5d630cef4f
Don't invalidate TB with the end of zero overhead loop when LBEG or LEND change. Instead encode the distance from the start of the page where the TB starts to the LEND in the TB cs_base and generate loopback code when the next PC matches encoded LEND. Distance to a destination within the same page and up to a maximum instruction length into the next page is encoded literally, otherwise it's zero. The distance from LEND to LBEG is also encoded in the cs_base: it's encoded literally when less than 256 or as 0 otherwise. This allows for TB chaining for the loopback branch at the end of a loop for the most common loop sizes. With this change the resulting emulation speed is about 10% higher in softmmu mode on uClibc-ng and LTP tests. Emulation speed in linux user mode is a few percent lower because there's no direct TB chaining between different memory pages. Testing with lower limit on direct TB chaining range shows gradual slowdown to ~15% for the block size of 64 bytes and ~50% for the block size of 32 bytes. Signed-off-by: Max Filippov <jcmvbkbc@gmail.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
5713 lines
175 KiB
C
5713 lines
175 KiB
C
/*
|
|
* Xtensa ISA:
|
|
* http://www.tensilica.com/products/literature-docs/documentation/xtensa-isa-databook.htm
|
|
*
|
|
* Copyright (c) 2011, Max Filippov, Open Source and Linux Lab.
|
|
* All rights reserved.
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without
|
|
* modification, are permitted provided that the following conditions are met:
|
|
* * Redistributions of source code must retain the above copyright
|
|
* notice, this list of conditions and the following disclaimer.
|
|
* * Redistributions in binary form must reproduce the above copyright
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
* documentation and/or other materials provided with the distribution.
|
|
* * Neither the name of the Open Source and Linux Lab nor the
|
|
* names of its contributors may be used to endorse or promote products
|
|
* derived from this software without specific prior written permission.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
|
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
|
|
* DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
|
|
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
|
|
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
|
|
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
|
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
|
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
*/
|
|
|
|
#include "qemu/osdep.h"
|
|
|
|
#include "cpu.h"
|
|
#include "exec/exec-all.h"
|
|
#include "disas/disas.h"
|
|
#include "tcg-op.h"
|
|
#include "qemu/log.h"
|
|
#include "sysemu/sysemu.h"
|
|
#include "exec/cpu_ldst.h"
|
|
#include "exec/semihost.h"
|
|
#include "exec/translator.h"
|
|
|
|
#include "exec/helper-proto.h"
|
|
#include "exec/helper-gen.h"
|
|
|
|
#include "trace-tcg.h"
|
|
#include "exec/log.h"
|
|
|
|
|
|
struct DisasContext {
|
|
DisasContextBase base;
|
|
const XtensaConfig *config;
|
|
uint32_t pc;
|
|
int cring;
|
|
int ring;
|
|
uint32_t lbeg_off;
|
|
uint32_t lend;
|
|
|
|
bool sar_5bit;
|
|
bool sar_m32_5bit;
|
|
bool sar_m32_allocated;
|
|
TCGv_i32 sar_m32;
|
|
|
|
unsigned window;
|
|
unsigned callinc;
|
|
bool cwoe;
|
|
|
|
bool debug;
|
|
bool icount;
|
|
TCGv_i32 next_icount;
|
|
|
|
unsigned cpenable;
|
|
|
|
uint32_t *raw_arg;
|
|
xtensa_insnbuf insnbuf;
|
|
xtensa_insnbuf slotbuf;
|
|
};
|
|
|
|
static TCGv_i32 cpu_pc;
|
|
static TCGv_i32 cpu_R[16];
|
|
static TCGv_i32 cpu_FR[16];
|
|
static TCGv_i32 cpu_SR[256];
|
|
static TCGv_i32 cpu_UR[256];
|
|
|
|
#include "exec/gen-icount.h"
|
|
|
|
typedef struct XtensaReg {
|
|
const char *name;
|
|
uint64_t opt_bits;
|
|
enum {
|
|
SR_R = 1,
|
|
SR_W = 2,
|
|
SR_X = 4,
|
|
SR_RW = 3,
|
|
SR_RWX = 7,
|
|
} access;
|
|
} XtensaReg;
|
|
|
|
#define XTENSA_REG_ACCESS(regname, opt, acc) { \
|
|
.name = (regname), \
|
|
.opt_bits = XTENSA_OPTION_BIT(opt), \
|
|
.access = (acc), \
|
|
}
|
|
|
|
#define XTENSA_REG(regname, opt) XTENSA_REG_ACCESS(regname, opt, SR_RWX)
|
|
|
|
#define XTENSA_REG_BITS_ACCESS(regname, opt, acc) { \
|
|
.name = (regname), \
|
|
.opt_bits = (opt), \
|
|
.access = (acc), \
|
|
}
|
|
|
|
#define XTENSA_REG_BITS(regname, opt) \
|
|
XTENSA_REG_BITS_ACCESS(regname, opt, SR_RWX)
|
|
|
|
static const XtensaReg sregnames[256] = {
|
|
[LBEG] = XTENSA_REG("LBEG", XTENSA_OPTION_LOOP),
|
|
[LEND] = XTENSA_REG("LEND", XTENSA_OPTION_LOOP),
|
|
[LCOUNT] = XTENSA_REG("LCOUNT", XTENSA_OPTION_LOOP),
|
|
[SAR] = XTENSA_REG_BITS("SAR", XTENSA_OPTION_ALL),
|
|
[BR] = XTENSA_REG("BR", XTENSA_OPTION_BOOLEAN),
|
|
[LITBASE] = XTENSA_REG("LITBASE", XTENSA_OPTION_EXTENDED_L32R),
|
|
[SCOMPARE1] = XTENSA_REG("SCOMPARE1", XTENSA_OPTION_CONDITIONAL_STORE),
|
|
[ACCLO] = XTENSA_REG("ACCLO", XTENSA_OPTION_MAC16),
|
|
[ACCHI] = XTENSA_REG("ACCHI", XTENSA_OPTION_MAC16),
|
|
[MR] = XTENSA_REG("MR0", XTENSA_OPTION_MAC16),
|
|
[MR + 1] = XTENSA_REG("MR1", XTENSA_OPTION_MAC16),
|
|
[MR + 2] = XTENSA_REG("MR2", XTENSA_OPTION_MAC16),
|
|
[MR + 3] = XTENSA_REG("MR3", XTENSA_OPTION_MAC16),
|
|
[WINDOW_BASE] = XTENSA_REG("WINDOW_BASE", XTENSA_OPTION_WINDOWED_REGISTER),
|
|
[WINDOW_START] = XTENSA_REG("WINDOW_START",
|
|
XTENSA_OPTION_WINDOWED_REGISTER),
|
|
[PTEVADDR] = XTENSA_REG("PTEVADDR", XTENSA_OPTION_MMU),
|
|
[MMID] = XTENSA_REG_BITS("MMID", XTENSA_OPTION_ALL),
|
|
[RASID] = XTENSA_REG("RASID", XTENSA_OPTION_MMU),
|
|
[ITLBCFG] = XTENSA_REG("ITLBCFG", XTENSA_OPTION_MMU),
|
|
[DTLBCFG] = XTENSA_REG("DTLBCFG", XTENSA_OPTION_MMU),
|
|
[IBREAKENABLE] = XTENSA_REG("IBREAKENABLE", XTENSA_OPTION_DEBUG),
|
|
[MEMCTL] = XTENSA_REG_BITS("MEMCTL", XTENSA_OPTION_ALL),
|
|
[CACHEATTR] = XTENSA_REG("CACHEATTR", XTENSA_OPTION_CACHEATTR),
|
|
[ATOMCTL] = XTENSA_REG("ATOMCTL", XTENSA_OPTION_ATOMCTL),
|
|
[DDR] = XTENSA_REG("DDR", XTENSA_OPTION_DEBUG),
|
|
[IBREAKA] = XTENSA_REG("IBREAKA0", XTENSA_OPTION_DEBUG),
|
|
[IBREAKA + 1] = XTENSA_REG("IBREAKA1", XTENSA_OPTION_DEBUG),
|
|
[DBREAKA] = XTENSA_REG("DBREAKA0", XTENSA_OPTION_DEBUG),
|
|
[DBREAKA + 1] = XTENSA_REG("DBREAKA1", XTENSA_OPTION_DEBUG),
|
|
[DBREAKC] = XTENSA_REG("DBREAKC0", XTENSA_OPTION_DEBUG),
|
|
[DBREAKC + 1] = XTENSA_REG("DBREAKC1", XTENSA_OPTION_DEBUG),
|
|
[CONFIGID0] = XTENSA_REG_BITS_ACCESS("CONFIGID0", XTENSA_OPTION_ALL, SR_R),
|
|
[EPC1] = XTENSA_REG("EPC1", XTENSA_OPTION_EXCEPTION),
|
|
[EPC1 + 1] = XTENSA_REG("EPC2", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPC1 + 2] = XTENSA_REG("EPC3", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPC1 + 3] = XTENSA_REG("EPC4", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPC1 + 4] = XTENSA_REG("EPC5", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPC1 + 5] = XTENSA_REG("EPC6", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPC1 + 6] = XTENSA_REG("EPC7", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[DEPC] = XTENSA_REG("DEPC", XTENSA_OPTION_EXCEPTION),
|
|
[EPS2] = XTENSA_REG("EPS2", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPS2 + 1] = XTENSA_REG("EPS3", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPS2 + 2] = XTENSA_REG("EPS4", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPS2 + 3] = XTENSA_REG("EPS5", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPS2 + 4] = XTENSA_REG("EPS6", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EPS2 + 5] = XTENSA_REG("EPS7", XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[CONFIGID1] = XTENSA_REG_BITS_ACCESS("CONFIGID1", XTENSA_OPTION_ALL, SR_R),
|
|
[EXCSAVE1] = XTENSA_REG("EXCSAVE1", XTENSA_OPTION_EXCEPTION),
|
|
[EXCSAVE1 + 1] = XTENSA_REG("EXCSAVE2",
|
|
XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EXCSAVE1 + 2] = XTENSA_REG("EXCSAVE3",
|
|
XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EXCSAVE1 + 3] = XTENSA_REG("EXCSAVE4",
|
|
XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EXCSAVE1 + 4] = XTENSA_REG("EXCSAVE5",
|
|
XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EXCSAVE1 + 5] = XTENSA_REG("EXCSAVE6",
|
|
XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[EXCSAVE1 + 6] = XTENSA_REG("EXCSAVE7",
|
|
XTENSA_OPTION_HIGH_PRIORITY_INTERRUPT),
|
|
[CPENABLE] = XTENSA_REG("CPENABLE", XTENSA_OPTION_COPROCESSOR),
|
|
[INTSET] = XTENSA_REG_ACCESS("INTSET", XTENSA_OPTION_INTERRUPT, SR_RW),
|
|
[INTCLEAR] = XTENSA_REG_ACCESS("INTCLEAR", XTENSA_OPTION_INTERRUPT, SR_W),
|
|
[INTENABLE] = XTENSA_REG("INTENABLE", XTENSA_OPTION_INTERRUPT),
|
|
[PS] = XTENSA_REG_BITS("PS", XTENSA_OPTION_ALL),
|
|
[VECBASE] = XTENSA_REG("VECBASE", XTENSA_OPTION_RELOCATABLE_VECTOR),
|
|
[EXCCAUSE] = XTENSA_REG("EXCCAUSE", XTENSA_OPTION_EXCEPTION),
|
|
[DEBUGCAUSE] = XTENSA_REG_ACCESS("DEBUGCAUSE", XTENSA_OPTION_DEBUG, SR_R),
|
|
[CCOUNT] = XTENSA_REG("CCOUNT", XTENSA_OPTION_TIMER_INTERRUPT),
|
|
[PRID] = XTENSA_REG_ACCESS("PRID", XTENSA_OPTION_PROCESSOR_ID, SR_R),
|
|
[ICOUNT] = XTENSA_REG("ICOUNT", XTENSA_OPTION_DEBUG),
|
|
[ICOUNTLEVEL] = XTENSA_REG("ICOUNTLEVEL", XTENSA_OPTION_DEBUG),
|
|
[EXCVADDR] = XTENSA_REG("EXCVADDR", XTENSA_OPTION_EXCEPTION),
|
|
[CCOMPARE] = XTENSA_REG("CCOMPARE0", XTENSA_OPTION_TIMER_INTERRUPT),
|
|
[CCOMPARE + 1] = XTENSA_REG("CCOMPARE1",
|
|
XTENSA_OPTION_TIMER_INTERRUPT),
|
|
[CCOMPARE + 2] = XTENSA_REG("CCOMPARE2",
|
|
XTENSA_OPTION_TIMER_INTERRUPT),
|
|
[MISC] = XTENSA_REG("MISC0", XTENSA_OPTION_MISC_SR),
|
|
[MISC + 1] = XTENSA_REG("MISC1", XTENSA_OPTION_MISC_SR),
|
|
[MISC + 2] = XTENSA_REG("MISC2", XTENSA_OPTION_MISC_SR),
|
|
[MISC + 3] = XTENSA_REG("MISC3", XTENSA_OPTION_MISC_SR),
|
|
};
|
|
|
|
static const XtensaReg uregnames[256] = {
|
|
[EXPSTATE] = XTENSA_REG_BITS("EXPSTATE", XTENSA_OPTION_ALL),
|
|
[THREADPTR] = XTENSA_REG("THREADPTR", XTENSA_OPTION_THREAD_POINTER),
|
|
[FCR] = XTENSA_REG("FCR", XTENSA_OPTION_FP_COPROCESSOR),
|
|
[FSR] = XTENSA_REG("FSR", XTENSA_OPTION_FP_COPROCESSOR),
|
|
};
|
|
|
|
void xtensa_translate_init(void)
|
|
{
|
|
static const char * const regnames[] = {
|
|
"ar0", "ar1", "ar2", "ar3",
|
|
"ar4", "ar5", "ar6", "ar7",
|
|
"ar8", "ar9", "ar10", "ar11",
|
|
"ar12", "ar13", "ar14", "ar15",
|
|
};
|
|
static const char * const fregnames[] = {
|
|
"f0", "f1", "f2", "f3",
|
|
"f4", "f5", "f6", "f7",
|
|
"f8", "f9", "f10", "f11",
|
|
"f12", "f13", "f14", "f15",
|
|
};
|
|
int i;
|
|
|
|
cpu_pc = tcg_global_mem_new_i32(cpu_env,
|
|
offsetof(CPUXtensaState, pc), "pc");
|
|
|
|
for (i = 0; i < 16; i++) {
|
|
cpu_R[i] = tcg_global_mem_new_i32(cpu_env,
|
|
offsetof(CPUXtensaState, regs[i]),
|
|
regnames[i]);
|
|
}
|
|
|
|
for (i = 0; i < 16; i++) {
|
|
cpu_FR[i] = tcg_global_mem_new_i32(cpu_env,
|
|
offsetof(CPUXtensaState, fregs[i].f32[FP_F32_LOW]),
|
|
fregnames[i]);
|
|
}
|
|
|
|
for (i = 0; i < 256; ++i) {
|
|
if (sregnames[i].name) {
|
|
cpu_SR[i] = tcg_global_mem_new_i32(cpu_env,
|
|
offsetof(CPUXtensaState, sregs[i]),
|
|
sregnames[i].name);
|
|
}
|
|
}
|
|
|
|
for (i = 0; i < 256; ++i) {
|
|
if (uregnames[i].name) {
|
|
cpu_UR[i] = tcg_global_mem_new_i32(cpu_env,
|
|
offsetof(CPUXtensaState, uregs[i]),
|
|
uregnames[i].name);
|
|
}
|
|
}
|
|
}
|
|
|
|
static inline bool option_enabled(DisasContext *dc, int opt)
|
|
{
|
|
return xtensa_option_enabled(dc->config, opt);
|
|
}
|
|
|
|
static void init_sar_tracker(DisasContext *dc)
|
|
{
|
|
dc->sar_5bit = false;
|
|
dc->sar_m32_5bit = false;
|
|
dc->sar_m32_allocated = false;
|
|
}
|
|
|
|
static void reset_sar_tracker(DisasContext *dc)
|
|
{
|
|
if (dc->sar_m32_allocated) {
|
|
tcg_temp_free(dc->sar_m32);
|
|
}
|
|
}
|
|
|
|
static void gen_right_shift_sar(DisasContext *dc, TCGv_i32 sa)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[SAR], sa, 0x1f);
|
|
if (dc->sar_m32_5bit) {
|
|
tcg_gen_discard_i32(dc->sar_m32);
|
|
}
|
|
dc->sar_5bit = true;
|
|
dc->sar_m32_5bit = false;
|
|
}
|
|
|
|
static void gen_left_shift_sar(DisasContext *dc, TCGv_i32 sa)
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(32);
|
|
if (!dc->sar_m32_allocated) {
|
|
dc->sar_m32 = tcg_temp_local_new_i32();
|
|
dc->sar_m32_allocated = true;
|
|
}
|
|
tcg_gen_andi_i32(dc->sar_m32, sa, 0x1f);
|
|
tcg_gen_sub_i32(cpu_SR[SAR], tmp, dc->sar_m32);
|
|
dc->sar_5bit = false;
|
|
dc->sar_m32_5bit = true;
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void gen_exception(DisasContext *dc, int excp)
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(excp);
|
|
gen_helper_exception(cpu_env, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void gen_exception_cause(DisasContext *dc, uint32_t cause)
|
|
{
|
|
TCGv_i32 tpc = tcg_const_i32(dc->pc);
|
|
TCGv_i32 tcause = tcg_const_i32(cause);
|
|
gen_helper_exception_cause(cpu_env, tpc, tcause);
|
|
tcg_temp_free(tpc);
|
|
tcg_temp_free(tcause);
|
|
if (cause == ILLEGAL_INSTRUCTION_CAUSE ||
|
|
cause == SYSCALL_CAUSE) {
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
}
|
|
}
|
|
|
|
static void gen_exception_cause_vaddr(DisasContext *dc, uint32_t cause,
|
|
TCGv_i32 vaddr)
|
|
{
|
|
TCGv_i32 tpc = tcg_const_i32(dc->pc);
|
|
TCGv_i32 tcause = tcg_const_i32(cause);
|
|
gen_helper_exception_cause_vaddr(cpu_env, tpc, tcause, vaddr);
|
|
tcg_temp_free(tpc);
|
|
tcg_temp_free(tcause);
|
|
}
|
|
|
|
static void gen_debug_exception(DisasContext *dc, uint32_t cause)
|
|
{
|
|
TCGv_i32 tpc = tcg_const_i32(dc->pc);
|
|
TCGv_i32 tcause = tcg_const_i32(cause);
|
|
gen_helper_debug_exception(cpu_env, tpc, tcause);
|
|
tcg_temp_free(tpc);
|
|
tcg_temp_free(tcause);
|
|
if (cause & (DEBUGCAUSE_IB | DEBUGCAUSE_BI | DEBUGCAUSE_BN)) {
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
}
|
|
}
|
|
|
|
static bool gen_check_privilege(DisasContext *dc)
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
if (!dc->cring) {
|
|
return true;
|
|
}
|
|
#endif
|
|
gen_exception_cause(dc, PRIVILEGED_CAUSE);
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
return false;
|
|
}
|
|
|
|
static bool gen_check_cpenable(DisasContext *dc, uint32_t cp_mask)
|
|
{
|
|
cp_mask &= ~dc->cpenable;
|
|
|
|
if (option_enabled(dc, XTENSA_OPTION_COPROCESSOR) && cp_mask) {
|
|
gen_exception_cause(dc, COPROCESSOR0_DISABLED + ctz32(cp_mask));
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
static void gen_jump_slot(DisasContext *dc, TCGv dest, int slot)
|
|
{
|
|
tcg_gen_mov_i32(cpu_pc, dest);
|
|
if (dc->icount) {
|
|
tcg_gen_mov_i32(cpu_SR[ICOUNT], dc->next_icount);
|
|
}
|
|
if (dc->base.singlestep_enabled) {
|
|
gen_exception(dc, EXCP_DEBUG);
|
|
} else {
|
|
if (slot >= 0) {
|
|
tcg_gen_goto_tb(slot);
|
|
tcg_gen_exit_tb(dc->base.tb, slot);
|
|
} else {
|
|
tcg_gen_exit_tb(NULL, 0);
|
|
}
|
|
}
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
}
|
|
|
|
static void gen_jump(DisasContext *dc, TCGv dest)
|
|
{
|
|
gen_jump_slot(dc, dest, -1);
|
|
}
|
|
|
|
static void gen_jumpi(DisasContext *dc, uint32_t dest, int slot)
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(dest);
|
|
if (((dc->base.pc_first ^ dest) & TARGET_PAGE_MASK) != 0) {
|
|
slot = -1;
|
|
}
|
|
gen_jump_slot(dc, tmp, slot);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void gen_callw_slot(DisasContext *dc, int callinc, TCGv_i32 dest,
|
|
int slot)
|
|
{
|
|
TCGv_i32 tcallinc = tcg_const_i32(callinc);
|
|
|
|
tcg_gen_deposit_i32(cpu_SR[PS], cpu_SR[PS],
|
|
tcallinc, PS_CALLINC_SHIFT, PS_CALLINC_LEN);
|
|
tcg_temp_free(tcallinc);
|
|
tcg_gen_movi_i32(cpu_R[callinc << 2],
|
|
(callinc << 30) | (dc->base.pc_next & 0x3fffffff));
|
|
gen_jump_slot(dc, dest, slot);
|
|
}
|
|
|
|
static void gen_callw(DisasContext *dc, int callinc, TCGv_i32 dest)
|
|
{
|
|
gen_callw_slot(dc, callinc, dest, -1);
|
|
}
|
|
|
|
static void gen_callwi(DisasContext *dc, int callinc, uint32_t dest, int slot)
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(dest);
|
|
if (((dc->base.pc_first ^ dest) & TARGET_PAGE_MASK) != 0) {
|
|
slot = -1;
|
|
}
|
|
gen_callw_slot(dc, callinc, tmp, slot);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static bool gen_check_loop_end(DisasContext *dc, int slot)
|
|
{
|
|
if (dc->base.pc_next == dc->lend) {
|
|
TCGLabel *label = gen_new_label();
|
|
|
|
tcg_gen_brcondi_i32(TCG_COND_EQ, cpu_SR[LCOUNT], 0, label);
|
|
tcg_gen_subi_i32(cpu_SR[LCOUNT], cpu_SR[LCOUNT], 1);
|
|
if (dc->lbeg_off) {
|
|
gen_jumpi(dc, dc->base.pc_next - dc->lbeg_off, slot);
|
|
} else {
|
|
gen_jump(dc, cpu_SR[LBEG]);
|
|
}
|
|
gen_set_label(label);
|
|
gen_jumpi(dc, dc->base.pc_next, -1);
|
|
return true;
|
|
}
|
|
return false;
|
|
}
|
|
|
|
static void gen_jumpi_check_loop_end(DisasContext *dc, int slot)
|
|
{
|
|
if (!gen_check_loop_end(dc, slot)) {
|
|
gen_jumpi(dc, dc->base.pc_next, slot);
|
|
}
|
|
}
|
|
|
|
static void gen_brcond(DisasContext *dc, TCGCond cond,
|
|
TCGv_i32 t0, TCGv_i32 t1, uint32_t addr)
|
|
{
|
|
TCGLabel *label = gen_new_label();
|
|
|
|
tcg_gen_brcond_i32(cond, t0, t1, label);
|
|
gen_jumpi_check_loop_end(dc, 0);
|
|
gen_set_label(label);
|
|
gen_jumpi(dc, addr, 1);
|
|
}
|
|
|
|
static void gen_brcondi(DisasContext *dc, TCGCond cond,
|
|
TCGv_i32 t0, uint32_t t1, uint32_t addr)
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(t1);
|
|
gen_brcond(dc, cond, t0, tmp, addr);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static bool check_sr(DisasContext *dc, uint32_t sr, unsigned access)
