-  mtpcr  x1,cr0
+  setpcr status, SR_EF
 
   # enable vec
-  mfpcr  x1,cr0
-  ori  x1,x1,0x4
-  mtpcr  x1,cr0
+  setpcr t0, status, SR_EV
 
   ## if that didn't stick, we don't have an FPU, so don't initialize it
-  mfpcr x1,cr0
-  andi x1,x1,0x2
-  beqz x1,1f
+  and t0, t0, SR_EF
+  beqz t0, 1f
 
-  mtfsr   x0
-  mxtf.s  f0, x0
-  mxtf.s  f1, x0
-  mxtf.s  f2, x0
-  mxtf.s  f3, x0
-  mxtf.s  f4, x0
-  mxtf.s  f5, x0
-  mxtf.s  f6, x0
-  mxtf.s  f7, x0
-  mxtf.s  f8, x0
-  mxtf.s  f9, x0
-  mxtf.s  f10,x0
-  mxtf.s  f11,x0
-  mxtf.s  f12,x0
-  mxtf.s  f13,x0
-  mxtf.s  f14,x0
-  mxtf.s  f15,x0
-  mxtf.s  f16,x0
-  mxtf.s  f17,x0
-  mxtf.s  f18,x0
-  mxtf.s  f19,x0
-  mxtf.s  f20,x0
-  mxtf.s  f21,x0
-  mxtf.s  f22,x0
-  mxtf.s  f23,x0
-  mxtf.s  f24,x0
-  mxtf.s  f25,x0
-  mxtf.s  f26,x0
-  mxtf.s  f27,x0
-  mxtf.s  f28,x0
-  mxtf.s  f29,x0
-  mxtf.s  f30,x0
-  mxtf.s  f31,x0
+  fssr    x0
+  fmv.s.x f0, x0
+  fmv.s.x f1, x0
+  fmv.s.x f2, x0
+  fmv.s.x f3, x0
+  fmv.s.x f4, x0
+  fmv.s.x f5, x0
+  fmv.s.x f6, x0
+  fmv.s.x f7, x0
+  fmv.s.x f8, x0
+  fmv.s.x f9, x0
+  fmv.s.x f10,x0
+  fmv.s.x f11,x0
+  fmv.s.x f12,x0
+  fmv.s.x f13,x0
+  fmv.s.x f14,x0
+  fmv.s.x f15,x0
+  fmv.s.x f16,x0
+  fmv.s.x f17,x0
+  fmv.s.x f18,x0
+  fmv.s.x f19,x0
+  fmv.s.x f20,x0
+  fmv.s.x f21,x0
+  fmv.s.x f22,x0
+  fmv.s.x f23,x0
+  fmv.s.x f24,x0
+  fmv.s.x f25,x0
+  fmv.s.x f26,x0
+  fmv.s.x f27,x0
+  fmv.s.x f28,x0
+  fmv.s.x f29,x0
+  fmv.s.x f30,x0
+  fmv.s.x f31,x0
 1:
 
   # only allow core 0 to proceed
-1:mfpcr a0, cr10
+1:mfpcr a0, hartid
   bnez a0, 1b
 
   la  sp,stacktop
diff --git a/benchmarks/common/pcr.h b/benchmarks/common/pcr.h
index 7659a97..8780cdd 100644
--- a/benchmarks/common/pcr.h
+++ b/benchmarks/common/pcr.h
@@ -1,42 +1,53 @@
+// See LICENSE for license details.
+
 #ifndef _RISCV_PCR_H
 #define _RISCV_PCR_H
 
-#define SR_ET    0x00000001
-#define SR_EF    0x00000002
-#define SR_EV    0x00000004
-#define SR_EC    0x00000008
-#define SR_PS    0x00000010
-#define SR_S     0x00000020
-#define SR_U64   0x00000040
-#define SR_S64   0x00000080
-#define SR_VM    0x00000100
+#define SR_S     0x00000001
+#define SR_PS    0x00000002
+#define SR_EI    0x00000004
+#define SR_PEI   0x00000008
+#define SR_EF    0x00000010
+#define SR_U64   0x00000020
+#define SR_S64   0x00000040
+#define SR_VM    0x00000080
+#define SR_EV    0x00000100
 #define SR_IM    0x00FF0000
-#define SR_ZERO  ~(SR_ET|SR_EF|SR_EV|SR_EC|SR_PS|SR_S|SR_U64|SR_S64|SR_VM|SR_IM)
+#define SR_IP    0xFF000000
+#define SR_ZERO  ~(SR_S|SR_PS|SR_EI|SR_PEI|SR_EF|SR_U64|SR_S64|SR_VM|SR_EV|SR_IM|SR_IP)
 #define SR_IM_SHIFT 16
+#define SR_IP_SHIFT 24
 
