/* Narrowing of power operator or math.pow. */
TRef lj_opt_narrow_pow(jit_State *J, TRef rb, TRef rc, TValue *vc)
{
lua_Number n;
if (tvisstr(vc) && !lj_str_tonum(strV(vc), vc))
lj_trace_err(J, LJ_TRERR_BADTYPE);
n = numV(vc);
/* Limit narrowing for pow to small exponents (or for two constants). */
if ((tref_isk(rc) && tref_isint(rc) && tref_isk(rb)) ||
((J->flags & JIT_F_OPT_NARROW) &&
(numisint(n) && n >= -65536.0 && n <= 65536.0))) {
TRef tmp;
if (!tref_isinteger(rc)) {
if (tref_isstr(rc))
rc = emitir(IRTG(IR_STRTO, IRT_NUM), rc, 0);
rc = emitir(IRTGI(IR_TOINT), rc, IRTOINT_CHECK); /* Guarded TOINT! */
}
if (!tref_isk(rc)) { /* Range guard: -65536 <= i <= 65536 */
tmp = emitir(IRTI(IR_ADD), rc, lj_ir_kint(J, 65536-2147483647-1));
emitir(IRTGI(IR_LE), tmp, lj_ir_kint(J, 2*65536-2147483647-1));
}
return emitir(IRTN(IR_POWI), rb, rc);
}
/* FOLD covers most cases, but some are easier to do here. */
if (tref_isk(rb) && tvispone(ir_knum(IR(tref_ref(rb)))))
return rb; /* 1 ^ x ==> 1 */
rc = lj_ir_tonum(J, rc);
if (tref_isk(rc) && ir_knum(IR(tref_ref(rc)))->n == 0.5)
return emitir(IRTN(IR_FPMATH), rb, IRFPM_SQRT); /* x ^ 0.5 ==> sqrt(x) */
/* Split up b^c into exp2(c*log2(b)). Assembler may rejoin later. */
rb = emitir(IRTN(IR_FPMATH), rb, IRFPM_LOG2);
rc = emitir(IRTN(IR_MUL), rb, rc);
return emitir(IRTN(IR_FPMATH), rc, IRFPM_EXP2);
}
static void LJ_FASTCALL recff_tonumber(jit_State *J, RecordFFData *rd)
{
TRef tr = J->base[0];
TRef base = J->base[1];
if (tr && base) {
base = lj_opt_narrow_toint(J, base);
if (!tref_isk(base) || IR(tref_ref(base))->i != 10)
recff_nyiu(J);
}
if (tref_isnumber_str(tr)) {
if (tref_isstr(tr)) {
TValue tmp;
if (!lj_str_tonum(strV(&rd->argv[0]), &tmp))
recff_nyiu(J); /* Would need an inverted STRTO for this case. */
tr = emitir(IRTG(IR_STRTO, IRT_NUM), tr, 0);
}
#if LJ_HASFFI
} else if (tref_iscdata(tr)) {
lj_crecord_tonumber(J, rd);
return;
#endif
} else {
tr = TREF_NIL;
}
J->base[0] = tr;
UNUSED(rd);
}
/* Add all modified slots to the snapshot. */
static MSize snapshot_slots(jit_State *J, SnapEntry *map, BCReg nslots)
{
IRRef retf = J->chain[IR_RETF]; /* Limits SLOAD restore elimination. */
BCReg s;
MSize n = 0;
for (s = 0; s < nslots; s++) {
TRef tr = J->slot[s];
IRRef ref = tref_ref(tr);
if (ref) {
SnapEntry sn = SNAP_TR(s, tr);
IRIns *ir = &J->cur.ir[ref];
if (!(sn & (SNAP_CONT|SNAP_FRAME)) &&
ir->o == IR_SLOAD && ir->op1 == s && ref > retf) {
/* No need to snapshot unmodified non-inherited slots. */
if (!(ir->op2 & IRSLOAD_INHERIT))
continue;
/* No need to restore readonly slots and unmodified non-parent slots. */
if (!(LJ_DUALNUM && (ir->op2 & IRSLOAD_CONVERT)) &&
(ir->op2 & (IRSLOAD_READONLY|IRSLOAD_PARENT)) != IRSLOAD_PARENT)
sn |= SNAP_NORESTORE;
}
if (LJ_SOFTFP && irt_isnum(ir->t))
sn |= SNAP_SOFTFPNUM;
map[n++] = sn;
}
}
return n;
}
/* Emit or eliminate collected PHIs. */
static void loop_emit_phi(jit_State *J, IRRef1 *subst, IRRef1 *phi, IRRef nphi)
{
int pass2 = 0;
IRRef i, nslots;
IRRef invar = J->chain[IR_LOOP];
/* Pass #1: mark redundant and potentially redundant PHIs. */
for (i = 0; i < nphi; i++) {
IRRef lref = phi[i];
IRRef rref = subst[lref];
if (lref == rref || rref == REF_DROP) { /* Invariants are redundant. */
irt_setmark(IR(lref)->t);
} else if (!(IR(rref)->op1 == lref || IR(rref)->op2 == lref)) {
/* Quick check for simple recurrences failed, need pass2. */
irt_setmark(IR(lref)->t);
pass2 = 1;
}
}
/* Pass #2: traverse variant part and clear marks of non-redundant PHIs. */
if (pass2) {