|
|
{
|
|
if (!xtensa_option_bits_enabled(dc->config, sregnames[sr].opt_bits)) {
|
|
if (sregnames[sr].name) {
|
|
qemu_log_mask(LOG_GUEST_ERROR, "SR %s is not configured\n", sregnames[sr].name);
|
|
} else {
|
|
qemu_log_mask(LOG_UNIMP, "SR %d is not implemented\n", sr);
|
|
}
|
|
return false;
|
|
} else if (!(sregnames[sr].access & access)) {
|
|
static const char * const access_text[] = {
|
|
[SR_R] = "rsr",
|
|
[SR_W] = "wsr",
|
|
[SR_X] = "xsr",
|
|
};
|
|
assert(access < ARRAY_SIZE(access_text) && access_text[access]);
|
|
qemu_log_mask(LOG_GUEST_ERROR, "SR %s is not available for %s\n", sregnames[sr].name,
|
|
access_text[access]);
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
#ifndef CONFIG_USER_ONLY
|
|
static void gen_rsr_ccount(DisasContext *dc, TCGv_i32 d, uint32_t sr)
|
|
{
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_start();
|
|
}
|
|
gen_helper_update_ccount(cpu_env);
|
|
tcg_gen_mov_i32(d, cpu_SR[sr]);
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_end();
|
|
}
|
|
}
|
|
|
|
static void gen_rsr_ptevaddr(DisasContext *dc, TCGv_i32 d, uint32_t sr)
|
|
{
|
|
tcg_gen_shri_i32(d, cpu_SR[EXCVADDR], 10);
|
|
tcg_gen_or_i32(d, d, cpu_SR[sr]);
|
|
tcg_gen_andi_i32(d, d, 0xfffffffc);
|
|
}
|
|
#endif
|
|
|
|
static void gen_rsr(DisasContext *dc, TCGv_i32 d, uint32_t sr)
|
|
{
|
|
static void (* const rsr_handler[256])(DisasContext *dc,
|
|
TCGv_i32 d, uint32_t sr) = {
|
|
#ifndef CONFIG_USER_ONLY
|
|
[CCOUNT] = gen_rsr_ccount,
|
|
[INTSET] = gen_rsr_ccount,
|
|
[PTEVADDR] = gen_rsr_ptevaddr,
|
|
#endif
|
|
};
|
|
|
|
if (rsr_handler[sr]) {
|
|
rsr_handler[sr](dc, d, sr);
|
|
} else {
|
|
tcg_gen_mov_i32(d, cpu_SR[sr]);
|
|
}
|
|
}
|
|
|
|
static void gen_wsr_sar(DisasContext *dc, uint32_t sr, TCGv_i32 s)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], s, 0x3f);
|
|
if (dc->sar_m32_5bit) {
|
|
tcg_gen_discard_i32(dc->sar_m32);
|
|
}
|
|
dc->sar_5bit = false;
|
|
dc->sar_m32_5bit = false;
|
|
}
|
|
|
|
static void gen_wsr_br(DisasContext *dc, uint32_t sr, TCGv_i32 s)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], s, 0xffff);
|
|
}
|
|
|
|
static void gen_wsr_litbase(DisasContext *dc, uint32_t sr, TCGv_i32 s)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], s, 0xfffff001);
|
|
}
|
|
|
|
static void gen_wsr_acchi(DisasContext *dc, uint32_t sr, TCGv_i32 s)
|
|
{
|
|
tcg_gen_ext8s_i32(cpu_SR[sr], s);
|
|
}
|
|
|
|
#ifndef CONFIG_USER_ONLY
|
|
static void gen_wsr_windowbase(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
gen_helper_wsr_windowbase(cpu_env, v);
|
|
}
|
|
|
|
static void gen_wsr_windowstart(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], v, (1 << dc->config->nareg / 4) - 1);
|
|
}
|
|
|
|
static void gen_wsr_ptevaddr(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], v, 0xffc00000);
|
|
}
|
|
|
|
static void gen_wsr_rasid(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
gen_helper_wsr_rasid(cpu_env, v);
|
|
}
|
|
|
|
static void gen_wsr_tlbcfg(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], v, 0x01130000);
|
|
}
|
|
|
|
static void gen_wsr_ibreakenable(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
gen_helper_wsr_ibreakenable(cpu_env, v);
|
|
}
|
|
|
|
static void gen_wsr_memctl(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
gen_helper_wsr_memctl(cpu_env, v);
|
|
}
|
|
|
|
static void gen_wsr_atomctl(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], v, 0x3f);
|
|
}
|
|
|
|
static void gen_wsr_ibreaka(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
unsigned id = sr - IBREAKA;
|
|
TCGv_i32 tmp = tcg_const_i32(id);
|
|
|
|
assert(id < dc->config->nibreak);
|
|
gen_helper_wsr_ibreaka(cpu_env, tmp, v);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void gen_wsr_dbreaka(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
unsigned id = sr - DBREAKA;
|
|
TCGv_i32 tmp = tcg_const_i32(id);
|
|
|
|
assert(id < dc->config->ndbreak);
|
|
gen_helper_wsr_dbreaka(cpu_env, tmp, v);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void gen_wsr_dbreakc(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
unsigned id = sr - DBREAKC;
|
|
TCGv_i32 tmp = tcg_const_i32(id);
|
|
|
|
assert(id < dc->config->ndbreak);
|
|
gen_helper_wsr_dbreakc(cpu_env, tmp, v);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void gen_wsr_cpenable(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], v, 0xff);
|
|
}
|
|
|
|
static void gen_check_interrupts(DisasContext *dc)
|
|
{
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_start();
|
|
}
|
|
gen_helper_check_interrupts(cpu_env);
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_end();
|
|
}
|
|
}
|
|
|
|
static void gen_wsr_intset(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], v,
|
|
dc->config->inttype_mask[INTTYPE_SOFTWARE]);
|
|
}
|
|
|
|
static void gen_wsr_intclear(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
|
|
tcg_gen_andi_i32(tmp, v,
|
|
dc->config->inttype_mask[INTTYPE_EDGE] |
|
|
dc->config->inttype_mask[INTTYPE_NMI] |
|
|
dc->config->inttype_mask[INTTYPE_SOFTWARE]);
|
|
tcg_gen_andc_i32(cpu_SR[INTSET], cpu_SR[INTSET], tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void gen_wsr_intenable(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_mov_i32(cpu_SR[sr], v);
|
|
}
|
|
|
|
static void gen_wsr_ps(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
uint32_t mask = PS_WOE | PS_CALLINC | PS_OWB |
|
|
PS_UM | PS_EXCM | PS_INTLEVEL;
|
|
|
|
if (option_enabled(dc, XTENSA_OPTION_MMU)) {
|
|
mask |= PS_RING;
|
|
}
|
|
tcg_gen_andi_i32(cpu_SR[sr], v, mask);
|
|
}
|
|
|
|
static void gen_wsr_ccount(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_start();
|
|
}
|
|
gen_helper_wsr_ccount(cpu_env, v);
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_end();
|
|
}
|
|
}
|
|
|
|
static void gen_wsr_icount(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
if (dc->icount) {
|
|
tcg_gen_mov_i32(dc->next_icount, v);
|
|
} else {
|
|
tcg_gen_mov_i32(cpu_SR[sr], v);
|
|
}
|
|
}
|
|
|
|
static void gen_wsr_icountlevel(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[sr], v, 0xf);
|
|
}
|
|
|
|
static void gen_wsr_ccompare(DisasContext *dc, uint32_t sr, TCGv_i32 v)
|
|
{
|
|
uint32_t id = sr - CCOMPARE;
|
|
uint32_t int_bit = 1 << dc->config->timerint[id];
|
|
TCGv_i32 tmp = tcg_const_i32(id);
|
|
|
|
assert(id < dc->config->nccompare);
|
|
tcg_gen_mov_i32(cpu_SR[sr], v);
|
|
tcg_gen_andi_i32(cpu_SR[INTSET], cpu_SR[INTSET], ~int_bit);
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_start();
|
|
}
|
|
gen_helper_update_ccompare(cpu_env, tmp);
|
|
tcg_temp_free(tmp);
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_end();
|
|
}
|
|
}
|
|
#else
|
|
static void gen_check_interrupts(DisasContext *dc)
|
|
{
|
|
}
|
|
#endif
|
|
|
|
static void gen_wsr(DisasContext *dc, uint32_t sr, TCGv_i32 s)
|
|
{
|
|
static void (* const wsr_handler[256])(DisasContext *dc,
|
|
uint32_t sr, TCGv_i32 v) = {
|
|
[SAR] = gen_wsr_sar,
|
|
[BR] = gen_wsr_br,
|
|
[LITBASE] = gen_wsr_litbase,
|
|
[ACCHI] = gen_wsr_acchi,
|
|
#ifndef CONFIG_USER_ONLY
|
|
[WINDOW_BASE] = gen_wsr_windowbase,
|
|
[WINDOW_START] = gen_wsr_windowstart,
|
|
[PTEVADDR] = gen_wsr_ptevaddr,
|
|
[RASID] = gen_wsr_rasid,
|
|
[ITLBCFG] = gen_wsr_tlbcfg,
|
|
[DTLBCFG] = gen_wsr_tlbcfg,
|
|
[IBREAKENABLE] = gen_wsr_ibreakenable,
|
|
[MEMCTL] = gen_wsr_memctl,
|
|
[ATOMCTL] = gen_wsr_atomctl,
|
|
[IBREAKA] = gen_wsr_ibreaka,
|
|
[IBREAKA + 1] = gen_wsr_ibreaka,
|
|
[DBREAKA] = gen_wsr_dbreaka,
|
|
[DBREAKA + 1] = gen_wsr_dbreaka,
|
|
[DBREAKC] = gen_wsr_dbreakc,
|
|
[DBREAKC + 1] = gen_wsr_dbreakc,
|
|
[CPENABLE] = gen_wsr_cpenable,
|
|
[INTSET] = gen_wsr_intset,
|
|
[INTCLEAR] = gen_wsr_intclear,
|
|
[INTENABLE] = gen_wsr_intenable,
|
|
[PS] = gen_wsr_ps,
|
|
[CCOUNT] = gen_wsr_ccount,
|
|
[ICOUNT] = gen_wsr_icount,
|
|
[ICOUNTLEVEL] = gen_wsr_icountlevel,
|
|
[CCOMPARE] = gen_wsr_ccompare,
|
|
[CCOMPARE + 1] = gen_wsr_ccompare,
|
|
[CCOMPARE + 2] = gen_wsr_ccompare,
|
|
#endif
|
|
};
|
|
|
|
if (wsr_handler[sr]) {
|
|
wsr_handler[sr](dc, sr, s);
|
|
} else {
|
|
tcg_gen_mov_i32(cpu_SR[sr], s);
|
|
}
|
|
}
|
|
|
|
static void gen_wur(uint32_t ur, TCGv_i32 s)
|
|
{
|
|
switch (ur) {
|
|
case FCR:
|
|
gen_helper_wur_fcr(cpu_env, s);
|
|
break;
|
|
|
|
case FSR:
|
|
tcg_gen_andi_i32(cpu_UR[ur], s, 0xffffff80);
|
|
break;
|
|
|
|
default:
|
|
tcg_gen_mov_i32(cpu_UR[ur], s);
|
|
break;
|
|
}
|
|
}
|
|
|
|
static void gen_load_store_alignment(DisasContext *dc, int shift,
|
|
TCGv_i32 addr, bool no_hw_alignment)
|
|
{
|
|
if (!option_enabled(dc, XTENSA_OPTION_UNALIGNED_EXCEPTION)) {
|
|
tcg_gen_andi_i32(addr, addr, ~0 << shift);
|
|
} else if (option_enabled(dc, XTENSA_OPTION_HW_ALIGNMENT) &&
|
|
no_hw_alignment) {
|
|
TCGLabel *label = gen_new_label();
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_andi_i32(tmp, addr, ~(~0 << shift));
|
|
tcg_gen_brcondi_i32(TCG_COND_EQ, tmp, 0, label);
|
|
gen_exception_cause_vaddr(dc, LOAD_STORE_ALIGNMENT_CAUSE, addr);
|
|
gen_set_label(label);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
}
|
|
|
|
#ifndef CONFIG_USER_ONLY
|
|
static void gen_waiti(DisasContext *dc, uint32_t imm4)
|
|
{
|
|
TCGv_i32 pc = tcg_const_i32(dc->base.pc_next);
|
|
TCGv_i32 intlevel = tcg_const_i32(imm4);
|
|
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_start();
|
|
}
|
|
gen_helper_waiti(cpu_env, pc, intlevel);
|
|
if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) {
|
|
gen_io_end();
|
|
}
|
|
tcg_temp_free(pc);
|
|
tcg_temp_free(intlevel);
|
|
}
|
|
#endif
|
|
|
|
static bool gen_window_check(DisasContext *dc, uint32_t mask)
|
|
{
|
|
unsigned r = 31 - clz32(mask);
|
|
|
|
if (r / 4 > dc->window) {
|
|
TCGv_i32 pc = tcg_const_i32(dc->pc);
|
|
TCGv_i32 w = tcg_const_i32(r / 4);
|
|
|
|
gen_helper_window_check(cpu_env, pc, w);
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
static TCGv_i32 gen_mac16_m(TCGv_i32 v, bool hi, bool is_unsigned)
|
|
{
|
|
TCGv_i32 m = tcg_temp_new_i32();
|
|
|
|
if (hi) {
|
|
(is_unsigned ? tcg_gen_shri_i32 : tcg_gen_sari_i32)(m, v, 16);
|
|
} else {
|
|
(is_unsigned ? tcg_gen_ext16u_i32 : tcg_gen_ext16s_i32)(m, v);
|
|
}
|
|
return m;
|
|
}
|
|
|
|
static void gen_zero_check(DisasContext *dc, const uint32_t arg[])
|
|
{
|
|
TCGLabel *label = gen_new_label();
|
|
|
|
tcg_gen_brcondi_i32(TCG_COND_NE, cpu_R[arg[2]], 0, label);
|
|
gen_exception_cause(dc, INTEGER_DIVIDE_BY_ZERO_CAUSE);
|
|
gen_set_label(label);
|
|
}
|
|
|
|
static inline unsigned xtensa_op0_insn_len(DisasContext *dc, uint8_t op0)
|
|
{
|
|
return xtensa_isa_length_from_chars(dc->config->isa, &op0);
|
|
}
|
|
|
|
static void disas_xtensa_insn(CPUXtensaState *env, DisasContext *dc)
|
|
{
|
|
xtensa_isa isa = dc->config->isa;
|
|
unsigned char b[MAX_INSN_LENGTH] = {cpu_ldub_code(env, dc->pc)};
|
|
unsigned len = xtensa_op0_insn_len(dc, b[0]);
|
|
xtensa_format fmt;
|
|
int slot, slots;
|
|
unsigned i;
|
|
uint32_t op_flags = 0;
|
|
struct {
|
|
XtensaOpcodeOps *ops;
|
|
uint32_t arg[MAX_OPCODE_ARGS];
|
|
uint32_t raw_arg[MAX_OPCODE_ARGS];
|
|
} slot_prop[MAX_INSN_SLOTS];
|
|
uint32_t debug_cause = 0;
|
|
uint32_t windowed_register = 0;
|
|
uint32_t coprocessor = 0;
|
|
|
|
if (len == XTENSA_UNDEFINED) {
|
|
qemu_log_mask(LOG_GUEST_ERROR,
|
|
"unknown instruction length (pc = %08x)\n",
|
|
dc->pc);
|
|
gen_exception_cause(dc, ILLEGAL_INSTRUCTION_CAUSE);
|
|
return;
|
|
}
|
|
|
|
dc->base.pc_next = dc->pc + len;
|
|
for (i = 1; i < len; ++i) {
|
|
b[i] = cpu_ldub_code(env, dc->pc + i);
|
|
}
|
|
xtensa_insnbuf_from_chars(isa, dc->insnbuf, b, len);
|
|
fmt = xtensa_format_decode(isa, dc->insnbuf);
|
|
if (fmt == XTENSA_UNDEFINED) {
|
|
qemu_log_mask(LOG_GUEST_ERROR,
|
|
"unrecognized instruction format (pc = %08x)\n",
|
|
dc->pc);
|
|
gen_exception_cause(dc, ILLEGAL_INSTRUCTION_CAUSE);
|
|
return;
|
|
}
|
|
slots = xtensa_format_num_slots(isa, fmt);
|
|
for (slot = 0; slot < slots; ++slot) {
|
|
xtensa_opcode opc;
|
|
int opnd, vopnd, opnds;
|
|
uint32_t *raw_arg = slot_prop[slot].raw_arg;
|
|
uint32_t *arg = slot_prop[slot].arg;
|
|
XtensaOpcodeOps *ops;
|
|
|
|
dc->raw_arg = raw_arg;
|
|
|
|
xtensa_format_get_slot(isa, fmt, slot, dc->insnbuf, dc->slotbuf);
|
|
opc = xtensa_opcode_decode(isa, fmt, slot, dc->slotbuf);
|
|
if (opc == XTENSA_UNDEFINED) {
|
|
qemu_log_mask(LOG_GUEST_ERROR,
|
|
"unrecognized opcode in slot %d (pc = %08x)\n",
|
|
slot, dc->pc);
|
|
gen_exception_cause(dc, ILLEGAL_INSTRUCTION_CAUSE);
|
|
return;
|
|
}
|
|
opnds = xtensa_opcode_num_operands(isa, opc);
|
|
|
|
for (opnd = vopnd = 0; opnd < opnds; ++opnd) {
|
|
if (xtensa_operand_is_visible(isa, opc, opnd)) {
|
|
uint32_t v;
|
|
|
|
xtensa_operand_get_field(isa, opc, opnd, fmt, slot,
|
|
dc->slotbuf, &v);
|
|
xtensa_operand_decode(isa, opc, opnd, &v);
|
|
raw_arg[vopnd] = v;
|
|
if (xtensa_operand_is_PCrelative(isa, opc, opnd)) {
|
|
xtensa_operand_undo_reloc(isa, opc, opnd, &v, dc->pc);
|
|
}
|
|
arg[vopnd] = v;
|
|
++vopnd;
|
|
}
|
|
}
|
|
ops = dc->config->opcode_ops[opc];
|
|
slot_prop[slot].ops = ops;
|
|
|
|
if (ops) {
|
|
op_flags |= ops->op_flags;
|
|
} else {
|
|
qemu_log_mask(LOG_UNIMP,
|
|
"unimplemented opcode '%s' in slot %d (pc = %08x)\n",
|
|
xtensa_opcode_name(isa, opc), slot, dc->pc);
|
|
op_flags |= XTENSA_OP_ILL;
|
|
}
|
|
if ((op_flags & XTENSA_OP_ILL) ||
|
|
(ops && ops->test_ill && ops->test_ill(dc, arg, ops->par))) {
|
|
gen_exception_cause(dc, ILLEGAL_INSTRUCTION_CAUSE);
|
|
return;
|
|
}
|
|
if (ops->op_flags & XTENSA_OP_DEBUG_BREAK) {
|
|
debug_cause |= ops->par[0];
|
|
}
|
|
if (ops->test_overflow) {
|
|
windowed_register |= ops->test_overflow(dc, arg, ops->par);
|
|
}
|
|
if (ops->windowed_register_op) {
|
|
uint32_t reg_opnd = ops->windowed_register_op;
|
|
|
|
while (reg_opnd) {
|
|
unsigned i = ctz32(reg_opnd);
|
|
|
|
windowed_register |= 1 << arg[i];
|
|
reg_opnd ^= 1 << i;
|
|
}
|
|
}
|
|
coprocessor |= ops->coprocessor;
|
|
}
|
|
|
|
if ((op_flags & XTENSA_OP_PRIVILEGED) &&
|
|
!gen_check_privilege(dc)) {
|
|
return;
|
|
}
|
|
|
|
if (op_flags & XTENSA_OP_SYSCALL) {
|
|
gen_exception_cause(dc, SYSCALL_CAUSE);
|
|
return;
|
|
}
|
|
|
|
if ((op_flags & XTENSA_OP_DEBUG_BREAK) && dc->debug) {
|
|
gen_debug_exception(dc, debug_cause);
|
|
return;
|
|
}
|
|
|
|
if (windowed_register && !gen_window_check(dc, windowed_register)) {
|
|
return;
|
|
}
|
|
|
|
if (op_flags & XTENSA_OP_UNDERFLOW) {
|
|
TCGv_i32 tmp = tcg_const_i32(dc->pc);
|
|
|
|
gen_helper_test_underflow_retw(cpu_env, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
if (op_flags & XTENSA_OP_ALLOCA) {
|
|
TCGv_i32 tmp = tcg_const_i32(dc->pc);
|
|
|
|
gen_helper_movsp(cpu_env, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
if (coprocessor && !gen_check_cpenable(dc, coprocessor)) {
|
|
return;
|
|
}
|
|
|
|
if (op_flags & XTENSA_OP_DIVIDE_BY_ZERO) {
|
|
for (slot = 0; slot < slots; ++slot) {
|
|
if (slot_prop[slot].ops->op_flags & XTENSA_OP_DIVIDE_BY_ZERO) {
|
|
gen_zero_check(dc, slot_prop[slot].arg);
|
|
}
|
|
}
|
|
}
|
|
|
|
for (slot = 0; slot < slots; ++slot) {
|
|
XtensaOpcodeOps *ops = slot_prop[slot].ops;
|
|
|
|
dc->raw_arg = slot_prop[slot].raw_arg;
|
|
ops->translate(dc, slot_prop[slot].arg, ops->par);
|
|
}
|
|
|
|
if (dc->base.is_jmp == DISAS_NEXT) {
|
|
if (op_flags & XTENSA_OP_CHECK_INTERRUPTS) {
|
|
gen_check_interrupts(dc);
|
|
}
|
|
|
|
if (op_flags & XTENSA_OP_EXIT_TB_M1) {
|
|
/* Change in mmu index, memory mapping or tb->flags; exit tb */
|
|
gen_jumpi_check_loop_end(dc, -1);
|
|
} else if (op_flags & XTENSA_OP_EXIT_TB_0) {
|
|
gen_jumpi_check_loop_end(dc, 0);
|
|
}
|
|
}
|
|
|
|
if (dc->base.is_jmp == DISAS_NEXT) {
|
|
gen_check_loop_end(dc, 0);
|
|
}
|
|
dc->pc = dc->base.pc_next;
|
|
}
|
|
|
|
static inline unsigned xtensa_insn_len(CPUXtensaState *env, DisasContext *dc)
|
|
{
|
|
uint8_t b0 = cpu_ldub_code(env, dc->pc);
|
|
return xtensa_op0_insn_len(dc, b0);
|
|
}
|
|
|
|
static void gen_ibreak_check(CPUXtensaState *env, DisasContext *dc)
|
|
{
|
|
unsigned i;
|
|
|
|
for (i = 0; i < dc->config->nibreak; ++i) {
|
|
if ((env->sregs[IBREAKENABLE] & (1 << i)) &&
|
|
env->sregs[IBREAKA + i] == dc->pc) {
|
|
gen_debug_exception(dc, DEBUGCAUSE_IB);
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
static void xtensa_tr_init_disas_context(DisasContextBase *dcbase,
|
|
CPUState *cpu)
|
|
{
|
|
DisasContext *dc = container_of(dcbase, DisasContext, base);
|
|
CPUXtensaState *env = cpu->env_ptr;
|
|
uint32_t tb_flags = dc->base.tb->flags;
|
|
|
|
dc->config = env->config;
|
|
dc->pc = dc->base.pc_first;
|
|
dc->ring = tb_flags & XTENSA_TBFLAG_RING_MASK;
|
|
dc->cring = (tb_flags & XTENSA_TBFLAG_EXCM) ? 0 : dc->ring;
|
|
dc->lbeg_off = (dc->base.tb->cs_base & XTENSA_CSBASE_LBEG_OFF_MASK) >>
|
|
XTENSA_CSBASE_LBEG_OFF_SHIFT;
|
|
dc->lend = (dc->base.tb->cs_base & XTENSA_CSBASE_LEND_MASK) +
|
|
(dc->base.pc_first & TARGET_PAGE_MASK);
|
|
dc->debug = tb_flags & XTENSA_TBFLAG_DEBUG;
|
|
dc->icount = tb_flags & XTENSA_TBFLAG_ICOUNT;
|
|
dc->cpenable = (tb_flags & XTENSA_TBFLAG_CPENABLE_MASK) >>
|
|
XTENSA_TBFLAG_CPENABLE_SHIFT;
|
|
dc->window = ((tb_flags & XTENSA_TBFLAG_WINDOW_MASK) >>
