qemu-e2k/tcg/optimize.c

2239 lines
57 KiB
C
Raw Normal View History

/*
* Optimizations for Tiny Code Generator for QEMU
*
* Copyright (c) 2010 Samsung Electronics.
* Contributed by Kirill Batuzov <batuzovk@ispras.ru>
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "qemu/osdep.h"
#include "qemu/int128.h"
#include "tcg/tcg-op.h"
#include "tcg-internal.h"
#define CASE_OP_32_64(x) \
glue(glue(case INDEX_op_, x), _i32): \
glue(glue(case INDEX_op_, x), _i64)
#define CASE_OP_32_64_VEC(x) \
glue(glue(case INDEX_op_, x), _i32): \
glue(glue(case INDEX_op_, x), _i64): \
glue(glue(case INDEX_op_, x), _vec)
typedef struct TempOptInfo {
bool is_const;
TCGTemp *prev_copy;
TCGTemp *next_copy;
uint64_t val;
uint64_t z_mask; /* mask bit is 0 if and only if value bit is 0 */
uint64_t s_mask; /* a left-aligned mask of clrsb(value) bits. */
} TempOptInfo;
typedef struct OptContext {
TCGContext *tcg;
TCGOp *prev_mb;
TCGTempSet temps_used;
/* In flight values from optimization. */
uint64_t a_mask; /* mask bit is 0 iff value identical to first input */
uint64_t z_mask; /* mask bit is 0 iff value bit is 0 */
uint64_t s_mask; /* mask of clrsb(value) bits */
TCGType type;
} OptContext;
/* Calculate the smask for a specific value. */
static uint64_t smask_from_value(uint64_t value)
{
int rep = clrsb64(value);
return ~(~0ull >> rep);
}
/*
* Calculate the smask for a given set of known-zeros.
* If there are lots of zeros on the left, we can consider the remainder
* an unsigned field, and thus the corresponding signed field is one bit
* larger.
*/
static uint64_t smask_from_zmask(uint64_t zmask)
{
/*
* Only the 0 bits are significant for zmask, thus the msb itself
* must be zero, else we have no sign information.
*/
int rep = clz64(zmask);
if (rep == 0) {
return 0;
}
rep -= 1;
return ~(~0ull >> rep);
}
/*
* Recreate a properly left-aligned smask after manipulation.
* Some bit-shuffling, particularly shifts and rotates, may
* retain sign bits on the left, but may scatter disconnected
* sign bits on the right. Retain only what remains to the left.
*/
static uint64_t smask_from_smask(int64_t smask)
{
/* Only the 1 bits are significant for smask */
return smask_from_zmask(~smask);
}
static inline TempOptInfo *ts_info(TCGTemp *ts)
{
return ts->state_ptr;
}
static inline TempOptInfo *arg_info(TCGArg arg)
{
return ts_info(arg_temp(arg));
}
static inline bool ts_is_const(TCGTemp *ts)
{
return ts_info(ts)->is_const;
}
static inline bool arg_is_const(TCGArg arg)
{
return ts_is_const(arg_temp(arg));
}
static inline bool ts_is_copy(TCGTemp *ts)
{
return ts_info(ts)->next_copy != ts;
}
/* Reset TEMP's state, possibly removing the temp for the list of copies. */
static void reset_ts(TCGTemp *ts)
{
TempOptInfo *ti = ts_info(ts);
TempOptInfo *pi = ts_info(ti->prev_copy);
TempOptInfo *ni = ts_info(ti->next_copy);
ni->prev_copy = ti->prev_copy;
pi->next_copy = ti->next_copy;
ti->next_copy = ts;
ti->prev_copy = ts;
ti->is_const = false;
ti->z_mask = -1;
ti->s_mask = 0;
}
static void reset_temp(TCGArg arg)
{
reset_ts(arg_temp(arg));
}
/* Initialize and activate a temporary. */
static void init_ts_info(OptContext *ctx, TCGTemp *ts)
{
size_t idx = temp_idx(ts);
TempOptInfo *ti;
if (test_bit(idx, ctx->temps_used.l)) {
return;
}
set_bit(idx, ctx->temps_used.l);
ti = ts->state_ptr;
if (ti == NULL) {
ti = tcg_malloc(sizeof(TempOptInfo));
ts->state_ptr = ti;
}
ti->next_copy = ts;
ti->prev_copy = ts;
if (ts->kind == TEMP_CONST) {
ti->is_const = true;
ti->val = ts->val;
ti->z_mask = ts->val;
ti->s_mask = smask_from_value(ts->val);
} else {
ti->is_const = false;
ti->z_mask = -1;
ti->s_mask = 0;
}
}
static TCGTemp *find_better_copy(TCGContext *s, TCGTemp *ts)
{
TCGTemp *i, *g, *l;
/* If this is already readonly, we can't do better. */
if (temp_readonly(ts)) {
return ts;
}
g = l = NULL;
for (i = ts_info(ts)->next_copy; i != ts; i = ts_info(i)->next_copy) {
if (temp_readonly(i)) {
return i;
} else if (i->kind > ts->kind) {
if (i->kind == TEMP_GLOBAL) {
g = i;
} else if (i->kind == TEMP_LOCAL) {
l = i;
}
}
}
/* If we didn't find a better representation, return the same temp. */
return g ? g : l ? l : ts;
}
static bool ts_are_copies(TCGTemp *ts1, TCGTemp *ts2)
{
TCGTemp *i;
if (ts1 == ts2) {
return true;
}
if (!ts_is_copy(ts1) || !ts_is_copy(ts2)) {
return false;
}
for (i = ts_info(ts1)->next_copy; i != ts1; i = ts_info(i)->next_copy) {
if (i == ts2) {
return true;
}
}
return false;
}
static bool args_are_copies(TCGArg arg1, TCGArg arg2)
{
return ts_are_copies(arg_temp(arg1), arg_temp(arg2));
}
static bool tcg_opt_gen_mov(OptContext *ctx, TCGOp *op, TCGArg dst, TCGArg src)
{
TCGTemp *dst_ts = arg_temp(dst);
TCGTemp *src_ts = arg_temp(src);
TempOptInfo *di;
TempOptInfo *si;
TCGOpcode new_op;
if (ts_are_copies(dst_ts, src_ts)) {
tcg_op_remove(ctx->tcg, op);
return true;
}
reset_ts(dst_ts);
di = ts_info(dst_ts);
si = ts_info(src_ts);
switch (ctx->type) {
case TCG_TYPE_I32:
new_op = INDEX_op_mov_i32;
break;
case TCG_TYPE_I64:
new_op = INDEX_op_mov_i64;
break;
case TCG_TYPE_V64:
case TCG_TYPE_V128:
case TCG_TYPE_V256:
/* TCGOP_VECL and TCGOP_VECE remain unchanged. */
new_op = INDEX_op_mov_vec;
break;
default:
g_assert_not_reached();
}
op->opc = new_op;
op->args[0] = dst;
op->args[1] = src;
di->z_mask = si->z_mask;
di->s_mask = si->s_mask;
if (src_ts->type == dst_ts->type) {
TempOptInfo *ni = ts_info(si->next_copy);
di->next_copy = si->next_copy;
di->prev_copy = src_ts;
ni->prev_copy = dst_ts;
si->next_copy = dst_ts;
di->is_const = si->is_const;
di->val = si->val;
}
return true;
}
static bool tcg_opt_gen_movi(OptContext *ctx, TCGOp *op,
TCGArg dst, uint64_t val)
{
TCGTemp *tv;
if (ctx->type == TCG_TYPE_I32) {
val = (int32_t)val;
}
/* Convert movi to mov with constant temp. */
tv = tcg_constant_internal(ctx->type, val);
init_ts_info(ctx, tv);
return tcg_opt_gen_mov(ctx, op, dst, temp_arg(tv));
}
static uint64_t do_constant_folding_2(TCGOpcode op, uint64_t x, uint64_t y)
{
uint64_t l64, h64;
switch (op) {
CASE_OP_32_64(add):
return x + y;
CASE_OP_32_64(sub):
return x - y;
CASE_OP_32_64(mul):