-#define PCR_SR       0
-#define PCR_EPC      1
-#define PCR_BADVADDR 2
-#define PCR_EVEC     3
-#define PCR_COUNT    4
-#define PCR_COMPARE  5
-#define PCR_CAUSE    6
-#define PCR_PTBR     7
-#define PCR_SEND_IPI 8
-#define PCR_CLR_IPI  9
-#define PCR_COREID   10
-#define PCR_IMPL     11
-#define PCR_K0       12
-#define PCR_K1       13
+#define PCR_SUP0     0
+#define PCR_SUP1     1
+#define PCR_EPC      2
+#define PCR_BADVADDR 3
+#define PCR_PTBR     4
+#define PCR_ASID     5
+#define PCR_COUNT    6
+#define PCR_COMPARE  7
+#define PCR_EVEC     8
+#define PCR_CAUSE    9
+#define PCR_SR       10
+#define PCR_HARTID   11
+#define PCR_IMPL     12
+#define PCR_FATC     13
+#define PCR_SEND_IPI 14
+#define PCR_CLR_IPI  15
 #define PCR_VECBANK  18
 #define PCR_VECCFG   19
 #define PCR_RESET    29
 #define PCR_TOHOST   30
 #define PCR_FROMHOST 31
 
+#define IRQ_COP   2
 #define IRQ_IPI   5
+#define IRQ_HOST  6
 #define IRQ_TIMER 7
 
+#define IMPL_SPIKE  1
+#define IMPL_ROCKET 2
+
 #define CAUSE_MISALIGNED_FETCH 0
 #define CAUSE_FAULT_FETCH 1
 #define CAUSE_ILLEGAL_INSTRUCTION 2
@@ -60,10 +71,29 @@
 #define CAUSE_VECTOR_FAULT_LOAD 30
 #define CAUSE_VECTOR_FAULT_STORE 31
 
+// page table entry (PTE) fields
+#define PTE_V    0x001 // Entry is a page Table descriptor
+#define PTE_T    0x002 // Entry is a page Table, not a terminal node
+#define PTE_G    0x004 // Global
+#define PTE_UR   0x008 // User Write permission
+#define PTE_UW   0x010 // User Read permission
+#define PTE_UX   0x020 // User eXecute permission
+#define PTE_SR   0x040 // Supervisor Read permission
+#define PTE_SW   0x080 // Supervisor Write permission
+#define PTE_SX   0x100 // Supervisor eXecute permission
+#define PTE_PERM (PTE_SR | PTE_SW | PTE_SX | PTE_UR | PTE_UW | PTE_UX)
+
 #ifdef __riscv
 
-#define ASM_CR(r)   _ASM_CR(r)
-#define _ASM_CR(r)  cr##r
+#ifdef __riscv64
+# define RISCV_PGLEVELS 3
+# define RISCV_PGSHIFT 13
+#else
+# define RISCV_PGLEVELS 2
+# define RISCV_PGSHIFT 12
+#endif
+#define RISCV_PGLEVEL_BITS 10
+#define RISCV_PGSIZE (1 << RISCV_PGSHIFT)
 
 #ifndef __ASSEMBLER__
 
@@ -83,6 +113,10 @@
           asm volatile ("clearpcr %0,cr%2,%1" : "=r"(__tmp) : "i"(val), "i"(reg)); \
           __tmp; })
 
+#define rdcycle() ({ unsigned long __tmp; \
+          asm volatile ("rdcycle %0" : "=r"(__tmp)); \
+          __tmp; })
+
 #endif
 