|
|
XTENSA_TBFLAG_WINDOW_SHIFT);
|
|
dc->cwoe = tb_flags & XTENSA_TBFLAG_CWOE;
|
|
dc->callinc = ((tb_flags & XTENSA_TBFLAG_CALLINC_MASK) >>
|
|
XTENSA_TBFLAG_CALLINC_SHIFT);
|
|
|
|
if (dc->config->isa) {
|
|
dc->insnbuf = xtensa_insnbuf_alloc(dc->config->isa);
|
|
dc->slotbuf = xtensa_insnbuf_alloc(dc->config->isa);
|
|
}
|
|
init_sar_tracker(dc);
|
|
}
|
|
|
|
static void xtensa_tr_tb_start(DisasContextBase *dcbase, CPUState *cpu)
|
|
{
|
|
DisasContext *dc = container_of(dcbase, DisasContext, base);
|
|
|
|
if (dc->icount) {
|
|
dc->next_icount = tcg_temp_local_new_i32();
|
|
}
|
|
}
|
|
|
|
static void xtensa_tr_insn_start(DisasContextBase *dcbase, CPUState *cpu)
|
|
{
|
|
tcg_gen_insn_start(dcbase->pc_next);
|
|
}
|
|
|
|
static bool xtensa_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cpu,
|
|
const CPUBreakpoint *bp)
|
|
{
|
|
DisasContext *dc = container_of(dcbase, DisasContext, base);
|
|
|
|
tcg_gen_movi_i32(cpu_pc, dc->base.pc_next);
|
|
gen_exception(dc, EXCP_DEBUG);
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
/* The address covered by the breakpoint must be included in
|
|
[tb->pc, tb->pc + tb->size) in order to for it to be
|
|
properly cleared -- thus we increment the PC here so that
|
|
the logic setting tb->size below does the right thing. */
|
|
dc->base.pc_next += 2;
|
|
return true;
|
|
}
|
|
|
|
static void xtensa_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu)
|
|
{
|
|
DisasContext *dc = container_of(dcbase, DisasContext, base);
|
|
CPUXtensaState *env = cpu->env_ptr;
|
|
target_ulong page_start;
|
|
|
|
/* These two conditions only apply to the first insn in the TB,
|
|
but this is the first TranslateOps hook that allows exiting. */
|
|
if ((tb_cflags(dc->base.tb) & CF_USE_ICOUNT)
|
|
&& (dc->base.tb->flags & XTENSA_TBFLAG_YIELD)) {
|
|
gen_exception(dc, EXCP_YIELD);
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
return;
|
|
}
|
|
if (dc->base.tb->flags & XTENSA_TBFLAG_EXCEPTION) {
|
|
gen_exception(dc, EXCP_DEBUG);
|
|
dc->base.is_jmp = DISAS_NORETURN;
|
|
return;
|
|
}
|
|
|
|
if (dc->icount) {
|
|
TCGLabel *label = gen_new_label();
|
|
|
|
tcg_gen_addi_i32(dc->next_icount, cpu_SR[ICOUNT], 1);
|
|
tcg_gen_brcondi_i32(TCG_COND_NE, dc->next_icount, 0, label);
|
|
tcg_gen_mov_i32(dc->next_icount, cpu_SR[ICOUNT]);
|
|
if (dc->debug) {
|
|
gen_debug_exception(dc, DEBUGCAUSE_IC);
|
|
}
|
|
gen_set_label(label);
|
|
}
|
|
|
|
if (dc->debug) {
|
|
gen_ibreak_check(env, dc);
|
|
}
|
|
|
|
disas_xtensa_insn(env, dc);
|
|
|
|
if (dc->icount) {
|
|
tcg_gen_mov_i32(cpu_SR[ICOUNT], dc->next_icount);
|
|
}
|
|
|
|
/* End the TB if the next insn will cross into the next page. */
|
|
page_start = dc->base.pc_first & TARGET_PAGE_MASK;
|
|
if (dc->base.is_jmp == DISAS_NEXT &&
|
|
(dc->pc - page_start >= TARGET_PAGE_SIZE ||
|
|
dc->pc - page_start + xtensa_insn_len(env, dc) > TARGET_PAGE_SIZE)) {
|
|
dc->base.is_jmp = DISAS_TOO_MANY;
|
|
}
|
|
}
|
|
|
|
static void xtensa_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu)
|
|
{
|
|
DisasContext *dc = container_of(dcbase, DisasContext, base);
|
|
|
|
reset_sar_tracker(dc);
|
|
if (dc->config->isa) {
|
|
xtensa_insnbuf_free(dc->config->isa, dc->insnbuf);
|
|
xtensa_insnbuf_free(dc->config->isa, dc->slotbuf);
|
|
}
|
|
if (dc->icount) {
|
|
tcg_temp_free(dc->next_icount);
|
|
}
|
|
|
|
switch (dc->base.is_jmp) {
|
|
case DISAS_NORETURN:
|
|
break;
|
|
case DISAS_TOO_MANY:
|
|
if (dc->base.singlestep_enabled) {
|
|
tcg_gen_movi_i32(cpu_pc, dc->pc);
|
|
gen_exception(dc, EXCP_DEBUG);
|
|
} else {
|
|
gen_jumpi(dc, dc->pc, 0);
|
|
}
|
|
break;
|
|
default:
|
|
g_assert_not_reached();
|
|
}
|
|
}
|
|
|
|
static void xtensa_tr_disas_log(const DisasContextBase *dcbase, CPUState *cpu)
|
|
{
|
|
qemu_log("IN: %s\n", lookup_symbol(dcbase->pc_first));
|
|
log_target_disas(cpu, dcbase->pc_first, dcbase->tb->size);
|
|
}
|
|
|
|
static const TranslatorOps xtensa_translator_ops = {
|
|
.init_disas_context = xtensa_tr_init_disas_context,
|
|
.tb_start = xtensa_tr_tb_start,
|
|
.insn_start = xtensa_tr_insn_start,
|
|
.breakpoint_check = xtensa_tr_breakpoint_check,
|
|
.translate_insn = xtensa_tr_translate_insn,
|
|
.tb_stop = xtensa_tr_tb_stop,
|
|
.disas_log = xtensa_tr_disas_log,
|
|
};
|
|
|
|
void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb)
|
|
{
|
|
DisasContext dc = {};
|
|
translator_loop(&xtensa_translator_ops, &dc.base, cpu, tb);
|
|
}
|
|
|
|
void xtensa_cpu_dump_state(CPUState *cs, FILE *f,
|
|
fprintf_function cpu_fprintf, int flags)
|
|
{
|
|
XtensaCPU *cpu = XTENSA_CPU(cs);
|
|
CPUXtensaState *env = &cpu->env;
|
|
int i, j;
|
|
|
|
cpu_fprintf(f, "PC=%08x\n\n", env->pc);
|
|
|
|
for (i = j = 0; i < 256; ++i) {
|
|
if (xtensa_option_bits_enabled(env->config, sregnames[i].opt_bits)) {
|
|
cpu_fprintf(f, "%12s=%08x%c", sregnames[i].name, env->sregs[i],
|
|
(j++ % 4) == 3 ? '\n' : ' ');
|
|
}
|
|
}
|
|
|
|
cpu_fprintf(f, (j % 4) == 0 ? "\n" : "\n\n");
|
|
|
|
for (i = j = 0; i < 256; ++i) {
|
|
if (xtensa_option_bits_enabled(env->config, uregnames[i].opt_bits)) {
|
|
cpu_fprintf(f, "%s=%08x%c", uregnames[i].name, env->uregs[i],
|
|
(j++ % 4) == 3 ? '\n' : ' ');
|
|
}
|
|
}
|
|
|
|
cpu_fprintf(f, (j % 4) == 0 ? "\n" : "\n\n");
|
|
|
|
for (i = 0; i < 16; ++i) {
|
|
cpu_fprintf(f, " A%02d=%08x%c", i, env->regs[i],
|
|
(i % 4) == 3 ? '\n' : ' ');
|
|
}
|
|
|
|
xtensa_sync_phys_from_window(env);
|
|
cpu_fprintf(f, "\n");
|
|
|
|
for (i = 0; i < env->config->nareg; ++i) {
|
|
cpu_fprintf(f, "AR%02d=%08x ", i, env->phys_regs[i]);
|
|
if (i % 4 == 3) {
|
|
bool ws = (env->sregs[WINDOW_START] & (1 << (i / 4))) != 0;
|
|
bool cw = env->sregs[WINDOW_BASE] == i / 4;
|
|
|
|
cpu_fprintf(f, "%c%c\n", ws ? '<' : ' ', cw ? '=' : ' ');
|
|
}
|
|
}
|
|
|
|
if ((flags & CPU_DUMP_FPU) &&
|
|
xtensa_option_enabled(env->config, XTENSA_OPTION_FP_COPROCESSOR)) {
|
|
cpu_fprintf(f, "\n");
|
|
|
|
for (i = 0; i < 16; ++i) {
|
|
cpu_fprintf(f, "F%02d=%08x (%+10.8e)%c", i,
|
|
float32_val(env->fregs[i].f32[FP_F32_LOW]),
|
|
*(float *)(env->fregs[i].f32 + FP_F32_LOW),
|
|
(i % 2) == 1 ? '\n' : ' ');
|
|
}
|
|
}
|
|
}
|
|
|
|
void restore_state_to_opc(CPUXtensaState *env, TranslationBlock *tb,
|
|
target_ulong *data)
|
|
{
|
|
env->pc = data[0];
|
|
}
|
|
|
|
static int compare_opcode_ops(const void *a, const void *b)
|
|
{
|
|
return strcmp((const char *)a,
|
|
((const XtensaOpcodeOps *)b)->name);
|
|
}
|
|
|
|
XtensaOpcodeOps *
|
|
xtensa_find_opcode_ops(const XtensaOpcodeTranslators *t,
|
|
const char *name)
|
|
{
|
|
return bsearch(name, t->opcode, t->num_opcodes,
|
|
sizeof(XtensaOpcodeOps), compare_opcode_ops);
|
|
}
|
|
|
|
static void translate_abs(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 zero = tcg_const_i32(0);
|
|
TCGv_i32 neg = tcg_temp_new_i32();
|
|
|
|
tcg_gen_neg_i32(neg, cpu_R[arg[1]]);
|
|
tcg_gen_movcond_i32(TCG_COND_GE, cpu_R[arg[0]],
|
|
cpu_R[arg[1]], zero, cpu_R[arg[1]], neg);
|
|
tcg_temp_free(neg);
|
|
tcg_temp_free(zero);
|
|
}
|
|
|
|
static void translate_add(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_add_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_addi(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_addi_i32(cpu_R[arg[0]], cpu_R[arg[1]], arg[2]);
|
|
}
|
|
|
|
static void translate_addx(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_shli_i32(tmp, cpu_R[arg[1]], par[0]);
|
|
tcg_gen_add_i32(cpu_R[arg[0]], tmp, cpu_R[arg[2]]);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_all(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
uint32_t shift = par[1];
|
|
TCGv_i32 mask = tcg_const_i32(((1 << shift) - 1) << arg[1]);
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
|
|
tcg_gen_and_i32(tmp, cpu_SR[BR], mask);
|
|
if (par[0]) {
|
|
tcg_gen_addi_i32(tmp, tmp, 1 << arg[1]);
|
|
} else {
|
|
tcg_gen_add_i32(tmp, tmp, mask);
|
|
}
|
|
tcg_gen_shri_i32(tmp, tmp, arg[1] + shift);
|
|
tcg_gen_deposit_i32(cpu_SR[BR], cpu_SR[BR],
|
|
tmp, arg[0], 1);
|
|
tcg_temp_free(mask);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_and(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_and_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_ball(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_and_i32(tmp, cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
gen_brcond(dc, par[0], tmp, cpu_R[arg[1]], arg[2]);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_bany(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_and_i32(tmp, cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
gen_brcondi(dc, par[0], tmp, 0, arg[2]);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_b(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_brcond(dc, par[0], cpu_R[arg[0]], cpu_R[arg[1]], arg[2]);
|
|
}
|
|
|
|
static void translate_bb(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifdef TARGET_WORDS_BIGENDIAN
|
|
TCGv_i32 bit = tcg_const_i32(0x80000000u);
|
|
#else
|
|
TCGv_i32 bit = tcg_const_i32(0x00000001u);
|
|
#endif
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_andi_i32(tmp, cpu_R[arg[1]], 0x1f);
|
|
#ifdef TARGET_WORDS_BIGENDIAN
|
|
tcg_gen_shr_i32(bit, bit, tmp);
|
|
#else
|
|
tcg_gen_shl_i32(bit, bit, tmp);
|
|
#endif
|
|
tcg_gen_and_i32(tmp, cpu_R[arg[0]], bit);
|
|
gen_brcondi(dc, par[0], tmp, 0, arg[2]);
|
|
tcg_temp_free(tmp);
|
|
tcg_temp_free(bit);
|
|
}
|
|
|
|
static void translate_bbi(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
#ifdef TARGET_WORDS_BIGENDIAN
|
|
tcg_gen_andi_i32(tmp, cpu_R[arg[0]], 0x80000000u >> arg[1]);
|
|
#else
|
|
tcg_gen_andi_i32(tmp, cpu_R[arg[0]], 0x00000001u << arg[1]);
|
|
#endif
|
|
gen_brcondi(dc, par[0], tmp, 0, arg[2]);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_bi(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_brcondi(dc, par[0], cpu_R[arg[0]], arg[1], arg[2]);
|
|
}
|
|
|
|
static void translate_bz(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_brcondi(dc, par[0], cpu_R[arg[0]], 0, arg[1]);
|
|
}
|
|
|
|
enum {
|
|
BOOLEAN_AND,
|
|
BOOLEAN_ANDC,
|
|
BOOLEAN_OR,
|
|
BOOLEAN_ORC,
|
|
BOOLEAN_XOR,
|
|
};
|
|
|
|
static void translate_boolean(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
static void (* const op[])(TCGv_i32, TCGv_i32, TCGv_i32) = {
|
|
[BOOLEAN_AND] = tcg_gen_and_i32,
|
|
[BOOLEAN_ANDC] = tcg_gen_andc_i32,
|
|
[BOOLEAN_OR] = tcg_gen_or_i32,
|
|
[BOOLEAN_ORC] = tcg_gen_orc_i32,
|
|
[BOOLEAN_XOR] = tcg_gen_xor_i32,
|
|
};
|
|
|
|
TCGv_i32 tmp1 = tcg_temp_new_i32();
|
|
TCGv_i32 tmp2 = tcg_temp_new_i32();
|
|
|
|
tcg_gen_shri_i32(tmp1, cpu_SR[BR], arg[1]);
|
|
tcg_gen_shri_i32(tmp2, cpu_SR[BR], arg[2]);
|
|
op[par[0]](tmp1, tmp1, tmp2);
|
|
tcg_gen_deposit_i32(cpu_SR[BR], cpu_SR[BR], tmp1, arg[0], 1);
|
|
tcg_temp_free(tmp1);
|
|
tcg_temp_free(tmp2);
|
|
}
|
|
|
|
static void translate_bp(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
|
|
tcg_gen_andi_i32(tmp, cpu_SR[BR], 1 << arg[0]);
|
|
gen_brcondi(dc, par[0], tmp, 0, arg[1]);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_call0(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_movi_i32(cpu_R[0], dc->base.pc_next);
|
|
gen_jumpi(dc, arg[0], 0);
|
|
}
|
|
|
|
static uint32_t test_overflow_callw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
return 1 << (par[0] * 4);
|
|
}
|
|
|
|
static void translate_callw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_callwi(dc, par[0], arg[0], 0);
|
|
}
|
|
|
|
static void translate_callx0(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_mov_i32(tmp, cpu_R[arg[0]]);
|
|
tcg_gen_movi_i32(cpu_R[0], dc->base.pc_next);
|
|
gen_jump(dc, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_callxw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
|
|
tcg_gen_mov_i32(tmp, cpu_R[arg[0]]);
|
|
gen_callw(dc, par[0], tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_clamps(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp1 = tcg_const_i32(-1u << arg[2]);
|
|
TCGv_i32 tmp2 = tcg_const_i32((1 << arg[2]) - 1);
|
|
|
|
tcg_gen_smax_i32(tmp1, tmp1, cpu_R[arg[1]]);
|
|
tcg_gen_smin_i32(cpu_R[arg[0]], tmp1, tmp2);
|
|
tcg_temp_free(tmp1);
|
|
tcg_temp_free(tmp2);
|
|
}
|
|
|
|
static void translate_clrb_expstate(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
/* TODO: GPIO32 may be a part of coprocessor */
|
|
tcg_gen_andi_i32(cpu_UR[EXPSTATE], cpu_UR[EXPSTATE], ~(1u << arg[0]));
|
|
}
|
|
|
|
static void translate_const16(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 c = tcg_const_i32(arg[1]);
|
|
|
|
tcg_gen_deposit_i32(cpu_R[arg[0]], c, cpu_R[arg[0]], 16, 16);
|
|
tcg_temp_free(c);
|
|
}
|
|
|
|
static void translate_dcache(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 addr = tcg_temp_new_i32();
|
|
TCGv_i32 res = tcg_temp_new_i32();
|
|
|
|
tcg_gen_addi_i32(addr, cpu_R[arg[0]], arg[1]);
|
|
tcg_gen_qemu_ld8u(res, addr, dc->cring);
|
|
tcg_temp_free(addr);
|
|
tcg_temp_free(res);
|
|
}
|
|
|
|
static void translate_depbits(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_deposit_i32(cpu_R[arg[1]], cpu_R[arg[1]], cpu_R[arg[0]],
|
|
arg[2], arg[3]);
|
|
}
|
|
|
|
static bool test_ill_entry(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (arg[0] > 3 || !dc->cwoe) {
|
|
qemu_log_mask(LOG_GUEST_ERROR,
|
|
"Illegal entry instruction(pc = %08x)\n", dc->pc);
|
|
return true;
|
|
} else {
|
|
return false;
|
|
}
|
|
}
|
|
|
|
static uint32_t test_overflow_entry(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
return 1 << (dc->callinc * 4);
|
|
}
|
|
|
|
static void translate_entry(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 pc = tcg_const_i32(dc->pc);
|
|
TCGv_i32 s = tcg_const_i32(arg[0]);
|
|
TCGv_i32 imm = tcg_const_i32(arg[1]);
|
|
gen_helper_entry(cpu_env, pc, s, imm);
|
|
tcg_temp_free(imm);
|
|
tcg_temp_free(s);
|
|
tcg_temp_free(pc);
|
|
}
|
|
|
|
static void translate_extui(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
int maskimm = (1 << arg[3]) - 1;
|
|
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_shri_i32(tmp, cpu_R[arg[1]], arg[2]);
|
|
tcg_gen_andi_i32(cpu_R[arg[0]], tmp, maskimm);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_icache(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
TCGv_i32 addr = tcg_temp_new_i32();
|
|
|
|
tcg_gen_movi_i32(cpu_pc, dc->pc);
|
|
tcg_gen_addi_i32(addr, cpu_R[arg[0]], arg[1]);
|
|
gen_helper_itlb_hit_test(cpu_env, addr);
|
|
tcg_temp_free(addr);
|
|
#endif
|
|
}
|
|
|
|
static void translate_itlb(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
TCGv_i32 dtlb = tcg_const_i32(par[0]);
|
|
|
|
gen_helper_itlb(cpu_env, cpu_R[arg[0]], dtlb);
|
|
tcg_temp_free(dtlb);
|
|
#endif
|
|
}
|
|
|
|
static void translate_j(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_jumpi(dc, arg[0], 0);
|
|
}
|
|
|
|
static void translate_jx(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_jump(dc, cpu_R[arg[0]]);
|
|
}
|
|
|
|
static void translate_l32e(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 addr = tcg_temp_new_i32();
|
|
|
|
tcg_gen_addi_i32(addr, cpu_R[arg[1]], arg[2]);
|
|
gen_load_store_alignment(dc, 2, addr, false);
|
|
tcg_gen_qemu_ld_tl(cpu_R[arg[0]], addr, dc->ring, MO_TEUL);
|
|
tcg_temp_free(addr);
|
|
}
|
|
|
|
static void translate_ldst(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 addr = tcg_temp_new_i32();
|
|
|
|
tcg_gen_addi_i32(addr, cpu_R[arg[1]], arg[2]);
|
|
if (par[0] & MO_SIZE) {
|
|
gen_load_store_alignment(dc, par[0] & MO_SIZE, addr, par[1]);
|
|
}
|
|
if (par[2]) {
|
|
if (par[1]) {
|
|
tcg_gen_mb(TCG_BAR_STRL | TCG_MO_ALL);
|
|
}
|
|
tcg_gen_qemu_st_tl(cpu_R[arg[0]], addr, dc->cring, par[0]);
|
|
} else {
|
|
tcg_gen_qemu_ld_tl(cpu_R[arg[0]], addr, dc->cring, par[0]);
|
|
if (par[1]) {
|
|
tcg_gen_mb(TCG_BAR_LDAQ | TCG_MO_ALL);
|
|
}
|
|
}
|
|
tcg_temp_free(addr);
|
|
}
|
|
|
|
static void translate_l32r(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp;
|
|
|
|
if (dc->base.tb->flags & XTENSA_TBFLAG_LITBASE) {
|
|
tmp = tcg_const_i32(dc->raw_arg[1] - 1);
|
|
tcg_gen_add_i32(tmp, cpu_SR[LITBASE], tmp);
|
|
} else {
|
|
tmp = tcg_const_i32(arg[1]);
|
|
}
|
|
tcg_gen_qemu_ld32u(cpu_R[arg[0]], tmp, dc->cring);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_loop(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
uint32_t lend = arg[1];
|
|
|
|
tcg_gen_subi_i32(cpu_SR[LCOUNT], cpu_R[arg[0]], 1);
|
|
tcg_gen_movi_i32(cpu_SR[LBEG], dc->base.pc_next);
|
|
tcg_gen_movi_i32(cpu_SR[LEND], lend);
|
|
|
|
if (par[0] != TCG_COND_NEVER) {
|
|
TCGLabel *label = gen_new_label();
|
|
tcg_gen_brcondi_i32(par[0], cpu_R[arg[0]], 0, label);
|
|
gen_jumpi(dc, lend, 1);
|
|
gen_set_label(label);
|
|
}
|
|
|
|
gen_jumpi(dc, dc->base.pc_next, 0);
|
|
}
|
|
|
|
enum {
|
|
MAC16_UMUL,
|
|
MAC16_MUL,
|
|
MAC16_MULA,
|
|
MAC16_MULS,
|
|
MAC16_NONE,
|
|
};
|
|
|
|
enum {
|
|
MAC16_LL,
|
|
MAC16_HL,
|
|
MAC16_LH,
|
|
MAC16_HH,
|
|
|
|
MAC16_HX = 0x1,
|
|
MAC16_XH = 0x2,
|
|
};
|
|
|
|
enum {
|
|
MAC16_AA,
|
|
MAC16_AD,
|
|
MAC16_DA,
|
|
MAC16_DD,
|
|
|
|
MAC16_XD = 0x1,
|
|
MAC16_DX = 0x2,
|
|
};
|
|
|
|
static void translate_mac16(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
int op = par[0];
|
|
bool is_m1_sr = par[1] & MAC16_DX;
|
|
bool is_m2_sr = par[1] & MAC16_XD;