return x * y;
CASE_OP_32_64_VEC(and):
return x & y;
CASE_OP_32_64_VEC(or):
return x | y;
CASE_OP_32_64_VEC(xor):
return x ^ y;
case INDEX_op_shl_i32:
return (uint32_t)x << (y & 31);
case INDEX_op_shl_i64:
return (uint64_t)x << (y & 63);
case INDEX_op_shr_i32:
return (uint32_t)x >> (y & 31);
case INDEX_op_shr_i64:
return (uint64_t)x >> (y & 63);
case INDEX_op_sar_i32:
return (int32_t)x >> (y & 31);
case INDEX_op_sar_i64:
return (int64_t)x >> (y & 63);
case INDEX_op_rotr_i32:
return ror32(x, y & 31);
case INDEX_op_rotr_i64:
return ror64(x, y & 63);
case INDEX_op_rotl_i32:
return rol32(x, y & 31);
case INDEX_op_rotl_i64:
return rol64(x, y & 63);
CASE_OP_32_64_VEC(not):
return ~x;
CASE_OP_32_64(neg):
return -x;
CASE_OP_32_64_VEC(andc):
return x & ~y;
CASE_OP_32_64_VEC(orc):
return x | ~y;
CASE_OP_32_64_VEC(eqv):
return ~(x ^ y);
CASE_OP_32_64_VEC(nand):
return ~(x & y);
CASE_OP_32_64_VEC(nor):
return ~(x | y);
case INDEX_op_clz_i32:
return (uint32_t)x ? clz32(x) : y;
case INDEX_op_clz_i64:
return x ? clz64(x) : y;
case INDEX_op_ctz_i32:
return (uint32_t)x ? ctz32(x) : y;
case INDEX_op_ctz_i64:
return x ? ctz64(x) : y;
case INDEX_op_ctpop_i32:
return ctpop32(x);
case INDEX_op_ctpop_i64:
return ctpop64(x);
CASE_OP_32_64(ext8s):
return (int8_t)x;
CASE_OP_32_64(ext16s):
return (int16_t)x;
CASE_OP_32_64(ext8u):
return (uint8_t)x;
CASE_OP_32_64(ext16u):
return (uint16_t)x;
CASE_OP_32_64(bswap16):
x = bswap16(x);
return y & TCG_BSWAP_OS ? (int16_t)x : x;
CASE_OP_32_64(bswap32):
x = bswap32(x);
return y & TCG_BSWAP_OS ? (int32_t)x : x;
case INDEX_op_bswap64_i64:
return bswap64(x);
case INDEX_op_ext_i32_i64:
case INDEX_op_ext32s_i64:
return (int32_t)x;
case INDEX_op_extu_i32_i64:
case INDEX_op_extrl_i64_i32:
case INDEX_op_ext32u_i64:
return (uint32_t)x;
case INDEX_op_extrh_i64_i32:
return (uint64_t)x >> 32;
case INDEX_op_muluh_i32:
return ((uint64_t)(uint32_t)x * (uint32_t)y) >> 32;
case INDEX_op_mulsh_i32:
return ((int64_t)(int32_t)x * (int32_t)y) >> 32;
case INDEX_op_muluh_i64:
mulu64(&l64, &h64, x, y);
return h64;
case INDEX_op_mulsh_i64:
muls64(&l64, &h64, x, y);
return h64;
case INDEX_op_div_i32:
/* Avoid crashing on divide by zero, otherwise undefined. */
return (int32_t)x / ((int32_t)y ? : 1);
case INDEX_op_divu_i32:
return (uint32_t)x / ((uint32_t)y ? : 1);
case INDEX_op_div_i64:
return (int64_t)x / ((int64_t)y ? : 1);
case INDEX_op_divu_i64:
return (uint64_t)x / ((uint64_t)y ? : 1);
case INDEX_op_rem_i32:
return (int32_t)x % ((int32_t)y ? : 1);
case INDEX_op_remu_i32:
return (uint32_t)x % ((uint32_t)y ? : 1);
case INDEX_op_rem_i64:
return (int64_t)x % ((int64_t)y ? : 1);
case INDEX_op_remu_i64:
return (uint64_t)x % ((uint64_t)y ? : 1);
default:
fprintf(stderr,
"Unrecognized operation %d in do_constant_folding.\n", op);
tcg_abort();
}
}
static uint64_t do_constant_folding(TCGOpcode op, TCGType type,
uint64_t x, uint64_t y)
{
uint64_t res = do_constant_folding_2(op, x, y);
if (type == TCG_TYPE_I32) {
tcg/optimize: fix constant signedness By convention, on a 64-bit host TCG internally stores 32-bit constants as sign-extended. This is not the case in the optimizer when a 32-bit constant is folded. This doesn't seem to have more consequences than suboptimal code generation. For instance the x86 backend assumes sign-extended constants, and in some rare cases uses a 32-bit unsigned immediate 0xffffffff instead of a 8-bit signed immediate 0xff for the constant -1. This is with a ppc guest: before ------ ---- 0x9f29cc movi_i32 tmp1,$0xffffffff movi_i32 tmp2,$0x0 add2_i32 tmp0,CA,CA,tmp2,r6,tmp2 add2_i32 tmp0,CA,tmp0,CA,tmp1,tmp2 mov_i32 r10,tmp0 0x7fd8c7dfe90c: xor %ebp,%ebp 0x7fd8c7dfe90e: mov %ebp,%r11d 0x7fd8c7dfe911: mov 0x18(%r14),%r9d 0x7fd8c7dfe915: add %r9d,%r10d 0x7fd8c7dfe918: adc %ebp,%r11d 0x7fd8c7dfe91b: add $0xffffffff,%r10d 0x7fd8c7dfe922: adc %ebp,%r11d 0x7fd8c7dfe925: mov %r11d,0x134(%r14) 0x7fd8c7dfe92c: mov %r10d,0x28(%r14) after ----- ---- 0x9f29cc movi_i32 tmp1,$0xffffffffffffffff movi_i32 tmp2,$0x0 add2_i32 tmp0,CA,CA,tmp2,r6,tmp2 add2_i32 tmp0,CA,tmp0,CA,tmp1,tmp2 mov_i32 r10,tmp0 0x7f37010d490c: xor %ebp,%ebp 0x7f37010d490e: mov %ebp,%r11d 0x7f37010d4911: mov 0x18(%r14),%r9d 0x7f37010d4915: add %r9d,%r10d 0x7f37010d4918: adc %ebp,%r11d 0x7f37010d491b: add $0xffffffffffffffff,%r10d 0x7f37010d491f: adc %ebp,%r11d 0x7f37010d4922: mov %r11d,0x134(%r14) 0x7f37010d4929: mov %r10d,0x28(%r14) Signed-off-by: Aurelien Jarno <aurelien@aurel32.net> Message-Id: <1436544211-2769-2-git-send-email-aurelien@aurel32.net> Signed-off-by: Richard Henderson <rth@twiddle.net>
2015-07-10 18:03:31 +02:00
res = (int32_t)res;
}
return res;
}
static bool do_constant_folding_cond_32(uint32_t x, uint32_t y, TCGCond c)
{
switch (c) {
case TCG_COND_EQ:
return x == y;
case TCG_COND_NE:
return x != y;
case TCG_COND_LT:
return (int32_t)x < (int32_t)y;
case TCG_COND_GE:
return (int32_t)x >= (int32_t)y;
case TCG_COND_LE:
return (int32_t)x <= (int32_t)y;
case TCG_COND_GT:
return (int32_t)x > (int32_t)y;
case TCG_COND_LTU:
return x < y;
case TCG_COND_GEU:
return x >= y;
case TCG_COND_LEU:
return x <= y;
case TCG_COND_GTU:
return x > y;
default:
tcg_abort();
}
}
static bool do_constant_folding_cond_64(uint64_t x, uint64_t y, TCGCond c)
{
switch (c) {
case TCG_COND_EQ:
return x == y;
case TCG_COND_NE:
return x != y;
case TCG_COND_LT:
return (int64_t)x < (int64_t)y;
case TCG_COND_GE:
return (int64_t)x >= (int64_t)y;
case TCG_COND_LE:
return (int64_t)x <= (int64_t)y;
case TCG_COND_GT:
return (int64_t)x > (int64_t)y;
case TCG_COND_LTU:
return x < y;
case TCG_COND_GEU:
return x >= y;
case TCG_COND_LEU:
return x <= y;
case TCG_COND_GTU:
return x > y;
default:
tcg_abort();
}
}
static bool do_constant_folding_cond_eq(TCGCond c)
{
switch (c) {
case TCG_COND_GT:
case TCG_COND_LTU:
case TCG_COND_LT:
case TCG_COND_GTU:
case TCG_COND_NE:
return 0;
case TCG_COND_GE:
case TCG_COND_GEU:
case TCG_COND_LE:
case TCG_COND_LEU:
case TCG_COND_EQ:
return 1;
default:
tcg_abort();
}
}
/*
* Return -1 if the condition can't be simplified,
* and the result of the condition (0 or 1) if it can.