 #endif
diff --git a/benchmarks/common/syscalls.S b/benchmarks/common/syscalls.S
deleted file mode 100644
index a0cdf6e..0000000
--- a/benchmarks/common/syscalls.S
+++ /dev/null
@@ -1,678 +0,0 @@
-	.file	1 "syscalls.c"
-	.section .mdebug.abi64
-	.previous
-	.section	.rodata.str1.8,"aMS",@progbits,1
-	.align	3
-$LC0:
-	.ascii	"0123456789abcdef\000"
-	.text
-	.align	2
-	.ent	printnum
-	.type	printnum, @function
-printnum:
-	.frame	x30,64,x1		# vars= 0, regs= 7/0, args= 0
-	.mask	0x03f00002,-8
-	.fmask	0x00000000,0
-	add	x30,x30,-64
-	sd	x22,32(x30)
-	sll	x22,x7,32
-	srl	x22,x22,32
-	sd	x24,48(x30)
-	sd	x23,40(x30)
-	sd	x21,24(x30)
-	sd	x20,16(x30)
-	sd	x25,56(x30)
-	sd	x1,8(x30)
-	move	x23,x6
-	move	x20,x4
-	move	x21,x5
-	move	x24,x9
-	bleu	x22,x6,$L2
-	addw	x8,x8,-1
-	move	x25,x8
-	ble	x8,x0,$L4
-$L6:
-	addw	x25,x25,-1
-	move	x4,x24
-	move	x5,x21
-	jalr	x20
-	bne	x25,x0,$L6
-$L4:
-	lui	x2,%hi($LC0)
-	add	x2,x2,%lo($LC0)
-	remu	x22,x23,x22
-	add	x22,x22,x2
-	lb	x4,0(x22)
-	move	x5,x21
-	move	x19,x20
-	ld	x25,56(x30)
-	ld	x24,48(x30)
-	ld	x23,40(x30)
-	ld	x22,32(x30)
-	ld	x21,24(x30)
-	ld	x20,16(x30)
-	ld	x1,8(x30)
-	add	x30,x30,64
-	jr	x19
-$L2:
-	addw	x8,x8,-1
-	divu	x6,x6,x22
-	jal	printnum
-	j	$L4
-	.end	printnum
-	.size	printnum, .-printnum
-	.align	2
-	.ent	getuint
-	.type	getuint, @function
-getuint:
-	.frame	x30,0,x1		# vars= 0, regs= 0/0, args= 0
-	.mask	0x00000000,0
-	.fmask	0x00000000,0
-	slt	x2,x5,2
-	bne	x2,x0,$L10
-$L13:
-	ld	x2,0(x4)
-	add	x3,x2,8
-	sd	x3,0(x4)
-	ld	x2,0(x2)
-	ret
-$L10:
-	bne	x5,x0,$L13
-	ld	x3,0(x4)
-	lwu	x2,0(x3)
-	add	x3,x3,8
-	sd	x3,0(x4)
-	ret
-	.end	getuint
-	.size	getuint, .-getuint
-	.align	2
-	.globl	putchar
-	.ent	putchar
-	.type	putchar, @function
-putchar:
-	.frame	x30,64,x1		# vars= 64, regs= 0/0, args= 0
-	.mask	0x00000000,0
-	.fmask	0x00000000,0
-	li	x2,-1			# 0xffffffffffffffff
-	add	x30,x30,-64
-	lui	x3,%hi(buflen.1596)
-	beq	x4,x2,$L21
-	lw	x5,%lo(buflen.1596)(x3)
-	lui	x2,%hi(buf.1595)
-	add	x2,x2,%lo(buf.1595)
-	add	x6,x2,x5
-	sb	x4,0(x6)
-	addw	x5,x5,1
-	li	x4,64			# 0x40
-	sw	x5,%lo(buflen.1596)(x3)
-	beq	x5,x4,$L15
-	move	x2,x0
-	add	x30,x30,64
-	j	x1
-$L21:
-	lui	x2,%hi(buf.1595)
-	add	x2,x2,%lo(buf.1595)
-$L15:
-	lw	x4,%lo(buflen.1596)(x3)
-	li	x5,4			# 0x4
-	sd	x0,0(x30)
-	sd	x0,8(x30)
-	sd	x0,16(x30)
-	sd	x0,24(x30)
-	sd	x0,32(x30)
-	sd	x0,40(x30)
-	sd	x0,48(x30)
-	sd	x0,56(x30)
-	sd	x5,0(x30)
-	li	x5,1			# 0x1
-	sd	x5,8(x30)
-	sd	x2,16(x30)
-	sd	x4,24(x30)
-	fence
- #APP
- # 45 "syscalls.c" 1
-	mtpcr x2,x30,cr30
- # 0 "" 2
- #NO_APP
-$L17:
- #APP
- # 46 "syscalls.c" 1
-	mfpcr x2,cr31
- # 0 "" 2
- #NO_APP