|
|
unsigned half = par[2];
|
|
uint32_t ld_offset = par[3];
|
|
unsigned off = ld_offset ? 2 : 0;
|
|
TCGv_i32 vaddr = tcg_temp_new_i32();
|
|
TCGv_i32 mem32 = tcg_temp_new_i32();
|
|
|
|
if (ld_offset) {
|
|
tcg_gen_addi_i32(vaddr, cpu_R[arg[1]], ld_offset);
|
|
gen_load_store_alignment(dc, 2, vaddr, false);
|
|
tcg_gen_qemu_ld32u(mem32, vaddr, dc->cring);
|
|
}
|
|
if (op != MAC16_NONE) {
|
|
TCGv_i32 m1 = gen_mac16_m(is_m1_sr ?
|
|
cpu_SR[MR + arg[off]] :
|
|
cpu_R[arg[off]],
|
|
half & MAC16_HX, op == MAC16_UMUL);
|
|
TCGv_i32 m2 = gen_mac16_m(is_m2_sr ?
|
|
cpu_SR[MR + arg[off + 1]] :
|
|
cpu_R[arg[off + 1]],
|
|
half & MAC16_XH, op == MAC16_UMUL);
|
|
|
|
if (op == MAC16_MUL || op == MAC16_UMUL) {
|
|
tcg_gen_mul_i32(cpu_SR[ACCLO], m1, m2);
|
|
if (op == MAC16_UMUL) {
|
|
tcg_gen_movi_i32(cpu_SR[ACCHI], 0);
|
|
} else {
|
|
tcg_gen_sari_i32(cpu_SR[ACCHI], cpu_SR[ACCLO], 31);
|
|
}
|
|
} else {
|
|
TCGv_i32 lo = tcg_temp_new_i32();
|
|
TCGv_i32 hi = tcg_temp_new_i32();
|
|
|
|
tcg_gen_mul_i32(lo, m1, m2);
|
|
tcg_gen_sari_i32(hi, lo, 31);
|
|
if (op == MAC16_MULA) {
|
|
tcg_gen_add2_i32(cpu_SR[ACCLO], cpu_SR[ACCHI],
|
|
cpu_SR[ACCLO], cpu_SR[ACCHI],
|
|
lo, hi);
|
|
} else {
|
|
tcg_gen_sub2_i32(cpu_SR[ACCLO], cpu_SR[ACCHI],
|
|
cpu_SR[ACCLO], cpu_SR[ACCHI],
|
|
lo, hi);
|
|
}
|
|
tcg_gen_ext8s_i32(cpu_SR[ACCHI], cpu_SR[ACCHI]);
|
|
|
|
tcg_temp_free_i32(lo);
|
|
tcg_temp_free_i32(hi);
|
|
}
|
|
tcg_temp_free(m1);
|
|
tcg_temp_free(m2);
|
|
}
|
|
if (ld_offset) {
|
|
tcg_gen_mov_i32(cpu_R[arg[1]], vaddr);
|
|
tcg_gen_mov_i32(cpu_SR[MR + arg[0]], mem32);
|
|
}
|
|
tcg_temp_free(vaddr);
|
|
tcg_temp_free(mem32);
|
|
}
|
|
|
|
static void translate_memw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mb(TCG_BAR_SC | TCG_MO_ALL);
|
|
}
|
|
|
|
static void translate_smin(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_smin_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_umin(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_umin_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_smax(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_smax_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_umax(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_umax_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_mov(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mov_i32(cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
}
|
|
|
|
static void translate_movcond(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 zero = tcg_const_i32(0);
|
|
|
|
tcg_gen_movcond_i32(par[0], cpu_R[arg[0]],
|
|
cpu_R[arg[2]], zero, cpu_R[arg[1]], cpu_R[arg[0]]);
|
|
tcg_temp_free(zero);
|
|
}
|
|
|
|
static void translate_movi(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_movi_i32(cpu_R[arg[0]], arg[1]);
|
|
}
|
|
|
|
static void translate_movp(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 zero = tcg_const_i32(0);
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
|
|
tcg_gen_andi_i32(tmp, cpu_SR[BR], 1 << arg[2]);
|
|
tcg_gen_movcond_i32(par[0],
|
|
cpu_R[arg[0]], tmp, zero,
|
|
cpu_R[arg[1]], cpu_R[arg[0]]);
|
|
tcg_temp_free(tmp);
|
|
tcg_temp_free(zero);
|
|
}
|
|
|
|
static void translate_movsp(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mov_i32(cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
}
|
|
|
|
static void translate_mul16(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 v1 = tcg_temp_new_i32();
|
|
TCGv_i32 v2 = tcg_temp_new_i32();
|
|
|
|
if (par[0]) {
|
|
tcg_gen_ext16s_i32(v1, cpu_R[arg[1]]);
|
|
tcg_gen_ext16s_i32(v2, cpu_R[arg[2]]);
|
|
} else {
|
|
tcg_gen_ext16u_i32(v1, cpu_R[arg[1]]);
|
|
tcg_gen_ext16u_i32(v2, cpu_R[arg[2]]);
|
|
}
|
|
tcg_gen_mul_i32(cpu_R[arg[0]], v1, v2);
|
|
tcg_temp_free(v2);
|
|
tcg_temp_free(v1);
|
|
}
|
|
|
|
static void translate_mull(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mul_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_mulh(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 lo = tcg_temp_new();
|
|
|
|
if (par[0]) {
|
|
tcg_gen_muls2_i32(lo, cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
} else {
|
|
tcg_gen_mulu2_i32(lo, cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
tcg_temp_free(lo);
|
|
}
|
|
|
|
static void translate_neg(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_neg_i32(cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
}
|
|
|
|
static void translate_nop(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
}
|
|
|
|
static void translate_nsa(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_clrsb_i32(cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
}
|
|
|
|
static void translate_nsau(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_clzi_i32(cpu_R[arg[0]], cpu_R[arg[1]], 32);
|
|
}
|
|
|
|
static void translate_or(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_or_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_ptlb(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
TCGv_i32 dtlb = tcg_const_i32(par[0]);
|
|
|
|
tcg_gen_movi_i32(cpu_pc, dc->pc);
|
|
gen_helper_ptlb(cpu_R[arg[0]], cpu_env, cpu_R[arg[1]], dtlb);
|
|
tcg_temp_free(dtlb);
|
|
#endif
|
|
}
|
|
|
|
static void translate_quos(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGLabel *label1 = gen_new_label();
|
|
TCGLabel *label2 = gen_new_label();
|
|
|
|
tcg_gen_brcondi_i32(TCG_COND_NE, cpu_R[arg[1]], 0x80000000,
|
|
label1);
|
|
tcg_gen_brcondi_i32(TCG_COND_NE, cpu_R[arg[2]], 0xffffffff,
|
|
label1);
|
|
tcg_gen_movi_i32(cpu_R[arg[0]],
|
|
par[0] ? 0x80000000 : 0);
|
|
tcg_gen_br(label2);
|
|
gen_set_label(label1);
|
|
if (par[0]) {
|
|
tcg_gen_div_i32(cpu_R[arg[0]],
|
|
cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
} else {
|
|
tcg_gen_rem_i32(cpu_R[arg[0]],
|
|
cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
gen_set_label(label2);
|
|
}
|
|
|
|
static void translate_quou(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_divu_i32(cpu_R[arg[0]],
|
|
cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_read_impwire(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
/* TODO: GPIO32 may be a part of coprocessor */
|
|
tcg_gen_movi_i32(cpu_R[arg[0]], 0);
|
|
}
|
|
|
|
static void translate_remu(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_remu_i32(cpu_R[arg[0]],
|
|
cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_rer(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_rer(cpu_R[arg[0]], cpu_env, cpu_R[arg[1]]);
|
|
}
|
|
|
|
static void translate_ret(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_jump(dc, cpu_R[0]);
|
|
}
|
|
|
|
static bool test_ill_retw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (!dc->cwoe) {
|
|
qemu_log_mask(LOG_GUEST_ERROR,
|
|
"Illegal retw instruction(pc = %08x)\n", dc->pc);
|
|
return true;
|
|
} else {
|
|
TCGv_i32 tmp = tcg_const_i32(dc->pc);
|
|
|
|
gen_helper_test_ill_retw(cpu_env, tmp);
|
|
tcg_temp_free(tmp);
|
|
return false;
|
|
}
|
|
}
|
|
|
|
static void translate_retw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(dc->pc);
|
|
gen_helper_retw(tmp, cpu_env, tmp);
|
|
gen_jump(dc, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_rfde(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_jump(dc, cpu_SR[dc->config->ndepc ? DEPC : EPC1]);
|
|
}
|
|
|
|
static void translate_rfe(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_andi_i32(cpu_SR[PS], cpu_SR[PS], ~PS_EXCM);
|
|
gen_jump(dc, cpu_SR[EPC1]);
|
|
}
|
|
|
|
static void translate_rfi(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mov_i32(cpu_SR[PS], cpu_SR[EPS2 + arg[0] - 2]);
|
|
gen_jump(dc, cpu_SR[EPC1 + arg[0] - 1]);
|
|
}
|
|
|
|
static void translate_rfw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(1);
|
|
|
|
tcg_gen_andi_i32(cpu_SR[PS], cpu_SR[PS], ~PS_EXCM);
|
|
tcg_gen_shl_i32(tmp, tmp, cpu_SR[WINDOW_BASE]);
|
|
|
|
if (par[0]) {
|
|
tcg_gen_andc_i32(cpu_SR[WINDOW_START],
|
|
cpu_SR[WINDOW_START], tmp);
|
|
} else {
|
|
tcg_gen_or_i32(cpu_SR[WINDOW_START],
|
|
cpu_SR[WINDOW_START], tmp);
|
|
}
|
|
|
|
tcg_temp_free(tmp);
|
|
gen_helper_restore_owb(cpu_env);
|
|
gen_jump(dc, cpu_SR[EPC1]);
|
|
}
|
|
|
|
static void translate_rotw(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(arg[0]);
|
|
gen_helper_rotw(cpu_env, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_rsil(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mov_i32(cpu_R[arg[0]], cpu_SR[PS]);
|
|
tcg_gen_andi_i32(cpu_SR[PS], cpu_SR[PS], ~PS_INTLEVEL);
|
|
tcg_gen_ori_i32(cpu_SR[PS], cpu_SR[PS], arg[1]);
|
|
}
|
|
|
|
static bool test_ill_rsr(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
return !check_sr(dc, par[0], SR_R);
|
|
}
|
|
|
|
static void translate_rsr(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_rsr(dc, cpu_R[arg[0]], par[0]);
|
|
}
|
|
|
|
static void translate_rtlb(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
static void (* const helper[])(TCGv_i32 r, TCGv_env env, TCGv_i32 a1,
|
|
TCGv_i32 a2) = {
|
|
gen_helper_rtlb0,
|
|
gen_helper_rtlb1,
|
|
};
|
|
TCGv_i32 dtlb = tcg_const_i32(par[0]);
|
|
|
|
helper[par[1]](cpu_R[arg[0]], cpu_env, cpu_R[arg[1]], dtlb);
|
|
tcg_temp_free(dtlb);
|
|
#endif
|
|
}
|
|
|
|
static void translate_rur(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (uregnames[par[0]].name) {
|
|
tcg_gen_mov_i32(cpu_R[arg[0]], cpu_UR[par[0]]);
|
|
} else {
|
|
qemu_log_mask(LOG_UNIMP, "RUR %d not implemented\n", par[0]);
|
|
}
|
|
}
|
|
|
|
static void translate_setb_expstate(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
/* TODO: GPIO32 may be a part of coprocessor */
|
|
tcg_gen_ori_i32(cpu_UR[EXPSTATE], cpu_UR[EXPSTATE], 1u << arg[0]);
|
|
}
|
|
|
|
#ifdef CONFIG_USER_ONLY
|
|
static void gen_check_atomctl(DisasContext *dc, TCGv_i32 addr)
|
|
{
|
|
}
|
|
#else
|
|
static void gen_check_atomctl(DisasContext *dc, TCGv_i32 addr)
|
|
{
|
|
TCGv_i32 tpc = tcg_const_i32(dc->pc);
|
|
|
|
gen_helper_check_atomctl(cpu_env, tpc, addr);
|
|
tcg_temp_free(tpc);
|
|
}
|
|
#endif
|
|
|
|
static void translate_s32c1i(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_local_new_i32();
|
|
TCGv_i32 addr = tcg_temp_local_new_i32();
|
|
|
|
tcg_gen_mov_i32(tmp, cpu_R[arg[0]]);
|
|
tcg_gen_addi_i32(addr, cpu_R[arg[1]], arg[2]);
|
|
gen_load_store_alignment(dc, 2, addr, true);
|
|
gen_check_atomctl(dc, addr);
|
|
tcg_gen_atomic_cmpxchg_i32(cpu_R[arg[0]], addr, cpu_SR[SCOMPARE1],
|
|
tmp, dc->cring, MO_TEUL);
|
|
tcg_temp_free(addr);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_s32e(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 addr = tcg_temp_new_i32();
|
|
|
|
tcg_gen_addi_i32(addr, cpu_R[arg[1]], arg[2]);
|
|
gen_load_store_alignment(dc, 2, addr, false);
|
|
tcg_gen_qemu_st_tl(cpu_R[arg[0]], addr, dc->ring, MO_TEUL);
|
|
tcg_temp_free(addr);
|
|
}
|
|
|
|
static void translate_salt(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_setcond_i32(par[0],
|
|
cpu_R[arg[0]],
|
|
cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_sext(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
int shift = 31 - arg[2];
|
|
|
|
if (shift == 24) {
|
|
tcg_gen_ext8s_i32(cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
} else if (shift == 16) {
|
|
tcg_gen_ext16s_i32(cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
} else {
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_shli_i32(tmp, cpu_R[arg[1]], shift);
|
|
tcg_gen_sari_i32(cpu_R[arg[0]], tmp, shift);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
}
|
|
|
|
static bool test_ill_simcall(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifdef CONFIG_USER_ONLY
|
|
bool ill = true;
|
|
#else
|
|
bool ill = !semihosting_enabled();
|
|
#endif
|
|
if (ill) {
|
|
qemu_log_mask(LOG_GUEST_ERROR, "SIMCALL but semihosting is disabled\n");
|
|
}
|
|
return ill;
|
|
}
|
|
|
|
static void translate_simcall(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
gen_helper_simcall(cpu_env);
|
|
#endif
|
|
}
|
|
|
|
/*
|
|
* Note: 64 bit ops are used here solely because SAR values
|
|
* have range 0..63
|
|
*/
|
|
#define gen_shift_reg(cmd, reg) do { \
|
|
TCGv_i64 tmp = tcg_temp_new_i64(); \
|
|
tcg_gen_extu_i32_i64(tmp, reg); \
|
|
tcg_gen_##cmd##_i64(v, v, tmp); \
|
|
tcg_gen_extrl_i64_i32(cpu_R[arg[0]], v); \
|
|
tcg_temp_free_i64(v); \
|
|
tcg_temp_free_i64(tmp); \
|
|
} while (0)
|
|
|
|
#define gen_shift(cmd) gen_shift_reg(cmd, cpu_SR[SAR])
|
|
|
|
static void translate_sll(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (dc->sar_m32_5bit) {
|
|
tcg_gen_shl_i32(cpu_R[arg[0]], cpu_R[arg[1]], dc->sar_m32);
|
|
} else {
|
|
TCGv_i64 v = tcg_temp_new_i64();
|
|
TCGv_i32 s = tcg_const_i32(32);
|
|
tcg_gen_sub_i32(s, s, cpu_SR[SAR]);
|
|
tcg_gen_andi_i32(s, s, 0x3f);
|
|
tcg_gen_extu_i32_i64(v, cpu_R[arg[1]]);
|
|
gen_shift_reg(shl, s);
|
|
tcg_temp_free(s);
|
|
}
|
|
}
|
|
|
|
static void translate_slli(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (arg[2] == 32) {
|
|
qemu_log_mask(LOG_GUEST_ERROR, "slli a%d, a%d, 32 is undefined\n",
|
|
arg[0], arg[1]);
|
|
}
|
|
tcg_gen_shli_i32(cpu_R[arg[0]], cpu_R[arg[1]], arg[2] & 0x1f);
|
|
}
|
|
|
|
static void translate_sra(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (dc->sar_m32_5bit) {
|
|
tcg_gen_sar_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_SR[SAR]);
|
|
} else {
|
|
TCGv_i64 v = tcg_temp_new_i64();
|
|
tcg_gen_ext_i32_i64(v, cpu_R[arg[1]]);
|
|
gen_shift(sar);
|
|
}
|
|
}
|
|
|
|
static void translate_srai(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_sari_i32(cpu_R[arg[0]], cpu_R[arg[1]], arg[2]);
|
|
}
|
|
|
|
static void translate_src(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i64 v = tcg_temp_new_i64();
|
|
tcg_gen_concat_i32_i64(v, cpu_R[arg[2]], cpu_R[arg[1]]);
|
|
gen_shift(shr);
|
|
}
|
|
|
|
static void translate_srl(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (dc->sar_m32_5bit) {
|
|
tcg_gen_shr_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_SR[SAR]);
|
|
} else {
|
|
TCGv_i64 v = tcg_temp_new_i64();
|
|
tcg_gen_extu_i32_i64(v, cpu_R[arg[1]]);
|
|
gen_shift(shr);
|
|
}
|
|
}
|
|
|
|
#undef gen_shift
|
|
#undef gen_shift_reg
|
|
|
|
static void translate_srli(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_shri_i32(cpu_R[arg[0]], cpu_R[arg[1]], arg[2]);
|
|
}
|
|
|
|
static void translate_ssa8b(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_shli_i32(tmp, cpu_R[arg[0]], 3);
|
|
gen_left_shift_sar(dc, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_ssa8l(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_shli_i32(tmp, cpu_R[arg[0]], 3);
|
|
gen_right_shift_sar(dc, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_ssai(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_const_i32(arg[0]);
|
|
gen_right_shift_sar(dc, tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_ssl(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_left_shift_sar(dc, cpu_R[arg[0]]);
|
|
}
|
|
|
|
static void translate_ssr(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_right_shift_sar(dc, cpu_R[arg[0]]);
|
|
}
|
|
|
|
static void translate_sub(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_sub_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static void translate_subx(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
tcg_gen_shli_i32(tmp, cpu_R[arg[1]], par[0]);
|
|
tcg_gen_sub_i32(cpu_R[arg[0]], tmp, cpu_R[arg[2]]);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static void translate_waiti(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
gen_waiti(dc, arg[0]);
|
|
#endif
|
|
}
|
|
|
|
static void translate_wtlb(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
#ifndef CONFIG_USER_ONLY
|
|
TCGv_i32 dtlb = tcg_const_i32(par[0]);
|
|
|
|
gen_helper_wtlb(cpu_env, cpu_R[arg[0]], cpu_R[arg[1]], dtlb);
|
|
tcg_temp_free(dtlb);
|
|
#endif
|
|
}
|
|
|
|
static void translate_wer(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_wer(cpu_env, cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
}
|
|
|
|
static void translate_wrmsk_expstate(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
/* TODO: GPIO32 may be a part of coprocessor */
|
|
tcg_gen_and_i32(cpu_UR[EXPSTATE], cpu_R[arg[0]], cpu_R[arg[1]]);
|
|
}
|
|
|
|
static bool test_ill_wsr(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
return !check_sr(dc, par[0], SR_W);
|
|
}
|
|
|
|
static void translate_wsr(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_wsr(dc, par[0], cpu_R[arg[0]]);
|
|
}
|
|
|
|
static void translate_wur(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
if (uregnames[par[0]].name) {
|
|
gen_wur(par[0], cpu_R[arg[0]]);
|
|
} else {
|
|
qemu_log_mask(LOG_UNIMP, "WUR %d not implemented\n", par[0]);
|
|
}
|
|
}
|
|
|
|
static void translate_xor(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_xor_i32(cpu_R[arg[0]], cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
}
|
|
|
|
static bool test_ill_xsr(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
return !check_sr(dc, par[0], SR_X);
|
|
}
|
|
|
|
static void translate_xsr(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
|
|
tcg_gen_mov_i32(tmp, cpu_R[arg[0]]);
|
|
gen_rsr(dc, cpu_R[arg[0]], par[0]);
|
|
gen_wsr(dc, par[0], tmp);
|
|