*/
static int do_constant_folding_cond(TCGType type, TCGArg x,
TCGArg y, TCGCond c)
{
if (arg_is_const(x) && arg_is_const(y)) {
uint64_t xv = arg_info(x)->val;
uint64_t yv = arg_info(y)->val;
switch (type) {
case TCG_TYPE_I32:
return do_constant_folding_cond_32(xv, yv, c);
case TCG_TYPE_I64:
return do_constant_folding_cond_64(xv, yv, c);
default:
/* Only scalar comparisons are optimizable */
return -1;
}
} else if (args_are_copies(x, y)) {
return do_constant_folding_cond_eq(c);
} else if (arg_is_const(y) && arg_info(y)->val == 0) {
switch (c) {
case TCG_COND_LTU:
return 0;
case TCG_COND_GEU:
return 1;
default:
return -1;
}
}
return -1;
}
/*
* Return -1 if the condition can't be simplified,
* and the result of the condition (0 or 1) if it can.
*/
static int do_constant_folding_cond2(TCGArg *p1, TCGArg *p2, TCGCond c)
{
TCGArg al = p1[0], ah = p1[1];
TCGArg bl = p2[0], bh = p2[1];
if (arg_is_const(bl) && arg_is_const(bh)) {
tcg_target_ulong blv = arg_info(bl)->val;
tcg_target_ulong bhv = arg_info(bh)->val;
uint64_t b = deposit64(blv, 32, 32, bhv);
if (arg_is_const(al) && arg_is_const(ah)) {
tcg_target_ulong alv = arg_info(al)->val;
tcg_target_ulong ahv = arg_info(ah)->val;
uint64_t a = deposit64(alv, 32, 32, ahv);
return do_constant_folding_cond_64(a, b, c);
}
if (b == 0) {
switch (c) {
case TCG_COND_LTU:
return 0;
case TCG_COND_GEU:
return 1;
default:
break;
}
}
}
if (args_are_copies(al, bl) && args_are_copies(ah, bh)) {
return do_constant_folding_cond_eq(c);
}
return -1;
}
/**
* swap_commutative:
* @dest: TCGArg of the destination argument, or NO_DEST.
* @p1: first paired argument
* @p2: second paired argument
*
* If *@p1 is a constant and *@p2 is not, swap.
* If *@p2 matches @dest, swap.
* Return true if a swap was performed.
*/
#define NO_DEST temp_arg(NULL)
static bool swap_commutative(TCGArg dest, TCGArg *p1, TCGArg *p2)
{
TCGArg a1 = *p1, a2 = *p2;
int sum = 0;
sum += arg_is_const(a1);
sum -= arg_is_const(a2);
/* Prefer the constant in second argument, and then the form
op a, a, b, which is better handled on non-RISC hosts. */
if (sum > 0 || (sum == 0 && dest == a2)) {
*p1 = a2;
*p2 = a1;
return true;
}
return false;
}
static bool swap_commutative2(TCGArg *p1, TCGArg *p2)
{
int sum = 0;
sum += arg_is_const(p1[0]);
sum += arg_is_const(p1[1]);
sum -= arg_is_const(p2[0]);
sum -= arg_is_const(p2[1]);
if (sum > 0) {
TCGArg t;
t = p1[0], p1[0] = p2[0], p2[0] = t;
t = p1[1], p1[1] = p2[1], p2[1] = t;
return true;
}
return false;
}
static void init_arguments(OptContext *ctx, TCGOp *op, int nb_args)
{
for (int i = 0; i < nb_args; i++) {
TCGTemp *ts = arg_temp(op->args[i]);
init_ts_info(ctx, ts);
}
}
static void copy_propagate(OptContext *ctx, TCGOp *op,
int nb_oargs, int nb_iargs)
{
TCGContext *s = ctx->tcg;
for (int i = nb_oargs; i < nb_oargs + nb_iargs; i++) {
TCGTemp *ts = arg_temp(op->args[i]);
if (ts_is_copy(ts)) {
op->args[i] = temp_arg(find_better_copy(s, ts));
}
}
}
static void finish_folding(OptContext *ctx, TCGOp *op)
{
const TCGOpDef *def = &tcg_op_defs[op->opc];
int i, nb_oargs;
/*
* For an opcode that ends a BB, reset all temp data.
* We do no cross-BB optimization.
*/
if (def->flags & TCG_OPF_BB_END) {
memset(&ctx->temps_used, 0, sizeof(ctx->temps_used));
ctx->prev_mb = NULL;
return;
}
nb_oargs = def->nb_oargs;
for (i = 0; i < nb_oargs; i++) {
TCGTemp *ts = arg_temp(op->args[i]);
reset_ts(ts);
/*
* Save the corresponding known-zero/sign bits mask for the
* first output argument (only one supported so far).
*/
if (i == 0) {
ts_info(ts)->z_mask = ctx->z_mask;
ts_info(ts)->s_mask = ctx->s_mask;
}
}
}
/*
* The fold_* functions return true when processing is complete,
* usually by folding the operation to a constant or to a copy,
* and calling tcg_opt_gen_{mov,movi}. They may do other things,
* like collect information about the value produced, for use in
* optimizing a subsequent operation.
*
* These first fold_* functions are all helpers, used by other
* folders for more specific operations.
*/
static bool fold_const1(OptContext *ctx, TCGOp *op)
{
if (arg_is_const(op->args[1])) {
uint64_t t;
t = arg_info(op->args[1])->val;
t = do_constant_folding(op->opc, ctx->type, t, 0);
return tcg_opt_gen_movi(ctx, op, op->args[0], t);
}
return false;
}
static bool fold_const2(OptContext *ctx, TCGOp *op)
{
if (arg_is_const(op->args[1]) && arg_is_const(op->args[2])) {
uint64_t t1 = arg_info(op->args[1])->val;
uint64_t t2 = arg_info(op->args[2])->val;
t1 = do_constant_folding(op->opc, ctx->type, t1, t2);
return tcg_opt_gen_movi(ctx, op, op->args[0], t1);
}
return false;
}
static bool fold_commutative(OptContext *ctx, TCGOp *op)
{
swap_commutative(op->args[0], &op->args[1], &op->args[2]);
return false;
}
static bool fold_const2_commutative(OptContext *ctx, TCGOp *op)
{
swap_commutative(op->args[0], &op->args[1], &op->args[2]);
return fold_const2(ctx, op);
}
static bool fold_masks(OptContext *ctx, TCGOp *op)
{
uint64_t a_mask = ctx->a_mask;
uint64_t z_mask = ctx->z_mask;
uint64_t s_mask = ctx->s_mask;
/*
* 32-bit ops generate 32-bit results, which for the purpose of
* simplifying tcg are sign-extended. Certainly that's how we
* represent our constants elsewhere. Note that the bits will
* be reset properly for a 64-bit value when encountering the
* type changing opcodes.
*/
if (ctx->type == TCG_TYPE_I32) {
a_mask = (int32_t)a_mask;
z_mask = (int32_t)z_mask;
s_mask |= MAKE_64BIT_MASK(32, 32);
ctx->z_mask = z_mask;
ctx->s_mask = s_mask;
}
if (z_mask == 0) {
return tcg_opt_gen_movi(ctx, op, op->args[0], 0);
}
if (a_mask == 0) {
return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]);
}
return false;
}
/*
* Convert @op to NOT, if NOT is supported by the host.
* Return true f the conversion is successful, which will still
* indicate that the processing is complete.