-	beq	x2,x0,$L17
-	move	x2,x0
-	sw	x0,%lo(buflen.1596)(x3)
-	add	x30,x30,64
-	j	x1
-	.end	putchar
-	.size	putchar, .-putchar
-	.align	2
-	.globl	exit
-	.ent	exit
-	.type	exit, @function
-exit:
-	.frame	x30,64,x1		# vars= 64, regs= 0/0, args= 0
-	.mask	0x00000000,0
-	.fmask	0x00000000,0
-	add	x30,x30,-64
-	li	x2,1			# 0x1
-	sd	x0,0(x30)
-	sd	x0,8(x30)
-	sd	x0,16(x30)
-	sd	x0,24(x30)
-	sd	x0,32(x30)
-	sd	x0,40(x30)
-	sd	x0,48(x30)
-	sd	x0,56(x30)
-	sd	x2,0(x30)
-	sd	x4,8(x30)
-	fence
- #APP
- # 12 "syscalls.c" 1
-	mtpcr x2,x30,cr30
- # 0 "" 2
- #NO_APP
-$L23:
-	j	$L23
-	.end	exit
-	.size	exit, .-exit
-	.align	2
-	.globl	printstr
-	.ent	printstr
-	.type	printstr, @function
-printstr:
-	.frame	x30,80,x1		# vars= 64, regs= 1/0, args= 0
-	.mask	0x00000002,-8
-	.fmask	0x00000000,0
-	add	x30,x30,-80
-	li	x3,4			# 0x4
-	sd	x0,0(x30)
-	sd	x0,8(x30)
-	sd	x0,16(x30)
-	sd	x0,24(x30)
-	sd	x0,32(x30)
-	sd	x0,40(x30)
-	sd	x0,48(x30)
-	sd	x0,56(x30)
-	sd	x3,0(x30)
-	li	x3,1			# 0x1
-	sd	x3,8(x30)
-	sd	x1,72(x30)
-	sd	x4,16(x30)
-	jal	strlen
-	sd	x2,24(x30)
-	fence
- #APP
- # 24 "syscalls.c" 1
-	mtpcr x2,x30,cr30
- # 0 "" 2
- #NO_APP
-$L25:
- #APP
- # 25 "syscalls.c" 1
-	mfpcr x2,cr31
- # 0 "" 2
- #NO_APP
-	beq	x2,x0,$L25
-	ld	x1,72(x30)
-	add	x30,x30,80
-	j	x1
-	.end	printstr
-	.size	printstr, .-printstr
-	.align	2
-	.globl	printhex
-	.ent	printhex
-	.type	printhex, @function
-printhex:
-	.frame	x30,48,x1		# vars= 32, regs= 1/0, args= 0
-	.mask	0x00000002,-8
-	.fmask	0x00000000,0
-	add	x30,x30,-48
-	sd	x1,40(x30)
-	add	x2,x30,15
-	add	x7,x30,-1
-$L29:
-	and	x3,x4,15
-	sltu	x6,x3,10
-	li	x5,87			# 0x57
-	beq	x6,x0,$L28
-	li	x5,48			# 0x30
-$L28:
-	add	x3,x5,x3
-	sb	x3,0(x2)
-	add	x2,x2,-1
-	srl	x4,x4,4
-	bne	x2,x7,$L29
-	move	x4,x30
-	sb	x0,16(x30)
-	jal	printstr
-	ld	x1,40(x30)
-	add	x30,x30,48
-	j	x1
-	.end	printhex
-	.size	printhex, .-printhex
-	.section	.rodata.str1.8
-	.align	3
-$LC1:
-	.ascii	"(null)\000"
-	.text
-	.align	2
-	.globl	vprintfmt
-	.ent	vprintfmt
-	.type	vprintfmt, @function
-vprintfmt:
-	.frame	x30,112,x1		# vars= 32, regs= 10/0, args= 0
-	.mask	0x2ff00002,-8
-	.fmask	0x00000000,0
-	add	x30,x30,-112
-	sd	x25,80(x30)
-	lui	x25,%hi($L53)
-	sd	x26,88(x30)
-	sd	x24,72(x30)
-	sd	x23,64(x30)
-	sd	x22,56(x30)
-	sd	x21,48(x30)
-	sd	x20,40(x30)
-	sd	x29,104(x30)
-	sd	x27,96(x30)
-	sd	x1,32(x30)
-	move	x21,x4
-	move	x20,x5
-	move	x23,x6
-	sd	x7,0(x30)
-	li	x22,37			# 0x25
-	add	x25,x25,%lo($L53)
-	li	x24,-1			# 0xffffffffffffffff
-	lui	x26,%hi($LC1)
-	j	$L84
-$L35:
-	beq	x2,x0,$L32
-	move	x5,x20
-	add	x23,x23,1
-	jalr	x21
-$L84:
-	lbu	x4,0(x23)
-	move	x2,x4