tcg_temp_free(tmp);
|
|
}
|
|
|
|
static const XtensaOpcodeOps core_ops[] = {
|
|
{
|
|
.name = "abs",
|
|
.translate = translate_abs,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "add",
|
|
.translate = translate_add,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "add.n",
|
|
.translate = translate_add,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "addi",
|
|
.translate = translate_addi,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "addi.n",
|
|
.translate = translate_addi,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "addmi",
|
|
.translate = translate_addi,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "addx2",
|
|
.translate = translate_addx,
|
|
.par = (const uint32_t[]){1},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "addx4",
|
|
.translate = translate_addx,
|
|
.par = (const uint32_t[]){2},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "addx8",
|
|
.translate = translate_addx,
|
|
.par = (const uint32_t[]){3},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "all4",
|
|
.translate = translate_all,
|
|
.par = (const uint32_t[]){true, 4},
|
|
}, {
|
|
.name = "all8",
|
|
.translate = translate_all,
|
|
.par = (const uint32_t[]){true, 8},
|
|
}, {
|
|
.name = "and",
|
|
.translate = translate_and,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "andb",
|
|
.translate = translate_boolean,
|
|
.par = (const uint32_t[]){BOOLEAN_AND},
|
|
}, {
|
|
.name = "andbc",
|
|
.translate = translate_boolean,
|
|
.par = (const uint32_t[]){BOOLEAN_ANDC},
|
|
}, {
|
|
.name = "any4",
|
|
.translate = translate_all,
|
|
.par = (const uint32_t[]){false, 4},
|
|
}, {
|
|
.name = "any8",
|
|
.translate = translate_all,
|
|
.par = (const uint32_t[]){false, 8},
|
|
}, {
|
|
.name = "ball",
|
|
.translate = translate_ball,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bany",
|
|
.translate = translate_bany,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bbc",
|
|
.translate = translate_bb,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bbci",
|
|
.translate = translate_bbi,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bbs",
|
|
.translate = translate_bb,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bbsi",
|
|
.translate = translate_bbi,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "beq",
|
|
.translate = translate_b,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "beqi",
|
|
.translate = translate_bi,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "beqz",
|
|
.translate = translate_bz,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "beqz.n",
|
|
.translate = translate_bz,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bf",
|
|
.translate = translate_bp,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
}, {
|
|
.name = "bge",
|
|
.translate = translate_b,
|
|
.par = (const uint32_t[]){TCG_COND_GE},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bgei",
|
|
.translate = translate_bi,
|
|
.par = (const uint32_t[]){TCG_COND_GE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bgeu",
|
|
.translate = translate_b,
|
|
.par = (const uint32_t[]){TCG_COND_GEU},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bgeui",
|
|
.translate = translate_bi,
|
|
.par = (const uint32_t[]){TCG_COND_GEU},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bgez",
|
|
.translate = translate_bz,
|
|
.par = (const uint32_t[]){TCG_COND_GE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "blt",
|
|
.translate = translate_b,
|
|
.par = (const uint32_t[]){TCG_COND_LT},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "blti",
|
|
.translate = translate_bi,
|
|
.par = (const uint32_t[]){TCG_COND_LT},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bltu",
|
|
.translate = translate_b,
|
|
.par = (const uint32_t[]){TCG_COND_LTU},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bltui",
|
|
.translate = translate_bi,
|
|
.par = (const uint32_t[]){TCG_COND_LTU},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bltz",
|
|
.translate = translate_bz,
|
|
.par = (const uint32_t[]){TCG_COND_LT},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bnall",
|
|
.translate = translate_ball,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bne",
|
|
.translate = translate_b,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "bnei",
|
|
.translate = translate_bi,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bnez",
|
|
.translate = translate_bz,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bnez.n",
|
|
.translate = translate_bz,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "bnone",
|
|
.translate = translate_bany,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "break",
|
|
.translate = translate_nop,
|
|
.par = (const uint32_t[]){DEBUGCAUSE_BI},
|
|
.op_flags = XTENSA_OP_DEBUG_BREAK,
|
|
}, {
|
|
.name = "break.n",
|
|
.translate = translate_nop,
|
|
.par = (const uint32_t[]){DEBUGCAUSE_BN},
|
|
.op_flags = XTENSA_OP_DEBUG_BREAK,
|
|
}, {
|
|
.name = "bt",
|
|
.translate = translate_bp,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
}, {
|
|
.name = "call0",
|
|
.translate = translate_call0,
|
|
}, {
|
|
.name = "call12",
|
|
.translate = translate_callw,
|
|
.test_overflow = test_overflow_callw,
|
|
.par = (const uint32_t[]){3},
|
|
}, {
|
|
.name = "call4",
|
|
.translate = translate_callw,
|
|
.test_overflow = test_overflow_callw,
|
|
.par = (const uint32_t[]){1},
|
|
}, {
|
|
.name = "call8",
|
|
.translate = translate_callw,
|
|
.test_overflow = test_overflow_callw,
|
|
.par = (const uint32_t[]){2},
|
|
}, {
|
|
.name = "callx0",
|
|
.translate = translate_callx0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "callx12",
|
|
.translate = translate_callxw,
|
|
.test_overflow = test_overflow_callw,
|
|
.par = (const uint32_t[]){3},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "callx4",
|
|
.translate = translate_callxw,
|
|
.test_overflow = test_overflow_callw,
|
|
.par = (const uint32_t[]){1},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "callx8",
|
|
.translate = translate_callxw,
|
|
.test_overflow = test_overflow_callw,
|
|
.par = (const uint32_t[]){2},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "clamps",
|
|
.translate = translate_clamps,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "clrb_expstate",
|
|
.translate = translate_clrb_expstate,
|
|
}, {
|
|
.name = "const16",
|
|
.translate = translate_const16,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "depbits",
|
|
.translate = translate_depbits,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "dhi",
|
|
.translate = translate_dcache,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dhu",
|
|
.translate = translate_dcache,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dhwb",
|
|
.translate = translate_dcache,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dhwbi",
|
|
.translate = translate_dcache,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dii",
|
|
.translate = translate_nop,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "diu",
|
|
.translate = translate_nop,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "diwb",
|
|
.translate = translate_nop,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "diwbi",
|
|
.translate = translate_nop,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dpfl",
|
|
.translate = translate_dcache,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dpfr",
|
|
.translate = translate_nop,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dpfro",
|
|
.translate = translate_nop,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dpfw",
|
|
.translate = translate_nop,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dpfwo",
|
|
.translate = translate_nop,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "dsync",
|
|
.translate = translate_nop,
|
|
}, {
|
|
.name = "entry",
|
|
.translate = translate_entry,
|
|
.test_ill = test_ill_entry,
|
|
.test_overflow = test_overflow_entry,
|
|
.op_flags = XTENSA_OP_EXIT_TB_M1,
|
|
}, {
|
|
.name = "esync",
|
|
.translate = translate_nop,
|
|
}, {
|
|
.name = "excw",
|
|
.translate = translate_nop,
|
|
}, {
|
|
.name = "extui",
|
|
.translate = translate_extui,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "extw",
|
|
.translate = translate_memw,
|
|
}, {
|
|
.name = "hwwdtlba",
|
|
.op_flags = XTENSA_OP_ILL,
|
|
}, {
|
|
.name = "hwwitlba",
|
|
.op_flags = XTENSA_OP_ILL,
|
|
}, {
|
|
.name = "idtlb",
|
|
.translate = translate_itlb,
|
|
.par = (const uint32_t[]){true},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "ihi",
|
|
.translate = translate_icache,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "ihu",
|
|
.translate = translate_icache,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "iii",
|
|
.translate = translate_nop,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "iitlb",
|
|
.translate = translate_itlb,
|
|
.par = (const uint32_t[]){false},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "iiu",
|
|
.translate = translate_nop,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "ill",
|
|
.op_flags = XTENSA_OP_ILL,
|
|
}, {
|
|
.name = "ill.n",
|
|
.op_flags = XTENSA_OP_ILL,
|
|
}, {
|
|
.name = "ipf",
|
|
.translate = translate_nop,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "ipfl",
|
|
.translate = translate_icache,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "isync",
|
|
.translate = translate_nop,
|
|
}, {
|
|
.name = "j",
|
|
.translate = translate_j,
|
|
}, {
|
|
.name = "jx",
|
|
.translate = translate_jx,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "l16si",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TESW, false, false},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "l16ui",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUW, false, false},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "l32ai",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUL, true, false},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "l32e",
|
|
.translate = translate_l32e,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "l32i",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUL, false, false},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "l32i.n",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUL, false, false},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "l32r",
|
|
.translate = translate_l32r,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "l8ui",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_UB, false, false},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_NONE, 0, 0, -4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_NONE, 0, 0, 4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "ldpte",
|
|
.op_flags = XTENSA_OP_ILL,
|
|
}, {
|
|
.name = "loop",
|
|
.translate = translate_loop,
|
|
.par = (const uint32_t[]){TCG_COND_NEVER},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "loopgtz",
|
|
.translate = translate_loop,
|
|
.par = (const uint32_t[]){TCG_COND_GT},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "loopnez",
|
|
.translate = translate_loop,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "max",
|
|
.translate = translate_smax,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "maxu",
|
|
.translate = translate_umax,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "memw",
|
|
.translate = translate_memw,
|
|
}, {
|
|
.name = "min",
|
|
.translate = translate_smin,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "minu",
|
|
.translate = translate_umin,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "mov",
|
|
.translate = translate_mov,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mov.n",
|
|
.translate = translate_mov,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "moveqz",
|
|
.translate = translate_movcond,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "movf",
|
|
.translate = translate_movp,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "movgez",
|
|
.translate = translate_movcond,
|
|
.par = (const uint32_t[]){TCG_COND_GE},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "movi",
|
|
.translate = translate_movi,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "movi.n",
|
|
.translate = translate_movi,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "movltz",
|
|
.translate = translate_movcond,
|
|
.par = (const uint32_t[]){TCG_COND_LT},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "movnez",
|
|
.translate = translate_movcond,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "movsp",
|
|
.translate = translate_movsp,
|
|
.windowed_register_op = 0x3,
|
|
.op_flags = XTENSA_OP_ALLOCA,
|
|
}, {
|
|
.name = "movt",
|
|
.translate = translate_movp,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mul.aa.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AA, MAC16_HH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mul.aa.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AA, MAC16_HL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mul.aa.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AA, MAC16_LH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mul.aa.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AA, MAC16_LL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mul.ad.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AD, MAC16_HH, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mul.ad.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AD, MAC16_HL, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mul.ad.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AD, MAC16_LH, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mul.ad.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_AD, MAC16_LL, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mul.da.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DA, MAC16_HH, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mul.da.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DA, MAC16_HL, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mul.da.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DA, MAC16_LH, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mul.da.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DA, MAC16_LL, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mul.dd.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DD, MAC16_HH, 0},
|
|
}, {
|
|
.name = "mul.dd.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DD, MAC16_HL, 0},
|
|
}, {
|
|
.name = "mul.dd.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DD, MAC16_LH, 0},
|
|
}, {
|
|
.name = "mul.dd.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MUL, MAC16_DD, MAC16_LL, 0},
|
|
}, {
|
|
.name = "mul16s",
|
|
.translate = translate_mul16,
|
|
.par = (const uint32_t[]){true},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "mul16u",
|
|
.translate = translate_mul16,
|
|
.par = (const uint32_t[]){false},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "mula.aa.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AA, MAC16_HH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mula.aa.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AA, MAC16_HL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mula.aa.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AA, MAC16_LH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mula.aa.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AA, MAC16_LL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "mula.ad.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AD, MAC16_HH, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mula.ad.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AD, MAC16_HL, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mula.ad.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AD, MAC16_LH, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mula.ad.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_AD, MAC16_LL, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "mula.da.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_HH, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.da.hh.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_HH, -4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.da.hh.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_HH, 4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.da.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_HL, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.da.hl.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_HL, -4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.da.hl.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_HL, 4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.da.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_LH, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.da.lh.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_LH, -4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.da.lh.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_LH, 4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.da.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_LL, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.da.ll.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_LL, -4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.da.ll.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DA, MAC16_LL, 4},