*/
static bool fold_not(OptContext *ctx, TCGOp *op);
static bool fold_to_not(OptContext *ctx, TCGOp *op, int idx)
{
TCGOpcode not_op;
bool have_not;
switch (ctx->type) {
case TCG_TYPE_I32:
not_op = INDEX_op_not_i32;
have_not = TCG_TARGET_HAS_not_i32;
break;
case TCG_TYPE_I64:
not_op = INDEX_op_not_i64;
have_not = TCG_TARGET_HAS_not_i64;
break;
case TCG_TYPE_V64:
case TCG_TYPE_V128:
case TCG_TYPE_V256:
not_op = INDEX_op_not_vec;
have_not = TCG_TARGET_HAS_not_vec;
break;
default:
g_assert_not_reached();
}
if (have_not) {
op->opc = not_op;
op->args[1] = op->args[idx];
return fold_not(ctx, op);
}
return false;
}
/* If the binary operation has first argument @i, fold to @i. */
static bool fold_ix_to_i(OptContext *ctx, TCGOp *op, uint64_t i)
{
if (arg_is_const(op->args[1]) && arg_info(op->args[1])->val == i) {
return tcg_opt_gen_movi(ctx, op, op->args[0], i);
}
return false;
}
/* If the binary operation has first argument @i, fold to NOT. */
static bool fold_ix_to_not(OptContext *ctx, TCGOp *op, uint64_t i)
{
if (arg_is_const(op->args[1]) && arg_info(op->args[1])->val == i) {
return fold_to_not(ctx, op, 2);
}
return false;
}
/* If the binary operation has second argument @i, fold to @i. */
static bool fold_xi_to_i(OptContext *ctx, TCGOp *op, uint64_t i)
{
if (arg_is_const(op->args[2]) && arg_info(op->args[2])->val == i) {
return tcg_opt_gen_movi(ctx, op, op->args[0], i);
}
return false;
}
/* If the binary operation has second argument @i, fold to identity. */
static bool fold_xi_to_x(OptContext *ctx, TCGOp *op, uint64_t i)
{
if (arg_is_const(op->args[2]) && arg_info(op->args[2])->val == i) {
return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]);
}
return false;
}
/* If the binary operation has second argument @i, fold to NOT. */
static bool fold_xi_to_not(OptContext *ctx, TCGOp *op, uint64_t i)
{
if (arg_is_const(op->args[2]) && arg_info(op->args[2])->val == i) {
return fold_to_not(ctx, op, 1);
}
return false;
}
/* If the binary operation has both arguments equal, fold to @i. */
static bool fold_xx_to_i(OptContext *ctx, TCGOp *op, uint64_t i)
{
if (args_are_copies(op->args[1], op->args[2])) {
return tcg_opt_gen_movi(ctx, op, op->args[0], i);
}
return false;
}
/* If the binary operation has both arguments equal, fold to identity. */
static bool fold_xx_to_x(OptContext *ctx, TCGOp *op)
{
if (args_are_copies(op->args[1], op->args[2])) {
return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]);
}
return false;
}
/*
* These outermost fold_<op> functions are sorted alphabetically.
*
* The ordering of the transformations should be:
* 1) those that produce a constant
* 2) those that produce a copy
* 3) those that produce information about the result value.
*/
static bool fold_add(OptContext *ctx, TCGOp *op)
{
if (fold_const2_commutative(ctx, op) ||
fold_xi_to_x(ctx, op, 0)) {
return true;
}
return false;
}
/* We cannot as yet do_constant_folding with vectors. */
static bool fold_add_vec(OptContext *ctx, TCGOp *op)
{
if (fold_commutative(ctx, op) ||
fold_xi_to_x(ctx, op, 0)) {
return true;
}
return false;
}
static bool fold_addsub2(OptContext *ctx, TCGOp *op, bool add)
{
if (arg_is_const(op->args[2]) && arg_is_const(op->args[3]) &&
arg_is_const(op->args[4]) && arg_is_const(op->args[5])) {
uint64_t al = arg_info(op->args[2])->val;
uint64_t ah = arg_info(op->args[3])->val;
uint64_t bl = arg_info(op->args[4])->val;
uint64_t bh = arg_info(op->args[5])->val;
TCGArg rl, rh;
TCGOp *op2;
if (ctx->type == TCG_TYPE_I32) {
uint64_t a = deposit64(al, 32, 32, ah);
uint64_t b = deposit64(bl, 32, 32, bh);
if (add) {
a += b;
} else {
a -= b;
}
al = sextract64(a, 0, 32);
ah = sextract64(a, 32, 32);
} else {
Int128 a = int128_make128(al, ah);
Int128 b = int128_make128(bl, bh);
if (add) {
a = int128_add(a, b);
} else {
a = int128_sub(a, b);
}
al = int128_getlo(a);
ah = int128_gethi(a);
}
rl = op->args[0];
rh = op->args[1];
/* The proper opcode is supplied by tcg_opt_gen_mov. */
op2 = tcg_op_insert_before(ctx->tcg, op, 0, 2);
tcg_opt_gen_movi(ctx, op, rl, al);
tcg_opt_gen_movi(ctx, op2, rh, ah);
return true;
}
return false;
}
static bool fold_add2(OptContext *ctx, TCGOp *op)
{
/* Note that the high and low parts may be independently swapped. */
swap_commutative(op->args[0], &op->args[2], &op->args[4]);
swap_commutative(op->args[1], &op->args[3], &op->args[5]);
return fold_addsub2(ctx, op, true);
}
static bool fold_and(OptContext *ctx, TCGOp *op)
{
uint64_t z1, z2;
if (fold_const2_commutative(ctx, op) ||
fold_xi_to_i(ctx, op, 0) ||
fold_xi_to_x(ctx, op, -1) ||
fold_xx_to_x(ctx, op)) {
return true;
}
z1 = arg_info(op->args[1])->z_mask;
z2 = arg_info(op->args[2])->z_mask;
ctx->z_mask = z1 & z2;
/*
* Sign repetitions are perforce all identical, whether they are 1 or 0.
* Bitwise operations preserve the relative quantity of the repetitions.
*/
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
/*
* Known-zeros does not imply known-ones. Therefore unless
* arg2 is constant, we can't infer affected bits from it.
*/
if (arg_is_const(op->args[2])) {
ctx->a_mask = z1 & ~z2;
}
return fold_masks(ctx, op);
}
static bool fold_andc(OptContext *ctx, TCGOp *op)
{
uint64_t z1;
if (fold_const2(ctx, op) ||
fold_xx_to_i(ctx, op, 0) ||
fold_xi_to_x(ctx, op, 0) ||
fold_ix_to_not(ctx, op, -1)) {
return true;
}
z1 = arg_info(op->args[1])->z_mask;
/*
* Known-zeros does not imply known-ones. Therefore unless
* arg2 is constant, we can't infer anything from it.
*/
if (arg_is_const(op->args[2])) {
uint64_t z2 = ~arg_info(op->args[2])->z_mask;
ctx->a_mask = z1 & ~z2;
z1 &= z2;
}
ctx->z_mask = z1;
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
return fold_masks(ctx, op);
}
static bool fold_brcond(OptContext *ctx, TCGOp *op)
{
TCGCond cond = op->args[2];
int i;
if (swap_commutative(NO_DEST, &op->args[0], &op->args[1])) {
op->args[2] = cond = tcg_swap_cond(cond);
}
i = do_constant_folding_cond(ctx->type, op->args[0], op->args[1], cond);
if (i == 0) {
tcg_op_remove(ctx->tcg, op);
return true;
}
if (i > 0) {
op->opc = INDEX_op_br;
op->args[0] = op->args[3];
}
return false;
}
static bool fold_brcond2(OptContext *ctx, TCGOp *op)
{
TCGCond cond = op->args[4];
TCGArg label = op->args[5];
int i, inv = 0;
if (swap_commutative2(&op->args[0], &op->args[2])) {
op->args[4] = cond = tcg_swap_cond(cond);
}
i = do_constant_folding_cond2(&op->args[0], &op->args[2], cond);
if (i >= 0) {
goto do_brcond_const;
}
switch (cond) {
case TCG_COND_LT:
case TCG_COND_GE:
/*
* Simplify LT/GE comparisons vs zero to a single compare
* vs the high word of the input.
*/
if (arg_is_const(op->args[2]) && arg_info(op->args[2])->val == 0 &&
arg_is_const(op->args[3]) && arg_info(op->args[3])->val == 0) {
goto do_brcond_high;
}
break;
case TCG_COND_NE:
inv = 1;
QEMU_FALLTHROUGH;
case TCG_COND_EQ:
/*
* Simplify EQ/NE comparisons where one of the pairs
* can be simplified.