-	bne	x4,x22,$L35
-	ld	x2,0(x30)
-	add	x6,x23,1
-	move	x7,x6
-	li	x27,32			# 0x20
-	sd	x0,8(x30)
-	li	x3,-1			# 0xffffffffffffffff
-	li	x29,-1			# 0xffffffffffffffff
-	move	x5,x0
-$L85:
-	lbu	x4,0(x7)
-	add	x23,x7,1
-	addw	x8,x4,-35
-	and	x9,x8,0xff
-	sltu	x9,x9,86
-	bne	x9,x0,$L90
-$L38:
-	li	x4,37			# 0x25
-	move	x5,x20
-	sd	x2,0(x30)
-	move	x23,x6
-	jalr	x21
-	j	$L84
-$L32:
-	ld	x1,32(x30)
-	ld	x29,104(x30)
-	ld	x27,96(x30)
-	ld	x26,88(x30)
-	ld	x25,80(x30)
-	ld	x24,72(x30)
-	ld	x23,64(x30)
-	ld	x22,56(x30)
-	ld	x21,48(x30)
-	ld	x20,40(x30)
-	add	x30,x30,112
-	j	x1
-$L90:
-	and	x8,x8,0xff
-	sll	x8,x8,3
-	add	x8,x25,x8
-	ld	x8,0(x8)
-	j	x8
-	.section	.rodata
-	.align	3
-	.align	2
-$L53:
-	.dword	$L39
-	.dword	$L38
-	.dword	$L40
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L41
-	.dword	$L38
-	.dword	$L38
-	.dword	$L42
-	.dword	$L43
-	.dword	$L38
-	.dword	$L74
-	.dword	$L44
-	.dword	$L44
-	.dword	$L44
-	.dword	$L44
-	.dword	$L44
-	.dword	$L44
-	.dword	$L44
-	.dword	$L44
-	.dword	$L44
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L45
-	.dword	$L46
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L38
-	.dword	$L47
-	.dword	$L38
-	.dword	$L38
-	.dword	$L48
-	.dword	$L49
-	.dword	$L38
-	.dword	$L38
-	.dword	$L50
-	.dword	$L38
-	.dword	$L51
-	.dword	$L38
-	.dword	$L38
-	.dword	$L52
-	.text
-$L52:
-	move	x4,x30
-	sd	x2,0(x30)
-	jal	getuint
-	move	x6,x2
-	li	x7,16			# 0x10
-$L73:
-	move	x4,x21
-	move	x5,x20
-	move	x8,x29
-	move	x9,x27
-	jal	printnum
-	j	$L84
-$L39:
-	li	x4,1			# 0x1
-	move	x7,x23
-	sd	x4,8(x30)
-	j	$L85
-$L40:
-	move	x5,x20
-	sd	x2,0(x30)
-	jalr	x21
-	j	$L84
-$L41:
-	lw	x3,0(x2)
-	move	x7,x23
-	add	x2,x2,8
-$L54:
-	bge	x29,x0,$L85
-	move	x29,x3
-	li	x3,-1			# 0xffffffffffffffff
-	j	$L85
-$L42:
-	move	x7,x23
-	li	x27,45			# 0x2d
-	j	$L85
-$L43:
-	blt	x29,x0,$L91
-	move	x7,x23
-	j	$L85
-$L74:
-	move	x7,x23
-	li	x27,48			# 0x30
-	j	$L85
-$L44:
-	lb	x8,1(x7)
-	addw	x3,x4,-48
-	move	x7,x23
-	addw	x4,x8,-48
-	sltu	x4,x4,10
-	beq	x4,x0,$L54
-$L55:
-	add	x7,x7,1
-	lb	x4,0(x7)
-	sllw	x9,x3,1
-	sllw	x3,x3,3
-	addw	x3,x9,x3
-	addw	x9,x4,-48
-	addw	x3,x3,x8
-	sltu	x9,x9,10
-	addw	x3,x3,-48
-	move	x8,x4
-	bne	x9,x0,$L55
-	j	$L54
-$L45:
-	lw	x4,0(x2)
-	add	x2,x2,8
-	move	x5,x20
-	sd	x2,0(x30)
-	jalr	x21
-	j	$L84
-$L46:
-	slt	x3,x5,2
-	bne	x3,x0,$L69
-$L88:
-	add	x3,x2,8
-	sd	x3,0(x30)
-	ld	x6,0(x2)
-	blt	x6,x0,$L72
-$L89:
-	li	x7,10			# 0xa
-	j	$L73
-$L47:
-	addw	x5,x5,1
-	move	x7,x23
-	j	$L85
-$L48:
-	move	x4,x30
-	sd	x2,0(x30)
-	jal	getuint
-	move	x6,x2
-	li	x7,8			# 0x8
-	j	$L73
-$L49:
-	sd	x2,0(x30)
-	li	x4,48			# 0x30
-	move	x5,x20
-	jalr	x21