|
|
.windowed_register_op = 0xa,
|
|
}, {
|
|
.name = "mula.dd.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_HH, 0},
|
|
}, {
|
|
.name = "mula.dd.hh.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_HH, -4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.dd.hh.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_HH, 4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.dd.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_HL, 0},
|
|
}, {
|
|
.name = "mula.dd.hl.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_HL, -4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.dd.hl.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_HL, 4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.dd.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_LH, 0},
|
|
}, {
|
|
.name = "mula.dd.lh.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_LH, -4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.dd.lh.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_LH, 4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.dd.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_LL, 0},
|
|
}, {
|
|
.name = "mula.dd.ll.lddec",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_LL, -4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mula.dd.ll.ldinc",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULA, MAC16_DD, MAC16_LL, 4},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "mull",
|
|
.translate = translate_mull,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "muls.aa.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AA, MAC16_HH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "muls.aa.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AA, MAC16_HL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "muls.aa.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AA, MAC16_LH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "muls.aa.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AA, MAC16_LL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "muls.ad.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AD, MAC16_HH, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "muls.ad.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AD, MAC16_HL, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "muls.ad.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AD, MAC16_LH, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "muls.ad.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_AD, MAC16_LL, 0},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "muls.da.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DA, MAC16_HH, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "muls.da.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DA, MAC16_HL, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "muls.da.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DA, MAC16_LH, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "muls.da.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DA, MAC16_LL, 0},
|
|
.windowed_register_op = 0x2,
|
|
}, {
|
|
.name = "muls.dd.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DD, MAC16_HH, 0},
|
|
}, {
|
|
.name = "muls.dd.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DD, MAC16_HL, 0},
|
|
}, {
|
|
.name = "muls.dd.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DD, MAC16_LH, 0},
|
|
}, {
|
|
.name = "muls.dd.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_MULS, MAC16_DD, MAC16_LL, 0},
|
|
}, {
|
|
.name = "mulsh",
|
|
.translate = translate_mulh,
|
|
.par = (const uint32_t[]){true},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "muluh",
|
|
.translate = translate_mulh,
|
|
.par = (const uint32_t[]){false},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "neg",
|
|
.translate = translate_neg,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "nop",
|
|
.translate = translate_nop,
|
|
}, {
|
|
.name = "nop.n",
|
|
.translate = translate_nop,
|
|
}, {
|
|
.name = "nsa",
|
|
.translate = translate_nsa,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "nsau",
|
|
.translate = translate_nsau,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "or",
|
|
.translate = translate_or,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "orb",
|
|
.translate = translate_boolean,
|
|
.par = (const uint32_t[]){BOOLEAN_OR},
|
|
}, {
|
|
.name = "orbc",
|
|
.translate = translate_boolean,
|
|
.par = (const uint32_t[]){BOOLEAN_ORC},
|
|
}, {
|
|
.name = "pdtlb",
|
|
.translate = translate_ptlb,
|
|
.par = (const uint32_t[]){true},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "pitlb",
|
|
.translate = translate_ptlb,
|
|
.par = (const uint32_t[]){false},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "quos",
|
|
.translate = translate_quos,
|
|
.par = (const uint32_t[]){true},
|
|
.op_flags = XTENSA_OP_DIVIDE_BY_ZERO,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "quou",
|
|
.translate = translate_quou,
|
|
.op_flags = XTENSA_OP_DIVIDE_BY_ZERO,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "rdtlb0",
|
|
.translate = translate_rtlb,
|
|
.par = (const uint32_t[]){true, 0},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "rdtlb1",
|
|
.translate = translate_rtlb,
|
|
.par = (const uint32_t[]){true, 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "read_impwire",
|
|
.translate = translate_read_impwire,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rems",
|
|
.translate = translate_quos,
|
|
.par = (const uint32_t[]){false},
|
|
.op_flags = XTENSA_OP_DIVIDE_BY_ZERO,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "remu",
|
|
.translate = translate_remu,
|
|
.op_flags = XTENSA_OP_DIVIDE_BY_ZERO,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "rer",
|
|
.translate = translate_rer,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "ret",
|
|
.translate = translate_ret,
|
|
}, {
|
|
.name = "ret.n",
|
|
.translate = translate_ret,
|
|
}, {
|
|
.name = "retw",
|
|
.translate = translate_retw,
|
|
.test_ill = test_ill_retw,
|
|
.op_flags = XTENSA_OP_UNDERFLOW,
|
|
}, {
|
|
.name = "retw.n",
|
|
.translate = translate_retw,
|
|
.test_ill = test_ill_retw,
|
|
.op_flags = XTENSA_OP_UNDERFLOW,
|
|
}, {
|
|
.name = "rfdd",
|
|
.op_flags = XTENSA_OP_ILL,
|
|
}, {
|
|
.name = "rfde",
|
|
.translate = translate_rfde,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
}, {
|
|
.name = "rfdo",
|
|
.op_flags = XTENSA_OP_ILL,
|
|
}, {
|
|
.name = "rfe",
|
|
.translate = translate_rfe,
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_CHECK_INTERRUPTS,
|
|
}, {
|
|
.name = "rfi",
|
|
.translate = translate_rfi,
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_CHECK_INTERRUPTS,
|
|
}, {
|
|
.name = "rfwo",
|
|
.translate = translate_rfw,
|
|
.par = (const uint32_t[]){true},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_CHECK_INTERRUPTS,
|
|
}, {
|
|
.name = "rfwu",
|
|
.translate = translate_rfw,
|
|
.par = (const uint32_t[]){false},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_CHECK_INTERRUPTS,
|
|
}, {
|
|
.name = "ritlb0",
|
|
.translate = translate_rtlb,
|
|
.par = (const uint32_t[]){false, 0},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "ritlb1",
|
|
.translate = translate_rtlb,
|
|
.par = (const uint32_t[]){false, 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "rotw",
|
|
.translate = translate_rotw,
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
}, {
|
|
.name = "rsil",
|
|
.translate = translate_rsil,
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.176",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){176},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.208",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){208},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.acchi",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){ACCHI},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.acclo",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){ACCLO},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.atomctl",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){ATOMCTL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.br",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){BR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.cacheattr",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CACHEATTR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ccompare0",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CCOMPARE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ccompare1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CCOMPARE + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ccompare2",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CCOMPARE + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ccount",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CCOUNT},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.configid0",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CONFIGID0},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.configid1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CONFIGID1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.cpenable",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){CPENABLE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.dbreaka0",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DBREAKA},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.dbreaka1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DBREAKA + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.dbreakc0",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DBREAKC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.dbreakc1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DBREAKC + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ddr",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.debugcause",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DEBUGCAUSE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.depc",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DEPC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.dtlbcfg",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){DTLBCFG},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.epc1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPC1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.epc2",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPC1 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.epc3",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPC1 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.epc4",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPC1 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.epc5",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPC1 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.epc6",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPC1 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.epc7",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPC1 + 6},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.eps2",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPS2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.eps3",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPS2 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.eps4",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPS2 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.eps5",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPS2 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.eps6",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPS2 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.eps7",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EPS2 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.exccause",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCCAUSE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excsave1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCSAVE1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excsave2",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excsave3",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excsave4",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excsave5",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excsave6",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excsave7",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 6},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.excvaddr",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){EXCVADDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ibreaka0",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){IBREAKA},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ibreaka1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){IBREAKA + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ibreakenable",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){IBREAKENABLE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.icount",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){ICOUNT},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.icountlevel",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){ICOUNTLEVEL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.intclear",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){INTCLEAR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.intenable",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){INTENABLE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.interrupt",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){INTSET},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.intset",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){INTSET},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.itlbcfg",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){ITLBCFG},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.lbeg",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){LBEG},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.lcount",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){LCOUNT},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.lend",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){LEND},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.litbase",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){LITBASE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.m0",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.m1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MR + 1},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.m2",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MR + 2},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.m3",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MR + 3},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.memctl",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MEMCTL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.misc0",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MISC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.misc1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MISC + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.misc2",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MISC + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.misc3",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){MISC + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.prid",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){PRID},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ps",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){PS},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.ptevaddr",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){PTEVADDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.rasid",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){RASID},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.sar",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){SAR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.scompare1",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){SCOMPARE1},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.vecbase",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){VECBASE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.windowbase",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){WINDOW_BASE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsr.windowstart",
|
|
.translate = translate_rsr,
|
|
.test_ill = test_ill_rsr,
|
|
.par = (const uint32_t[]){WINDOW_START},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rsync",
|
|
.translate = translate_nop,
|
|
}, {
|
|
.name = "rur.expstate",
|
|
.translate = translate_rur,
|
|
.par = (const uint32_t[]){EXPSTATE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "rur.fcr",
|
|
.translate = translate_rur,
|
|
.par = (const uint32_t[]){FCR},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "rur.fsr",
|
|
.translate = translate_rur,
|
|
.par = (const uint32_t[]){FSR},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "rur.threadptr",
|
|
.translate = translate_rur,
|
|