*/
i = do_constant_folding_cond(TCG_TYPE_I32, op->args[0],
op->args[2], cond);
switch (i ^ inv) {
case 0:
goto do_brcond_const;
case 1:
goto do_brcond_high;
}
i = do_constant_folding_cond(TCG_TYPE_I32, op->args[1],
op->args[3], cond);
switch (i ^ inv) {
case 0:
goto do_brcond_const;
case 1:
op->opc = INDEX_op_brcond_i32;
op->args[1] = op->args[2];
op->args[2] = cond;
op->args[3] = label;
break;
}
break;
default:
break;
do_brcond_high:
op->opc = INDEX_op_brcond_i32;
op->args[0] = op->args[1];
op->args[1] = op->args[3];
op->args[2] = cond;
op->args[3] = label;
break;
do_brcond_const:
if (i == 0) {
tcg_op_remove(ctx->tcg, op);
return true;
}
op->opc = INDEX_op_br;
op->args[0] = label;
break;
}
return false;
}
static bool fold_bswap(OptContext *ctx, TCGOp *op)
{
uint64_t z_mask, s_mask, sign;
if (arg_is_const(op->args[1])) {
uint64_t t = arg_info(op->args[1])->val;
t = do_constant_folding(op->opc, ctx->type, t, op->args[2]);
return tcg_opt_gen_movi(ctx, op, op->args[0], t);
}
z_mask = arg_info(op->args[1])->z_mask;
switch (op->opc) {
case INDEX_op_bswap16_i32:
case INDEX_op_bswap16_i64:
z_mask = bswap16(z_mask);
sign = INT16_MIN;
break;
case INDEX_op_bswap32_i32:
case INDEX_op_bswap32_i64:
z_mask = bswap32(z_mask);
sign = INT32_MIN;
break;
case INDEX_op_bswap64_i64:
z_mask = bswap64(z_mask);
sign = INT64_MIN;
break;
default:
g_assert_not_reached();
}
s_mask = smask_from_zmask(z_mask);
switch (op->args[2] & (TCG_BSWAP_OZ | TCG_BSWAP_OS)) {
case TCG_BSWAP_OZ:
break;
case TCG_BSWAP_OS:
/* If the sign bit may be 1, force all the bits above to 1. */
if (z_mask & sign) {
z_mask |= sign;
s_mask = sign << 1;
}
break;
default:
/* The high bits are undefined: force all bits above the sign to 1. */
z_mask |= sign << 1;
s_mask = 0;
break;
}
ctx->z_mask = z_mask;
ctx->s_mask = s_mask;
return fold_masks(ctx, op);
}
static bool fold_call(OptContext *ctx, TCGOp *op)
{
TCGContext *s = ctx->tcg;
int nb_oargs = TCGOP_CALLO(op);
int nb_iargs = TCGOP_CALLI(op);
int flags, i;
init_arguments(ctx, op, nb_oargs + nb_iargs);
copy_propagate(ctx, op, nb_oargs, nb_iargs);
/* If the function reads or writes globals, reset temp data. */
flags = tcg_call_flags(op);
if (!(flags & (TCG_CALL_NO_READ_GLOBALS | TCG_CALL_NO_WRITE_GLOBALS))) {
int nb_globals = s->nb_globals;
for (i = 0; i < nb_globals; i++) {
if (test_bit(i, ctx->temps_used.l)) {
reset_ts(&ctx->tcg->temps[i]);
}
}
}
/* Reset temp data for outputs. */
for (i = 0; i < nb_oargs; i++) {
reset_temp(op->args[i]);
}
/* Stop optimizing MB across calls. */
ctx->prev_mb = NULL;
return true;
}
static bool fold_count_zeros(OptContext *ctx, TCGOp *op)
{
uint64_t z_mask;
if (arg_is_const(op->args[1])) {
uint64_t t = arg_info(op->args[1])->val;
if (t != 0) {
t = do_constant_folding(op->opc, ctx->type, t, 0);
return tcg_opt_gen_movi(ctx, op, op->args[0], t);
}
return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[2]);
}
switch (ctx->type) {
case TCG_TYPE_I32:
z_mask = 31;
break;
case TCG_TYPE_I64:
z_mask = 63;
break;
default:
g_assert_not_reached();
}
ctx->z_mask = arg_info(op->args[2])->z_mask | z_mask;
ctx->s_mask = smask_from_zmask(ctx->z_mask);
return false;
}
static bool fold_ctpop(OptContext *ctx, TCGOp *op)
{
if (fold_const1(ctx, op)) {
return true;
}
switch (ctx->type) {
case TCG_TYPE_I32:
ctx->z_mask = 32 | 31;
break;
case TCG_TYPE_I64:
ctx->z_mask = 64 | 63;
break;
default:
g_assert_not_reached();
}
ctx->s_mask = smask_from_zmask(ctx->z_mask);
return false;
}
static bool fold_deposit(OptContext *ctx, TCGOp *op)
{
if (arg_is_const(op->args[1]) && arg_is_const(op->args[2])) {
uint64_t t1 = arg_info(op->args[1])->val;
uint64_t t2 = arg_info(op->args[2])->val;
t1 = deposit64(t1, op->args[3], op->args[4], t2);
return tcg_opt_gen_movi(ctx, op, op->args[0], t1);
}
ctx->z_mask = deposit64(arg_info(op->args[1])->z_mask,
op->args[3], op->args[4],
arg_info(op->args[2])->z_mask);
return false;
}
static bool fold_divide(OptContext *ctx, TCGOp *op)
{
if (fold_const2(ctx, op) ||
fold_xi_to_x(ctx, op, 1)) {
return true;
}
return false;
}
static bool fold_dup(OptContext *ctx, TCGOp *op)
{
if (arg_is_const(op->args[1])) {
uint64_t t = arg_info(op->args[1])->val;
t = dup_const(TCGOP_VECE(op), t);
return tcg_opt_gen_movi(ctx, op, op->args[0], t);
}
return false;
}
static bool fold_dup2(OptContext *ctx, TCGOp *op)
{
if (arg_is_const(op->args[1]) && arg_is_const(op->args[2])) {
uint64_t t = deposit64(arg_info(op->args[1])->val, 32, 32,
arg_info(op->args[2])->val);
return tcg_opt_gen_movi(ctx, op, op->args[0], t);
}
if (args_are_copies(op->args[1], op->args[2])) {
op->opc = INDEX_op_dup_vec;
TCGOP_VECE(op) = MO_32;
}
return false;
}
static bool fold_eqv(OptContext *ctx, TCGOp *op)
{
if (fold_const2_commutative(ctx, op) ||
fold_xi_to_x(ctx, op, -1) ||
fold_xi_to_not(ctx, op, 0)) {
return true;
}
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
return false;
}
static bool fold_extract(OptContext *ctx, TCGOp *op)
{
uint64_t z_mask_old, z_mask;
int pos = op->args[2];
int len = op->args[3];
if (arg_is_const(op->args[1])) {
uint64_t t;
t = arg_info(op->args[1])->val;
t = extract64(t, pos, len);
return tcg_opt_gen_movi(ctx, op, op->args[0], t);
}
z_mask_old = arg_info(op->args[1])->z_mask;
z_mask = extract64(z_mask_old, pos, len);
if (pos == 0) {
ctx->a_mask = z_mask_old ^ z_mask;
}
ctx->z_mask = z_mask;
ctx->s_mask = smask_from_zmask(z_mask);
return fold_masks(ctx, op);
}
static bool fold_extract2(OptContext *ctx, TCGOp *op)
{
if (arg_is_const(op->args[1]) && arg_is_const(op->args[2])) {
uint64_t v1 = arg_info(op->args[1])->val;
uint64_t v2 = arg_info(op->args[2])->val;
int shr = op->args[3];
if (op->opc == INDEX_op_extract2_i64) {
v1 >>= shr;
v2 <<= 64 - shr;
} else {
v1 = (uint32_t)v1 >> shr;
v2 = (uint64_t)((int32_t)v2 << (32 - shr));
}
return tcg_opt_gen_movi(ctx, op, op->args[0], v1 | v2);
}
return false;
}
static bool fold_exts(OptContext *ctx, TCGOp *op)
{
uint64_t s_mask_old, s_mask, z_mask, sign;
bool type_change = false;
if (fold_const1(ctx, op)) {
return true;
}
z_mask = arg_info(op->args[1])->z_mask;
s_mask = arg_info(op->args[1])->s_mask;
s_mask_old = s_mask;
switch (op->opc) {
CASE_OP_32_64(ext8s):
sign = INT8_MIN;
z_mask = (uint8_t)z_mask;
break;
CASE_OP_32_64(ext16s):
sign = INT16_MIN;
z_mask = (uint16_t)z_mask;
break;
case INDEX_op_ext_i32_i64:
type_change = true;
QEMU_FALLTHROUGH;
case INDEX_op_ext32s_i64:
sign = INT32_MIN;
z_mask = (uint32_t)z_mask;
break;
default:
g_assert_not_reached();
}
if (z_mask & sign) {
z_mask |= sign;
}
s_mask |= sign << 1;
ctx->z_mask = z_mask;
ctx->s_mask = s_mask;
if (!type_change) {
ctx->a_mask = s_mask & ~s_mask_old;
}
return fold_masks(ctx, op);
}
static bool fold_extu(OptContext *ctx, TCGOp *op)
{
uint64_t z_mask_old, z_mask;
bool type_change = false;
if (fold_const1(ctx, op)) {
return true;
}
z_mask_old = z_mask = arg_info(op->args[1])->z_mask;
switch (op->opc) {
CASE_OP_32_64(ext8u):
z_mask = (uint8_t)z_mask;
break;
CASE_OP_32_64(ext16u):
z_mask = (uint16_t)z_mask;
break;
case INDEX_op_extrl_i64_i32:
case INDEX_op_extu_i32_i64:
type_change = true;
QEMU_FALLTHROUGH;
case INDEX_op_ext32u_i64:
z_mask = (uint32_t)z_mask;
break;
case INDEX_op_extrh_i64_i32:
type_change = true;
z_mask >>= 32;
break;
default:
g_assert_not_reached();
}
ctx->z_mask = z_mask;
ctx->s_mask = smask_from_zmask(z_mask);
if (!type_change) {
ctx->a_mask = z_mask_old ^ z_mask;
}
return fold_masks(ctx, op);
}
static bool fold_mb(OptContext *ctx, TCGOp *op)
{
/* Eliminate duplicate and redundant fence instructions. */
if (ctx->prev_mb) {
/*
* Merge two barriers of the same type into one,
* or a weaker barrier into a stronger one,
* or two weaker barriers into a stronger one.