-	li	x4,120			# 0x78
-	move	x5,x20
-	jalr	x21
-	ld	x2,0(x30)
-	li	x7,16			# 0x10
-	add	x3,x2,8
-	sd	x3,0(x30)
-	ld	x6,0(x2)
-	j	$L73
-$L50:
-	add	x4,x2,8
-	sd	x4,0(x30)
-	ld	x2,0(x2)
-	sd	x2,16(x30)
-	beq	x2,x0,$L92
-$L58:
-	ble	x29,x0,$L59
-	li	x2,45			# 0x2d
-	beq	x27,x2,$L59
-	ld	x4,16(x30)
-	move	x5,x3
-	sd	x3,24(x30)
-	jal	strnlen
-	sllw	x2,x2,0
-	subw	x2,x29,x2
-	move	x29,x2
-	ld	x3,24(x30)
-	ble	x2,x0,$L59
-	move	x29,x2
-$L60:
-	sd	x3,24(x30)
-	move	x4,x27
-	move	x5,x20
-	addw	x29,x29,-1
-	jalr	x21
-	ld	x3,24(x30)
-	bne	x29,x0,$L60
-$L59:
-	ld	x2,16(x30)
-	lb	x4,0(x2)
-	add	x27,x2,1
-	beq	x4,x0,$L62
-$L80:
-	blt	x3,x0,$L67
-	addw	x2,x3,-1
-	move	x3,x2
-	beq	x2,x24,$L62
-$L67:
-	ld	x2,8(x30)
-	beq	x2,x0,$L63
-	addw	x2,x4,-32
-	sltu	x2,x2,95
-	beq	x2,x0,$L93
-$L63:
-	sd	x3,24(x30)
-	move	x5,x20
-	jalr	x21
-	ld	x3,24(x30)
-$L64:
-	lb	x4,0(x27)
-	addw	x29,x29,-1
-	add	x27,x27,1
-	bne	x4,x0,$L80
-$L62:
-	ble	x29,x0,$L84
-$L79:
-	addw	x29,x29,-1
-	li	x4,32			# 0x20
-	move	x5,x20
-	jalr	x21
-	bne	x29,x0,$L79
-	j	$L84
-$L51:
-	move	x4,x30
-	sd	x2,0(x30)
-	jal	getuint
-	move	x6,x2
-	li	x7,10			# 0xa
-	j	$L73
-$L93:
-	sd	x3,24(x30)
-	li	x4,63			# 0x3f
-	move	x5,x20
-	jalr	x21
-	ld	x3,24(x30)
-	j	$L64
-$L91:
-	move	x7,x23
-	move	x29,x0
-	j	$L85
-$L92:
-	add	x2,x26,%lo($LC1)
-	sd	x2,16(x30)
-	j	$L58
-$L72:
-	sd	x6,24(x30)
-	li	x4,45			# 0x2d
-	move	x5,x20
-	jalr	x21
-	ld	x6,24(x30)
-	li	x7,10			# 0xa
-	sub	x6,zero,x6
-	j	$L73
-$L69:
-	bne	x5,x0,$L88
-	lw	x6,0(x2)
-	add	x2,x2,8
-	sd	x2,0(x30)
-	bge	x6,x0,$L89
-	j	$L72
-	.end	vprintfmt
-	.size	vprintfmt, .-vprintfmt
-	.align	2
-	.globl	printf
-	.ent	printf
-	.type	printf, @function
-printf:
-	.frame	x30,96,x1		# vars= 16, regs= 1/0, args= 0
-	.mask	0x00000002,-72
-	.fmask	0x00000000,0
-	add	x30,x30,-96
-	add	x2,x30,40
-	move	x3,x4
-	lui	x4,%hi(putchar)
-	sd	x5,40(x30)
-	sd	x6,48(x30)
-	sd	x7,56(x30)
-	add	x4,x4,%lo(putchar)
-	move	x5,x0
-	move	x6,x3
-	move	x7,x2
-	sd	x1,24(x30)
-	sd	x8,64(x30)
-	sd	x9,72(x30)
-	sd	x10,80(x30)
-	sd	x11,88(x30)
-	sd	x2,0(x30)
-	jal	vprintfmt
-	li	x4,-1			# 0xffffffffffffffff
-	jal	putchar
-	ld	x1,24(x30)
-	move	x2,x0
-	add	x30,x30,96
-	j	x1
-	.end	printf
-	.size	printf, .-printf
-	.local	buflen.1596
-	.comm	buflen.1596,4,4
-	.local	buf.1595
-	.comm	buf.1595,64,8
-	.ident	"GCC: (GNU) 4.6.1"
diff --git a/benchmarks/common/syscalls.c b/benchmarks/common/syscalls.c
index f95dde4..0c1bc7f 100644
--- a/benchmarks/common/syscalls.c
+++ b/benchmarks/common/syscalls.c
@@ -1,12 +1,13 @@
 #include <stdint.h>
 #include <string.h>
 #include <stdarg.h>
+#include <machine/syscall.h>
 #include "pcr.h"
 