.par = (const uint32_t[]){THREADPTR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "s16i",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUW, false, true},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "s32c1i",
|
|
.translate = translate_s32c1i,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "s32e",
|
|
.translate = translate_s32e,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "s32i",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUL, false, true},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "s32i.n",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUL, false, true},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "s32nb",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUL, false, true},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "s32ri",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_TEUL, true, true},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "s8i",
|
|
.translate = translate_ldst,
|
|
.par = (const uint32_t[]){MO_UB, false, true},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "salt",
|
|
.translate = translate_salt,
|
|
.par = (const uint32_t[]){TCG_COND_LT},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "saltu",
|
|
.translate = translate_salt,
|
|
.par = (const uint32_t[]){TCG_COND_LTU},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "setb_expstate",
|
|
.translate = translate_setb_expstate,
|
|
}, {
|
|
.name = "sext",
|
|
.translate = translate_sext,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "simcall",
|
|
.translate = translate_simcall,
|
|
.test_ill = test_ill_simcall,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
}, {
|
|
.name = "sll",
|
|
.translate = translate_sll,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "slli",
|
|
.translate = translate_slli,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "sra",
|
|
.translate = translate_sra,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "srai",
|
|
.translate = translate_srai,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "src",
|
|
.translate = translate_src,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "srl",
|
|
.translate = translate_srl,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "srli",
|
|
.translate = translate_srli,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "ssa8b",
|
|
.translate = translate_ssa8b,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "ssa8l",
|
|
.translate = translate_ssa8l,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "ssai",
|
|
.translate = translate_ssai,
|
|
}, {
|
|
.name = "ssl",
|
|
.translate = translate_ssl,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "ssr",
|
|
.translate = translate_ssr,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "sub",
|
|
.translate = translate_sub,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "subx2",
|
|
.translate = translate_subx,
|
|
.par = (const uint32_t[]){1},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "subx4",
|
|
.translate = translate_subx,
|
|
.par = (const uint32_t[]){2},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "subx8",
|
|
.translate = translate_subx,
|
|
.par = (const uint32_t[]){3},
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "syscall",
|
|
.op_flags = XTENSA_OP_SYSCALL,
|
|
}, {
|
|
.name = "umul.aa.hh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_UMUL, MAC16_AA, MAC16_HH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "umul.aa.hl",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_UMUL, MAC16_AA, MAC16_HL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "umul.aa.lh",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_UMUL, MAC16_AA, MAC16_LH, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "umul.aa.ll",
|
|
.translate = translate_mac16,
|
|
.par = (const uint32_t[]){MAC16_UMUL, MAC16_AA, MAC16_LL, 0},
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "waiti",
|
|
.translate = translate_waiti,
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
}, {
|
|
.name = "wdtlb",
|
|
.translate = translate_wtlb,
|
|
.par = (const uint32_t[]){true},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "wer",
|
|
.translate = translate_wer,
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "witlb",
|
|
.translate = translate_wtlb,
|
|
.par = (const uint32_t[]){false},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "wrmsk_expstate",
|
|
.translate = translate_wrmsk_expstate,
|
|
.windowed_register_op = 0x3,
|
|
}, {
|
|
.name = "wsr.176",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){176},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.208",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){208},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.acchi",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){ACCHI},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.acclo",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){ACCLO},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.atomctl",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){ATOMCTL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.br",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){BR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.cacheattr",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CACHEATTR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ccompare0",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CCOMPARE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ccompare1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CCOMPARE + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ccompare2",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CCOMPARE + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ccount",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CCOUNT},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.configid0",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CONFIGID0},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.configid1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CONFIGID1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.cpenable",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){CPENABLE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.dbreaka0",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DBREAKA},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.dbreaka1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DBREAKA + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.dbreakc0",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DBREAKC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.dbreakc1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DBREAKC + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ddr",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.debugcause",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DEBUGCAUSE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.depc",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DEPC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.dtlbcfg",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){DTLBCFG},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.epc1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPC1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.epc2",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPC1 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.epc3",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPC1 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.epc4",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPC1 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.epc5",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPC1 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.epc6",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPC1 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.epc7",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPC1 + 6},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.eps2",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPS2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.eps3",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPS2 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.eps4",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPS2 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.eps5",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPS2 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.eps6",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPS2 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.eps7",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EPS2 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.exccause",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCCAUSE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excsave1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCSAVE1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excsave2",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excsave3",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excsave4",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excsave5",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excsave6",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excsave7",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 6},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.excvaddr",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){EXCVADDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ibreaka0",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){IBREAKA},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ibreaka1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){IBREAKA + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ibreakenable",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){IBREAKENABLE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.icount",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){ICOUNT},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.icountlevel",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){ICOUNTLEVEL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.intclear",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){INTCLEAR},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.intenable",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){INTENABLE},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.interrupt",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){INTSET},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.intset",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){INTSET},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.itlbcfg",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){ITLBCFG},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.lbeg",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){LBEG},
|
|
.op_flags = XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.lcount",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){LCOUNT},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.lend",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){LEND},
|
|
.op_flags = XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.litbase",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){LITBASE},
|
|
.op_flags = XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.m0",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.m1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MR + 1},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.m2",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MR + 2},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.m3",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MR + 3},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.memctl",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MEMCTL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.misc0",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MISC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.misc1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MISC + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.misc2",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MISC + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.misc3",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MISC + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.mmid",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){MMID},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.prid",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){PRID},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ps",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){PS},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_M1 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.ptevaddr",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){PTEVADDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.rasid",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){RASID},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.sar",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){SAR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.scompare1",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){SCOMPARE1},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.vecbase",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){VECBASE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.windowbase",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){WINDOW_BASE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wsr.windowstart",
|
|
.translate = translate_wsr,
|
|
.test_ill = test_ill_wsr,
|
|
.par = (const uint32_t[]){WINDOW_START},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wur.expstate",
|
|
.translate = translate_wur,
|
|
.par = (const uint32_t[]){EXPSTATE},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "wur.fcr",
|
|
.translate = translate_wur,
|
|
.par = (const uint32_t[]){FCR},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "wur.fsr",
|
|
.translate = translate_wur,
|
|
.par = (const uint32_t[]){FSR},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "wur.threadptr",
|
|
.translate = translate_wur,
|
|
.par = (const uint32_t[]){THREADPTR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xor",
|
|
.translate = translate_xor,
|
|
.windowed_register_op = 0x7,
|
|
}, {
|
|
.name = "xorb",
|
|
.translate = translate_boolean,
|
|
.par = (const uint32_t[]){BOOLEAN_XOR},
|
|
}, {
|
|
.name = "xsr.176",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){176},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.208",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){208},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.acchi",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){ACCHI},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.acclo",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){ACCLO},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.atomctl",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){ATOMCTL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.br",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){BR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.cacheattr",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CACHEATTR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ccompare0",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CCOMPARE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ccompare1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CCOMPARE + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ccompare2",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CCOMPARE + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ccount",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CCOUNT},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.configid0",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CONFIGID0},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.configid1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CONFIGID1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.cpenable",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){CPENABLE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.dbreaka0",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DBREAKA},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.dbreaka1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DBREAKA + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.dbreakc0",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DBREAKC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.dbreakc1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DBREAKC + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ddr",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.debugcause",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DEBUGCAUSE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.depc",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DEPC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.dtlbcfg",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){DTLBCFG},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.epc1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPC1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.epc2",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPC1 