* mb X; mb Y => mb X|Y
* mb; strl => mb; st
* ldaq; mb => ld; mb
* ldaq; strl => ld; mb; st
* Other combinations are also merged into a strong
* barrier. This is stricter than specified but for
* the purposes of TCG is better than not optimizing.
*/
ctx->prev_mb->args[0] |= op->args[0];
tcg_op_remove(ctx->tcg, op);
} else {
ctx->prev_mb = op;
}
return true;
}
static bool fold_mov(OptContext *ctx, TCGOp *op)
{
return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]);
}
static bool fold_movcond(OptContext *ctx, TCGOp *op)
{
TCGCond cond = op->args[5];
int i;
if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) {
op->args[5] = cond = tcg_swap_cond(cond);
}
/*
* Canonicalize the "false" input reg to match the destination reg so
* that the tcg backend can implement a "move if true" operation.
*/
if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) {
op->args[5] = cond = tcg_invert_cond(cond);
}
i = do_constant_folding_cond(ctx->type, op->args[1], op->args[2], cond);
if (i >= 0) {
return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[4 - i]);
}
ctx->z_mask = arg_info(op->args[3])->z_mask
| arg_info(op->args[4])->z_mask;
ctx->s_mask = arg_info(op->args[3])->s_mask
& arg_info(op->args[4])->s_mask;
if (arg_is_const(op->args[3]) && arg_is_const(op->args[4])) {
uint64_t tv = arg_info(op->args[3])->val;
uint64_t fv = arg_info(op->args[4])->val;
TCGOpcode opc;
switch (ctx->type) {
case TCG_TYPE_I32:
opc = INDEX_op_setcond_i32;
break;
case TCG_TYPE_I64:
opc = INDEX_op_setcond_i64;
break;
default:
g_assert_not_reached();
}
if (tv == 1 && fv == 0) {
op->opc = opc;
op->args[3] = cond;
} else if (fv == 1 && tv == 0) {
op->opc = opc;
op->args[3] = tcg_invert_cond(cond);
}
}
return false;
}
static bool fold_mul(OptContext *ctx, TCGOp *op)
{
if (fold_const2(ctx, op) ||
fold_xi_to_i(ctx, op, 0) ||
fold_xi_to_x(ctx, op, 1)) {
return true;
}
return false;
}
static bool fold_mul_highpart(OptContext *ctx, TCGOp *op)
{
if (fold_const2_commutative(ctx, op) ||
fold_xi_to_i(ctx, op, 0)) {
return true;
}
return false;
}
static bool fold_multiply2(OptContext *ctx, TCGOp *op)
{
swap_commutative(op->args[0], &op->args[2], &op->args[3]);
if (arg_is_const(op->args[2]) && arg_is_const(op->args[3])) {
uint64_t a = arg_info(op->args[2])->val;
uint64_t b = arg_info(op->args[3])->val;
uint64_t h, l;
TCGArg rl, rh;
TCGOp *op2;
switch (op->opc) {
case INDEX_op_mulu2_i32:
l = (uint64_t)(uint32_t)a * (uint32_t)b;
h = (int32_t)(l >> 32);
l = (int32_t)l;
break;
case INDEX_op_muls2_i32:
l = (int64_t)(int32_t)a * (int32_t)b;
h = l >> 32;
l = (int32_t)l;
break;
case INDEX_op_mulu2_i64:
mulu64(&l, &h, a, b);
break;
case INDEX_op_muls2_i64:
muls64(&l, &h, a, b);
break;
default:
g_assert_not_reached();
}
rl = op->args[0];
rh = op->args[1];
/* The proper opcode is supplied by tcg_opt_gen_mov. */
op2 = tcg_op_insert_before(ctx->tcg, op, 0, 2);
tcg_opt_gen_movi(ctx, op, rl, l);
tcg_opt_gen_movi(ctx, op2, rh, h);
return true;
}
return false;
}
static bool fold_nand(OptContext *ctx, TCGOp *op)
{
if (fold_const2_commutative(ctx, op) ||
fold_xi_to_not(ctx, op, -1)) {
return true;
}
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
return false;
}
static bool fold_neg(OptContext *ctx, TCGOp *op)
{
uint64_t z_mask;
if (fold_const1(ctx, op)) {
return true;
}
/* Set to 1 all bits to the left of the rightmost. */
z_mask = arg_info(op->args[1])->z_mask;
ctx->z_mask = -(z_mask & -z_mask);
/*
* Because of fold_sub_to_neg, we want to always return true,
* via finish_folding.
*/
finish_folding(ctx, op);
return true;
}
static bool fold_nor(OptContext *ctx, TCGOp *op)
{
if (fold_const2_commutative(ctx, op) ||
fold_xi_to_not(ctx, op, 0)) {
return true;
}
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
return false;
}
static bool fold_not(OptContext *ctx, TCGOp *op)
{
if (fold_const1(ctx, op)) {
return true;
}
ctx->s_mask = arg_info(op->args[1])->s_mask;
/* Because of fold_to_not, we want to always return true, via finish. */
finish_folding(ctx, op);
return true;
}
static bool fold_or(OptContext *ctx, TCGOp *op)
{
if (fold_const2_commutative(ctx, op) ||
fold_xi_to_x(ctx, op, 0) ||
fold_xx_to_x(ctx, op)) {
return true;
}
ctx->z_mask = arg_info(op->args[1])->z_mask
| arg_info(op->args[2])->z_mask;
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
return fold_masks(ctx, op);
}
static bool fold_orc(OptContext *ctx, TCGOp *op)
{
if (fold_const2(ctx, op) ||
fold_xx_to_i(ctx, op, -1) ||
fold_xi_to_x(ctx, op, -1) ||
fold_ix_to_not(ctx, op, 0)) {
return true;
}
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
return false;
}
static bool fold_qemu_ld(OptContext *ctx, TCGOp *op)
{
const TCGOpDef *def = &tcg_op_defs[op->opc];
MemOpIdx oi = op->args[def->nb_oargs + def->nb_iargs];
MemOp mop = get_memop(oi);
int width = 8 * memop_size(mop);
if (width < 64) {
ctx->s_mask = MAKE_64BIT_MASK(width, 64 - width);
if (!(mop & MO_SIGN)) {
ctx->z_mask = MAKE_64BIT_MASK(0, width);
ctx->s_mask <<= 1;
}
}
/* Opcodes that touch guest memory stop the mb optimization. */
ctx->prev_mb = NULL;
return false;
}
static bool fold_qemu_st(OptContext *ctx, TCGOp *op)
{
/* Opcodes that touch guest memory stop the mb optimization. */
ctx->prev_mb = NULL;
return false;
}
static bool fold_remainder(OptContext *ctx, TCGOp *op)
{
if (fold_const2(ctx, op) ||
fold_xx_to_i(ctx, op, 0)) {
return true;
}
return false;
}
static bool fold_setcond(OptContext *ctx, TCGOp *op)
{
TCGCond cond = op->args[3];
int i;
if (swap_commutative(op->args[0], &op->args[1], &op->args[2])) {
op->args[3] = cond = tcg_swap_cond(cond);
}
i = do_constant_folding_cond(ctx->type, op->args[1], op->args[2], cond);
if (i >= 0) {
return tcg_opt_gen_movi(ctx, op, op->args[0], i);
}
ctx->z_mask = 1;
ctx->s_mask = smask_from_zmask(1);
return false;
}
static bool fold_setcond2(OptContext *ctx, TCGOp *op)
{
TCGCond cond = op->args[5];
int i, inv = 0;
if (swap_commutative2(&op->args[1], &op->args[3])) {
op->args[5] = cond = tcg_swap_cond(cond);
}
i = do_constant_folding_cond2(&op->args[1], &op->args[3], cond);
if (i >= 0) {
goto do_setcond_const;
}
switch (cond) {
case TCG_COND_LT:
case TCG_COND_GE:
/*
* Simplify LT/GE comparisons vs zero to a single compare
* vs the high word of the input.