 void exit(int code)
 {
   volatile uint64_t magic_mem[8] = {0};
-  magic_mem[0] = 1;
+  magic_mem[0] = SYS_exit;
   magic_mem[1] = code;
   __sync_synchronize();
   mtpcr(PCR_TOHOST, (long)magic_mem);
@@ -16,7 +17,7 @@ void exit(int code)
 void printstr(const char* s)
 {
   volatile uint64_t magic_mem[8] = {0};
-  magic_mem[0] = 4;
+  magic_mem[0] = SYS_write;
   magic_mem[1] = 1;
   magic_mem[2] = (unsigned long)s;
   magic_mem[3] = strlen(s);
@@ -27,17 +28,16 @@ void printstr(const char* s)
 
 int putchar(int ch)
 {
-  #define buffered_putch_bufsize 64
-  static char buf[buffered_putch_bufsize];
+  static char buf[64];
   static int buflen = 0;
 
   if(ch != -1)
     buf[buflen++] = ch;
 
-  if(ch == -1 || buflen == buffered_putch_bufsize)
+  if(ch == -1 || buflen == sizeof(buf))
   {
     volatile uint64_t magic_mem[8] = {0};
-    magic_mem[0] = 4;
+    magic_mem[0] = SYS_write;
     magic_mem[1] = 1;
     magic_mem[2] = (long)buf;
     magic_mem[3] = buflen;
diff --git a/benchmarks/dgemm/dgemm_main.c b/benchmarks/dgemm/dgemm_main.c
index 7c8ce7c..72c90ef 100644
--- a/benchmarks/dgemm/dgemm_main.c
+++ b/benchmarks/dgemm/dgemm_main.c
@@ -70,7 +70,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }
diff --git a/benchmarks/dhrystone/dhrystone_main.c b/benchmarks/dhrystone/dhrystone_main.c
index 7e8c6af..c2a8038 100644
--- a/benchmarks/dhrystone/dhrystone_main.c
+++ b/benchmarks/dhrystone/dhrystone_main.c
@@ -53,7 +53,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }
diff --git a/benchmarks/median/median_main.c b/benchmarks/median/median_main.c
index 6decd3d..5c435dd 100644
--- a/benchmarks/median/median_main.c
+++ b/benchmarks/median/median_main.c
@@ -76,7 +76,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }
diff --git a/benchmarks/multiply/multiply_main.c b/benchmarks/multiply/multiply_main.c
index fcb6b27..f2ce12b 100644
--- a/benchmarks/multiply/multiply_main.c
+++ b/benchmarks/multiply/multiply_main.c
@@ -85,7 +85,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }
diff --git a/benchmarks/qsort/qsort_main.c b/benchmarks/qsort/qsort_main.c
index 6de1613..486b8fc 100644
--- a/benchmarks/qsort/qsort_main.c
+++ b/benchmarks/qsort/qsort_main.c
@@ -92,7 +92,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }
diff --git a/benchmarks/spmv/spmv_main.c b/benchmarks/spmv/spmv_main.c
index 11d9540..afb1cf1 100644
--- a/benchmarks/spmv/spmv_main.c
+++ b/benchmarks/spmv/spmv_main.c
@@ -70,7 +70,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }
diff --git a/benchmarks/towers/towers_main.c b/benchmarks/towers/towers_main.c
index 36526a2..aa61665 100644
--- a/benchmarks/towers/towers_main.c
+++ b/benchmarks/towers/towers_main.c
@@ -59,7 +59,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }
diff --git a/benchmarks/vec-cmplxmult/vec_cmplxmult_asm.S b/benchmarks/vec-cmplxmult/vec_cmplxmult_asm.S
index 0771965..9ccd6c2 100644
--- a/benchmarks/vec-cmplxmult/vec_cmplxmult_asm.S
+++ b/benchmarks/vec-cmplxmult/vec_cmplxmult_asm.S
@@ -21,9 +21,9 @@
 #define rVlen   a6
 #define rStride a7
 