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.epc3",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPC1 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.epc4",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPC1 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.epc5",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPC1 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.epc6",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPC1 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.epc7",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPC1 + 6},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.eps2",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPS2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.eps3",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPS2 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.eps4",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPS2 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.eps5",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPS2 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.eps6",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPS2 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.eps7",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EPS2 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.exccause",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCCAUSE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excsave1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCSAVE1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excsave2",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excsave3",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excsave4",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excsave5",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 4},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excsave6",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 5},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excsave7",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCSAVE1 + 6},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.excvaddr",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){EXCVADDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ibreaka0",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){IBREAKA},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ibreaka1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){IBREAKA + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ibreakenable",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){IBREAKENABLE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_0,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.icount",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){ICOUNT},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.icountlevel",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){ICOUNTLEVEL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.intclear",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){INTCLEAR},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.intenable",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){INTENABLE},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.interrupt",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){INTSET},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.intset",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){INTSET},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_0 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.itlbcfg",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){ITLBCFG},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.lbeg",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){LBEG},
|
|
.op_flags = XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.lcount",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){LCOUNT},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.lend",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){LEND},
|
|
.op_flags = XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.litbase",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){LITBASE},
|
|
.op_flags = XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.m0",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.m1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MR + 1},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.m2",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MR + 2},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.m3",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MR + 3},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.memctl",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MEMCTL},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.misc0",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MISC},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.misc1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MISC + 1},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.misc2",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MISC + 2},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.misc3",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){MISC + 3},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.prid",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){PRID},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ps",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){PS},
|
|
.op_flags =
|
|
XTENSA_OP_PRIVILEGED |
|
|
XTENSA_OP_EXIT_TB_M1 |
|
|
XTENSA_OP_CHECK_INTERRUPTS,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.ptevaddr",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){PTEVADDR},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.rasid",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){RASID},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.sar",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){SAR},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.scompare1",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){SCOMPARE1},
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.vecbase",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){VECBASE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.windowbase",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){WINDOW_BASE},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
}, {
|
|
.name = "xsr.windowstart",
|
|
.translate = translate_xsr,
|
|
.test_ill = test_ill_xsr,
|
|
.par = (const uint32_t[]){WINDOW_START},
|
|
.op_flags = XTENSA_OP_PRIVILEGED | XTENSA_OP_EXIT_TB_M1,
|
|
.windowed_register_op = 0x1,
|
|
},
|
|
};
|
|
|
|
const XtensaOpcodeTranslators xtensa_core_opcodes = {
|
|
.num_opcodes = ARRAY_SIZE(core_ops),
|
|
.opcode = core_ops,
|
|
};
|
|
|
|
|
|
static void translate_abs_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_abs_s(cpu_FR[arg[0]], cpu_FR[arg[1]]);
|
|
}
|
|
|
|
static void translate_add_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_add_s(cpu_FR[arg[0]], cpu_env,
|
|
cpu_FR[arg[1]], cpu_FR[arg[2]]);
|
|
}
|
|
|
|
enum {
|
|
COMPARE_UN,
|
|
COMPARE_OEQ,
|
|
COMPARE_UEQ,
|
|
COMPARE_OLT,
|
|
COMPARE_ULT,
|
|
COMPARE_OLE,
|
|
COMPARE_ULE,
|
|
};
|
|
|
|
static void translate_compare_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
static void (* const helper[])(TCGv_env env, TCGv_i32 bit,
|
|
TCGv_i32 s, TCGv_i32 t) = {
|
|
[COMPARE_UN] = gen_helper_un_s,
|
|
[COMPARE_OEQ] = gen_helper_oeq_s,
|
|
[COMPARE_UEQ] = gen_helper_ueq_s,
|
|
[COMPARE_OLT] = gen_helper_olt_s,
|
|
[COMPARE_ULT] = gen_helper_ult_s,
|
|
[COMPARE_OLE] = gen_helper_ole_s,
|
|
[COMPARE_ULE] = gen_helper_ule_s,
|
|
};
|
|
TCGv_i32 bit = tcg_const_i32(1 << arg[0]);
|
|
|
|
helper[par[0]](cpu_env, bit, cpu_FR[arg[1]], cpu_FR[arg[2]]);
|
|
tcg_temp_free(bit);
|
|
}
|
|
|
|
static void translate_float_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 scale = tcg_const_i32(-arg[2]);
|
|
|
|
if (par[0]) {
|
|
gen_helper_uitof(cpu_FR[arg[0]], cpu_env, cpu_R[arg[1]], scale);
|
|
} else {
|
|
gen_helper_itof(cpu_FR[arg[0]], cpu_env, cpu_R[arg[1]], scale);
|
|
}
|
|
tcg_temp_free(scale);
|
|
}
|
|
|
|
static void translate_ftoi_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 rounding_mode = tcg_const_i32(par[0]);
|
|
TCGv_i32 scale = tcg_const_i32(arg[2]);
|
|
|
|
if (par[1]) {
|
|
gen_helper_ftoui(cpu_R[arg[0]], cpu_FR[arg[1]],
|
|
rounding_mode, scale);
|
|
} else {
|
|
gen_helper_ftoi(cpu_R[arg[0]], cpu_FR[arg[1]],
|
|
rounding_mode, scale);
|
|
}
|
|
tcg_temp_free(rounding_mode);
|
|
tcg_temp_free(scale);
|
|
}
|
|
|
|
static void translate_ldsti(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 addr = tcg_temp_new_i32();
|
|
|
|
tcg_gen_addi_i32(addr, cpu_R[arg[1]], arg[2]);
|
|
gen_load_store_alignment(dc, 2, addr, false);
|
|
if (par[0]) {
|
|
tcg_gen_qemu_st32(cpu_FR[arg[0]], addr, dc->cring);
|
|
} else {
|
|
tcg_gen_qemu_ld32u(cpu_FR[arg[0]], addr, dc->cring);
|
|
}
|
|
if (par[1]) {
|
|
tcg_gen_mov_i32(cpu_R[arg[1]], addr);
|
|
}
|
|
tcg_temp_free(addr);
|
|
}
|
|
|
|
static void translate_ldstx(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 addr = tcg_temp_new_i32();
|
|
|
|
tcg_gen_add_i32(addr, cpu_R[arg[1]], cpu_R[arg[2]]);
|
|
gen_load_store_alignment(dc, 2, addr, false);
|
|
if (par[0]) {
|
|
tcg_gen_qemu_st32(cpu_FR[arg[0]], addr, dc->cring);
|
|
} else {
|
|
tcg_gen_qemu_ld32u(cpu_FR[arg[0]], addr, dc->cring);
|
|
}
|
|
if (par[1]) {
|
|
tcg_gen_mov_i32(cpu_R[arg[1]], addr);
|
|
}
|
|
tcg_temp_free(addr);
|
|
}
|
|
|
|
static void translate_madd_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_madd_s(cpu_FR[arg[0]], cpu_env,
|
|
cpu_FR[arg[0]], cpu_FR[arg[1]], cpu_FR[arg[2]]);
|
|
}
|
|
|
|
static void translate_mov_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mov_i32(cpu_FR[arg[0]], cpu_FR[arg[1]]);
|
|
}
|
|
|
|
static void translate_movcond_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 zero = tcg_const_i32(0);
|
|
|
|
tcg_gen_movcond_i32(par[0], cpu_FR[arg[0]],
|
|
cpu_R[arg[2]], zero,
|
|
cpu_FR[arg[1]], cpu_FR[arg[0]]);
|
|
tcg_temp_free(zero);
|
|
}
|
|
|
|
static void translate_movp_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
TCGv_i32 zero = tcg_const_i32(0);
|
|
TCGv_i32 tmp = tcg_temp_new_i32();
|
|
|
|
tcg_gen_andi_i32(tmp, cpu_SR[BR], 1 << arg[2]);
|
|
tcg_gen_movcond_i32(par[0],
|
|
cpu_FR[arg[0]], tmp, zero,
|
|
cpu_FR[arg[1]], cpu_FR[arg[0]]);
|
|
tcg_temp_free(tmp);
|
|
tcg_temp_free(zero);
|
|
}
|
|
|
|
static void translate_mul_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_mul_s(cpu_FR[arg[0]], cpu_env,
|
|
cpu_FR[arg[1]], cpu_FR[arg[2]]);
|
|
}
|
|
|
|
static void translate_msub_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_msub_s(cpu_FR[arg[0]], cpu_env,
|
|
cpu_FR[arg[0]], cpu_FR[arg[1]], cpu_FR[arg[2]]);
|
|
}
|
|
|
|
static void translate_neg_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_neg_s(cpu_FR[arg[0]], cpu_FR[arg[1]]);
|
|
}
|
|
|
|
static void translate_rfr_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mov_i32(cpu_R[arg[0]], cpu_FR[arg[1]]);
|
|
}
|
|
|
|
static void translate_sub_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
gen_helper_sub_s(cpu_FR[arg[0]], cpu_env,
|
|
cpu_FR[arg[1]], cpu_FR[arg[2]]);
|
|
}
|
|
|
|
static void translate_wfr_s(DisasContext *dc, const uint32_t arg[],
|
|
const uint32_t par[])
|
|
{
|
|
tcg_gen_mov_i32(cpu_FR[arg[0]], cpu_R[arg[1]]);
|
|
}
|
|
|
|
static const XtensaOpcodeOps fpu2000_ops[] = {
|
|
{
|
|
.name = "abs.s",
|
|
.translate = translate_abs_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "add.s",
|
|
.translate = translate_add_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ceil.s",
|
|
.translate = translate_ftoi_s,
|
|
.par = (const uint32_t[]){float_round_up, false},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "float.s",
|
|
.translate = translate_float_s,
|
|
.par = (const uint32_t[]){false},
|
|
.windowed_register_op = 0x2,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "floor.s",
|
|
.translate = translate_ftoi_s,
|
|
.par = (const uint32_t[]){float_round_down, false},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "lsi",
|
|
.translate = translate_ldsti,
|
|
.par = (const uint32_t[]){false, false},
|
|
.windowed_register_op = 0x2,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "lsiu",
|
|
.translate = translate_ldsti,
|
|
.par = (const uint32_t[]){false, true},
|
|
.windowed_register_op = 0x2,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "lsx",
|
|
.translate = translate_ldstx,
|
|
.par = (const uint32_t[]){false, false},
|
|
.windowed_register_op = 0x6,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "lsxu",
|
|
.translate = translate_ldstx,
|
|
.par = (const uint32_t[]){false, true},
|
|
.windowed_register_op = 0x6,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "madd.s",
|
|
.translate = translate_madd_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "mov.s",
|
|
.translate = translate_mov_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "moveqz.s",
|
|
.translate = translate_movcond_s,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.windowed_register_op = 0x4,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "movf.s",
|
|
.translate = translate_movp_s,
|
|
.par = (const uint32_t[]){TCG_COND_EQ},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "movgez.s",
|
|
.translate = translate_movcond_s,
|
|
.par = (const uint32_t[]){TCG_COND_GE},
|
|
.windowed_register_op = 0x4,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "movltz.s",
|
|
.translate = translate_movcond_s,
|
|
.par = (const uint32_t[]){TCG_COND_LT},
|
|
.windowed_register_op = 0x4,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "movnez.s",
|
|
.translate = translate_movcond_s,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.windowed_register_op = 0x4,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "movt.s",
|
|
.translate = translate_movp_s,
|
|
.par = (const uint32_t[]){TCG_COND_NE},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "msub.s",
|
|
.translate = translate_msub_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "mul.s",
|
|
.translate = translate_mul_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "neg.s",
|
|
.translate = translate_neg_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "oeq.s",
|
|
.translate = translate_compare_s,
|
|
.par = (const uint32_t[]){COMPARE_OEQ},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ole.s",
|
|
.translate = translate_compare_s,
|
|
.par = (const uint32_t[]){COMPARE_OLE},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "olt.s",
|
|
.translate = translate_compare_s,
|
|
.par = (const uint32_t[]){COMPARE_OLT},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "rfr",
|
|
.translate = translate_rfr_s,
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "round.s",
|
|
.translate = translate_ftoi_s,
|
|
.par = (const uint32_t[]){float_round_nearest_even, false},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ssi",
|
|
.translate = translate_ldsti,
|
|
.par = (const uint32_t[]){true, false},
|
|
.windowed_register_op = 0x2,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ssiu",
|
|
.translate = translate_ldsti,
|
|
.par = (const uint32_t[]){true, true},
|
|
.windowed_register_op = 0x2,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ssx",
|
|
.translate = translate_ldstx,
|
|
.par = (const uint32_t[]){true, false},
|
|
.windowed_register_op = 0x6,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ssxu",
|
|
.translate = translate_ldstx,
|
|
.par = (const uint32_t[]){true, true},
|
|
.windowed_register_op = 0x6,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "sub.s",
|
|
.translate = translate_sub_s,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "trunc.s",
|
|
.translate = translate_ftoi_s,
|
|
.par = (const uint32_t[]){float_round_to_zero, false},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ueq.s",
|
|
.translate = translate_compare_s,
|
|
.par = (const uint32_t[]){COMPARE_UEQ},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ufloat.s",
|
|
.translate = translate_float_s,
|
|
.par = (const uint32_t[]){true},
|
|
.windowed_register_op = 0x2,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ule.s",
|
|
.translate = translate_compare_s,
|
|
.par = (const uint32_t[]){COMPARE_ULE},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "ult.s",
|
|
.translate = translate_compare_s,
|
|
.par = (const uint32_t[]){COMPARE_ULT},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "un.s",
|
|
.translate = translate_compare_s,
|
|
.par = (const uint32_t[]){COMPARE_UN},
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "utrunc.s",
|
|
.translate = translate_ftoi_s,
|
|
.par = (const uint32_t[]){float_round_to_zero, true},
|
|
.windowed_register_op = 0x1,
|
|
.coprocessor = 0x1,
|
|
}, {
|
|
.name = "wfr",
|
|
.translate = translate_wfr_s,
|
|
.windowed_register_op = 0x2,
|
|
.coprocessor = 0x1,
|
|
},
|
|
};
|
|
|
|
const XtensaOpcodeTranslators xtensa_fpu2000_opcodes = {
|
|
.num_opcodes = ARRAY_SIZE(fpu2000_ops),
|
|
.opcode = fpu2000_ops,
|
|
};
|