*/
if (arg_is_const(op->args[3]) && arg_info(op->args[3])->val == 0 &&
arg_is_const(op->args[4]) && arg_info(op->args[4])->val == 0) {
goto do_setcond_high;
}
break;
case TCG_COND_NE:
inv = 1;
QEMU_FALLTHROUGH;
case TCG_COND_EQ:
/*
* Simplify EQ/NE comparisons where one of the pairs
* can be simplified.
*/
i = do_constant_folding_cond(TCG_TYPE_I32, op->args[1],
op->args[3], cond);
switch (i ^ inv) {
case 0:
goto do_setcond_const;
case 1:
goto do_setcond_high;
}
i = do_constant_folding_cond(TCG_TYPE_I32, op->args[2],
op->args[4], cond);
switch (i ^ inv) {
case 0:
goto do_setcond_const;
case 1:
op->args[2] = op->args[3];
op->args[3] = cond;
op->opc = INDEX_op_setcond_i32;
break;
}
break;
default:
break;
do_setcond_high:
op->args[1] = op->args[2];
op->args[2] = op->args[4];
op->args[3] = cond;
op->opc = INDEX_op_setcond_i32;
break;
}
ctx->z_mask = 1;
ctx->s_mask = smask_from_zmask(1);
return false;
do_setcond_const:
return tcg_opt_gen_movi(ctx, op, op->args[0], i);
}
static bool fold_sextract(OptContext *ctx, TCGOp *op)
{
uint64_t z_mask, s_mask, s_mask_old;
int pos = op->args[2];
int len = op->args[3];
if (arg_is_const(op->args[1])) {
uint64_t t;
t = arg_info(op->args[1])->val;
t = sextract64(t, pos, len);
return tcg_opt_gen_movi(ctx, op, op->args[0], t);
}
z_mask = arg_info(op->args[1])->z_mask;
z_mask = sextract64(z_mask, pos, len);
ctx->z_mask = z_mask;
s_mask_old = arg_info(op->args[1])->s_mask;
s_mask = sextract64(s_mask_old, pos, len);
s_mask |= MAKE_64BIT_MASK(len, 64 - len);
ctx->s_mask = s_mask;
if (pos == 0) {
ctx->a_mask = s_mask & ~s_mask_old;
}
return fold_masks(ctx, op);
}
static bool fold_shift(OptContext *ctx, TCGOp *op)
{
uint64_t s_mask, z_mask, sign;
if (fold_const2(ctx, op) ||
fold_ix_to_i(ctx, op, 0) ||
fold_xi_to_x(ctx, op, 0)) {
return true;
}
s_mask = arg_info(op->args[1])->s_mask;
z_mask = arg_info(op->args[1])->z_mask;
if (arg_is_const(op->args[2])) {
int sh = arg_info(op->args[2])->val;
ctx->z_mask = do_constant_folding(op->opc, ctx->type, z_mask, sh);
s_mask = do_constant_folding(op->opc, ctx->type, s_mask, sh);
ctx->s_mask = smask_from_smask(s_mask);
return fold_masks(ctx, op);
}
switch (op->opc) {
CASE_OP_32_64(sar):
/*
* Arithmetic right shift will not reduce the number of
* input sign repetitions.
*/
ctx->s_mask = s_mask;
break;
CASE_OP_32_64(shr):
/*
* If the sign bit is known zero, then logical right shift
* will not reduced the number of input sign repetitions.
*/
sign = (s_mask & -s_mask) >> 1;
if (!(z_mask & sign)) {
ctx->s_mask = s_mask;
}
break;
default:
break;
}
return false;
}
static bool fold_sub_to_neg(OptContext *ctx, TCGOp *op)
{
TCGOpcode neg_op;
bool have_neg;
if (!arg_is_const(op->args[1]) || arg_info(op->args[1])->val != 0) {
return false;
}
switch (ctx->type) {
case TCG_TYPE_I32:
neg_op = INDEX_op_neg_i32;
have_neg = TCG_TARGET_HAS_neg_i32;
break;
case TCG_TYPE_I64:
neg_op = INDEX_op_neg_i64;
have_neg = TCG_TARGET_HAS_neg_i64;
break;
case TCG_TYPE_V64:
case TCG_TYPE_V128:
case TCG_TYPE_V256:
neg_op = INDEX_op_neg_vec;
have_neg = (TCG_TARGET_HAS_neg_vec &&
tcg_can_emit_vec_op(neg_op, ctx->type, TCGOP_VECE(op)) > 0);
break;
default:
g_assert_not_reached();
}
if (have_neg) {
op->opc = neg_op;
op->args[1] = op->args[2];
return fold_neg(ctx, op);
}
return false;
}
/* We cannot as yet do_constant_folding with vectors. */
static bool fold_sub_vec(OptContext *ctx, TCGOp *op)
{
if (fold_xx_to_i(ctx, op, 0) ||
fold_xi_to_x(ctx, op, 0) ||
fold_sub_to_neg(ctx, op)) {
return true;
}
return false;
}
static bool fold_sub(OptContext *ctx, TCGOp *op)
{
return fold_const2(ctx, op) || fold_sub_vec(ctx, op);
}
static bool fold_sub2(OptContext *ctx, TCGOp *op)
{
return fold_addsub2(ctx, op, false);
}
static bool fold_tcg_ld(OptContext *ctx, TCGOp *op)
{
/* We can't do any folding with a load, but we can record bits. */
switch (op->opc) {
CASE_OP_32_64(ld8s):
ctx->s_mask = MAKE_64BIT_MASK(8, 56);
break;
CASE_OP_32_64(ld8u):
ctx->z_mask = MAKE_64BIT_MASK(0, 8);
ctx->s_mask = MAKE_64BIT_MASK(9, 55);
break;
CASE_OP_32_64(ld16s):
ctx->s_mask = MAKE_64BIT_MASK(16, 48);
break;
CASE_OP_32_64(ld16u):
ctx->z_mask = MAKE_64BIT_MASK(0, 16);
ctx->s_mask = MAKE_64BIT_MASK(17, 47);
break;
case INDEX_op_ld32s_i64:
ctx->s_mask = MAKE_64BIT_MASK(32, 32);
break;
case INDEX_op_ld32u_i64:
ctx->z_mask = MAKE_64BIT_MASK(0, 32);
ctx->s_mask = MAKE_64BIT_MASK(33, 31);
break;
default:
g_assert_not_reached();
}
return false;
}
static bool fold_xor(OptContext *ctx, TCGOp *op)
{
if (fold_const2_commutative(ctx, op) ||
fold_xx_to_i(ctx, op, 0) ||
fold_xi_to_x(ctx, op, 0) ||
fold_xi_to_not(ctx, op, -1)) {
return true;
}
ctx->z_mask = arg_info(op->args[1])->z_mask
| arg_info(op->args[2])->z_mask;
ctx->s_mask = arg_info(op->args[1])->s_mask
& arg_info(op->args[2])->s_mask;
return fold_masks(ctx, op);
}
/* Propagate constants and copies, fold constant expressions. */
void tcg_optimize(TCGContext *s)
{
int nb_temps, i;
TCGOp *op, *op_next;
OptContext ctx = { .tcg = s };
/* Array VALS has an element for each temp.
If this temp holds a constant then its value is kept in VALS' element.