-#define rAI a8
-#define rBI a9
-#define rCI a10
+#define rAI t0
+#define rBI t1
+#define rCI t2
 
 # WARNING: do not write to the s0,...,s9 registers without first saving them to
 # the stack.  
diff --git a/benchmarks/vec-matmul/vec_matmul_asm.S b/benchmarks/vec-matmul/vec_matmul_asm.S
index f14d186..5135772 100644
--- a/benchmarks/vec-matmul/vec_matmul_asm.S
+++ b/benchmarks/vec-matmul/vec_matmul_asm.S
@@ -7,6 +7,8 @@
 # Headers and Defines
 #--------------------------------------------------------------------------
 
+#include "pcr.h"
+
 # Here are some defines that make writing assembly code easier.
 
 # I'm using the knowledge that rLda will be placed in register a0, rA will be
@@ -26,15 +28,15 @@
 
 # address of VT function
 #define rVTAddr v1
-#define rTemp0  a8
+#define rTemp0  t0
 
 # desired app vector length (number of elements to vectorize)
-#define rNum    a9
+#define rNum    t1
 
-#define rATemp  a10
-#define rBTemp  a11
-#define rCTemp  a12
-#define rI      a13
+#define rATemp  t2
+#define rBTemp  t3
+#define rCTemp  t4
+#define rI      t5
 #define rJ      s1
 #define rK      s2
 #define rLda4   a4
@@ -118,9 +120,7 @@ vt_matmul_asm:
 
         
         # turn on vector unit
-        mfpcr a13,cr0
-        ori a13,a13,4
-        mtpcr x0,a13,cr0
+        setpcr status, SR_EV
 
         blez rLda, cpdone    # exit early if lda < 0
 
@@ -289,9 +289,7 @@ vtcode:
 
 transpose:
         # turn on vector unit
-        mfpcr a13,cr0
-        ori a13,a13,4
-        mtpcr x0,a13,cr0
+        setpcr status, SR_EV
 
         
         blez rLda, cpdone    # exit early if lda < 0
diff --git a/benchmarks/vvadd/vvadd_main.c b/benchmarks/vvadd/vvadd_main.c
index 9738118..70b3f75 100644
--- a/benchmarks/vvadd/vvadd_main.c
+++ b/benchmarks/vvadd/vvadd_main.c
@@ -77,7 +77,7 @@ void finishTest( int toHostValue )
     printf( "*** FAILED *** (tohost = %d)\n", toHostValue );
   exit(0);
 #else
-  asm( "mtpcr %0, cr30" : : "r" (toHostValue) );
+  asm( "mtpcr %0, tohost" : : "r" (toHostValue) );
   while ( 1 ) { }
 #endif
 }