If this temp is a copy of other ones then the other copies are
available through the doubly linked circular list. */
nb_temps = s->nb_temps;
for (i = 0; i < nb_temps; ++i) {
s->temps[i].state_ptr = NULL;
}
QTAILQ_FOREACH_SAFE(op, &s->ops, link, op_next) {
TCGOpcode opc = op->opc;
const TCGOpDef *def;
bool done = false;
/* Calls are special. */
if (opc == INDEX_op_call) {
fold_call(&ctx, op);
continue;
}
def = &tcg_op_defs[opc];
init_arguments(&ctx, op, def->nb_oargs + def->nb_iargs);
copy_propagate(&ctx, op, def->nb_oargs, def->nb_iargs);
/* Pre-compute the type of the operation. */
if (def->flags & TCG_OPF_VECTOR) {
ctx.type = TCG_TYPE_V64 + TCGOP_VECL(op);
} else if (def->flags & TCG_OPF_64BIT) {
ctx.type = TCG_TYPE_I64;
} else {
ctx.type = TCG_TYPE_I32;
}
/* Assume all bits affected, no bits known zero, no sign reps. */
ctx.a_mask = -1;
ctx.z_mask = -1;
ctx.s_mask = 0;
optimize: optimize using nonzero bits This adds two optimizations using the non-zero bit mask. In some cases involving shifts or ANDs the value can become zero, and can thus be optimized to a move of zero. Second, useless zero-extension or an AND with constant can be detected that would only zero bits that are already zero. The main advantage of this optimization is that it turns zero-extensions into moves, thus enabling much better copy propagation (around 1% code reduction). Here is for example a "test $0xff0000,%ecx + je" before optimization: mov_i64 tmp0,rcx movi_i64 tmp1,$0xff0000 discard cc_src and_i64 cc_dst,tmp0,tmp1 movi_i32 cc_op,$0x1c ext32u_i64 tmp0,cc_dst movi_i64 tmp12,$0x0 brcond_i64 tmp0,tmp12,eq,$0x0 and after (without patch on the left, with on the right): movi_i64 tmp1,$0xff0000 movi_i64 tmp1,$0xff0000 discard cc_src discard cc_src and_i64 cc_dst,rcx,tmp1 and_i64 cc_dst,rcx,tmp1 movi_i32 cc_op,$0x1c movi_i32 cc_op,$0x1c ext32u_i64 tmp0,cc_dst movi_i64 tmp12,$0x0 movi_i64 tmp12,$0x0 brcond_i64 tmp0,tmp12,eq,$0x0 brcond_i64 cc_dst,tmp12,eq,$0x0 Other similar cases: "test %eax, %eax + jne" where eax is already 32-bit (after optimization, without patch on the left, with on the right): discard cc_src discard cc_src mov_i64 cc_dst,rax mov_i64 cc_dst,rax movi_i32 cc_op,$0x1c movi_i32 cc_op,$0x1c ext32u_i64 tmp0,cc_dst movi_i64 tmp12,$0x0 movi_i64 tmp12,$0x0 brcond_i64 tmp0,tmp12,ne,$0x0 brcond_i64 rax,tmp12,ne,$0x0 "test $0x1, %dl + je": movi_i64 tmp1,$0x1 movi_i64 tmp1,$0x1 discard cc_src discard cc_src and_i64 cc_dst,rdx,tmp1 and_i64 cc_dst,rdx,tmp1 movi_i32 cc_op,$0x1a movi_i32 cc_op,$0x1a ext8u_i64 tmp0,cc_dst movi_i64 tmp12,$0x0 movi_i64 tmp12,$0x0 brcond_i64 tmp0,tmp12,eq,$0x0 brcond_i64 cc_dst,tmp12,eq,$0x0 In some cases TCG even outsmarts GCC. :) Here the input code has "and $0x2,%eax + movslq %eax,%rbx + test %rbx, %rbx" and the optimizer, thanks to copy propagation, does the following: movi_i64 tmp12,$0x2 movi_i64 tmp12,$0x2 and_i64 rax,rax,tmp12 and_i64 rax,rax,tmp12 mov_i64 cc_dst,rax mov_i64 cc_dst,rax ext32s_i64 tmp0,rax -> nop mov_i64 rbx,tmp0 -> mov_i64 rbx,cc_dst and_i64 cc_dst,rbx,rbx -> nop Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2013-01-12 00:42:53 +01:00
/*
* Process each opcode.
* Sorted alphabetically by opcode as much as possible.
*/
switch (opc) {
CASE_OP_32_64(add):
done = fold_add(&ctx, op);
break;
case INDEX_op_add_vec:
done = fold_add_vec(&ctx, op);
break;
CASE_OP_32_64(add2):
done = fold_add2(&ctx, op);
break;
CASE_OP_32_64_VEC(and):
done = fold_and(&ctx, op);
break;
CASE_OP_32_64_VEC(andc):
done = fold_andc(&ctx, op);
break;
CASE_OP_32_64(brcond):
done = fold_brcond(&ctx, op);
break;
case INDEX_op_brcond2_i32:
done = fold_brcond2(&ctx, op);
break;
CASE_OP_32_64(bswap16):
CASE_OP_32_64(bswap32):
case INDEX_op_bswap64_i64:
done = fold_bswap(&ctx, op);
break;
CASE_OP_32_64(clz):
CASE_OP_32_64(ctz):
done = fold_count_zeros(&ctx, op);
break;
CASE_OP_32_64(ctpop):
done = fold_ctpop(&ctx, op);
break;
CASE_OP_32_64(deposit):
done = fold_deposit(&ctx, op);
break;
CASE_OP_32_64(div):
CASE_OP_32_64(divu):
done = fold_divide(&ctx, op);
break;
case INDEX_op_dup_vec:
done = fold_dup(&ctx, op);
break;
case INDEX_op_dup2_vec:
done = fold_dup2(&ctx, op);
break;
CASE_OP_32_64_VEC(eqv):
done = fold_eqv(&ctx, op);
break;
CASE_OP_32_64(extract):
done = fold_extract(&ctx, op);
break;
CASE_OP_32_64(extract2):
done = fold_extract2(&ctx, op);
break;
CASE_OP_32_64(ext8s):
CASE_OP_32_64(ext16s):
case INDEX_op_ext32s_i64:
case INDEX_op_ext_i32_i64:
done = fold_exts(&ctx, op);
break;
CASE_OP_32_64(ext8u):
CASE_OP_32_64(ext16u):
case INDEX_op_ext32u_i64:
case INDEX_op_extu_i32_i64:
case INDEX_op_extrl_i64_i32:
case INDEX_op_extrh_i64_i32:
done = fold_extu(&ctx, op);
break;
CASE_OP_32_64(ld8s):
CASE_OP_32_64(ld8u):
CASE_OP_32_64(ld16s):
CASE_OP_32_64(ld16u):
case INDEX_op_ld32s_i64:
case INDEX_op_ld32u_i64:
done = fold_tcg_ld(&ctx, op);
break;
case INDEX_op_mb:
done = fold_mb(&ctx, op);
break;
CASE_OP_32_64_VEC(mov):
done = fold_mov(&ctx, op);
break;
CASE_OP_32_64(movcond):
done = fold_movcond(&ctx, op);
break;
CASE_OP_32_64(mul):
done = fold_mul(&ctx, op);
break;
CASE_OP_32_64(mulsh):
CASE_OP_32_64(muluh):
done = fold_mul_highpart(&ctx, op);
break;
CASE_OP_32_64(muls2):
CASE_OP_32_64(mulu2):
done = fold_multiply2(&ctx, op);
break;
CASE_OP_32_64_VEC(nand):
done = fold_nand(&ctx, op);
break;
CASE_OP_32_64(neg):
done = fold_neg(&ctx, op);
break;
CASE_OP_32_64_VEC(nor):
done = fold_nor(&ctx, op);
break;
CASE_OP_32_64_VEC(not):
done = fold_not(&ctx, op);
break;
CASE_OP_32_64_VEC(or):
done = fold_or(&ctx, op);
break;
CASE_OP_32_64_VEC(orc):
done = fold_orc(&ctx, op);
break;
case INDEX_op_qemu_ld_i32:
case INDEX_op_qemu_ld_i64:
done = fold_qemu_ld(&ctx, op);
break;
case INDEX_op_qemu_st_i32:
case INDEX_op_qemu_st8_i32:
case INDEX_op_qemu_st_i64:
done = fold_qemu_st(&ctx, op);
break;
CASE_OP_32_64(rem):
CASE_OP_32_64(remu):
done = fold_remainder(&ctx, op);
break;
CASE_OP_32_64(rotl):
CASE_OP_32_64(rotr):
CASE_OP_32_64(sar):
CASE_OP_32_64(shl):
CASE_OP_32_64(shr):
done = fold_shift(&ctx, op);
break;
CASE_OP_32_64(setcond):
done = fold_setcond(&ctx, op);
break;
case INDEX_op_setcond2_i32:
done = fold_setcond2(&ctx, op);
break;
CASE_OP_32_64(sextract):
done = fold_sextract(&ctx, op);
break;
CASE_OP_32_64(sub):
done = fold_sub(&ctx, op);
break;
case INDEX_op_sub_vec:
done = fold_sub_vec(&ctx, op);
break;
CASE_OP_32_64(sub2):
done = fold_sub2(&ctx, op);
break;
CASE_OP_32_64_VEC(xor):
done = fold_xor(&ctx, op);
break;
default:
break;
}
if (!done) {
finish_folding(&ctx, op);
}
}
}