for (i = J->cur.nins-1; i > invar; i--) {
IRIns *ir = IR(i);
if (!irref_isk(ir->op1)) irt_clearmark(IR(ir->op1)->t);
if (!irref_isk(ir->op2)) irt_clearmark(IR(ir->op2)->t);
}
}
/* Pass #3: add PHIs for variant slots without a corresponding SLOAD. */
nslots = J->baseslot+J->maxslot;
for (i = 1; i < nslots; i++) {
IRRef ref = tref_ref(J->slot[i]);
while (!irref_isk(ref) && ref != subst[ref]) {
IRIns *ir = IR(ref);
irt_clearmark(ir->t); /* Unmark potential uses, too. */
if (irt_isphi(ir->t) || irt_ispri(ir->t))
break;
irt_setphi(ir->t);
if (nphi >= LJ_MAX_PHI)
lj_trace_err(J, LJ_TRERR_PHIOV);
phi[nphi++] = (IRRef1)ref;
ref = subst[ref];
if (ref > invar)
break;
}
}
/* Pass #4: emit PHI instructions or eliminate PHIs. */
for (i = 0; i < nphi; i++) {
IRRef lref = phi[i];
IRIns *ir = IR(lref);
if (!irt_ismarked(ir->t)) { /* Emit PHI if not marked. */
IRRef rref = subst[lref];
if (rref > invar)
irt_setphi(IR(rref)->t);
emitir_raw(IRT(IR_PHI, irt_type(ir->t)), lref, rref);
} else { /* Otherwise eliminate PHI. */
irt_clearmark(ir->t);
irt_clearphi(ir->t);
}
}
}
/* Recursively strip overflow checks. */
static TRef narrow_stripov(jit_State *J, TRef tr, int lastop, IRRef mode)
{
IRRef ref = tref_ref(tr);
IRIns *ir = IR(ref);
int op = ir->o;
if (op >= IR_ADDOV && op <= lastop) {
BPropEntry *bp = narrow_bpc_get(J, ref, mode);
if (bp) {
return TREF(bp->val, irt_t(IR(bp->val)->t));
} else {
IRRef op1 = ir->op1, op2 = ir->op2; /* The IR may be reallocated. */
op1 = narrow_stripov(J, op1, lastop, mode);
op2 = narrow_stripov(J, op2, lastop, mode);
tr = emitir(IRT(op - IR_ADDOV + IR_ADD,
((mode & IRCONV_DSTMASK) >> IRCONV_DSH)), op1, op2);
narrow_bpc_set(J, ref, tref_ref(tr), mode);
}
} else if (LJ_64 && (mode & IRCONV_SEXT) && !irt_is64(ir->t)) {
/* Add all modified slots to the snapshot. */
static void snapshot_slots(jit_State *J, IRRef2 *map, BCReg nslots)
{
BCReg s;
for (s = 0; s < nslots; s++) {
IRRef ref = tref_ref(J->slot[s]);
if (ref) {
IRIns *ir = IR(ref);
if (ir->o == IR_SLOAD && ir->op1 == s && !(ir->op2 & IRSLOAD_INHERIT))
ref = 0;
}
map[s] = (IRRef2)ref;
}
}
/* Reassociate index references. */
static IRRef reassoc_xref(jit_State *J, IRIns *ir)
{
ptrdiff_t ofs = 0;
if (ir->o == IR_ADD && irref_isk(ir->op2)) { /* Get constant offset. */
IRIns *irk = IR(ir->op2);
ofs = (LJ_64 && irk->o == IR_KINT64) ? (ptrdiff_t)ir_k64(irk)->u64 :
(ptrdiff_t)irk->i;
ir = IR(ir->op1);
}
if (ir->o == IR_ADD) { /* Add of base + index. */
/* Index ref > base ref for loop-carried dependences. Only check op1. */
IRIns *ir2, *ir1 = IR(ir->op1);
int32_t shift = 0;
IRRef idxref;
/* Determine index shifts. Don't bother with IR_MUL here. */
if (ir1->o == IR_BSHL && irref_isk(ir1->op2))
shift = IR(ir1->op2)->i;
else if (ir1->o == IR_ADD && ir1->op1 == ir1->op2)
shift = 1;
else
ir1 = ir;
ir2 = IR(ir1->op1);
/* A non-reassociated add. Must be a loop-carried dependence. */
if (ir2->o == IR_ADD && irt_isint(ir2->t) && irref_isk(ir2->op2))
ofs += (ptrdiff_t)IR(ir2->op2)->i << shift;
else
return 0;
idxref = ir2->op1;
/* Try to CSE the reassociated chain. Give up if not found. */
if (ir1 != ir &&
!(idxref = reassoc_trycse(J, ir1->o, idxref,
ir1->o == IR_BSHL ? ir1->op2 : idxref)))
return 0;
if (!(idxref = reassoc_trycse(J, IR_ADD, idxref, ir->op2)))
return 0;
if (ofs != 0) {
IRRef refk = tref_ref(lj_ir_kintp(J, ofs));
if (!(idxref = reassoc_trycse(J, IR_ADD, idxref, refk)))
return 0;
}
return idxref; /* Success, found a reassociated index reference. Phew. */
}
return 0; /* Failure. */
}
/* Narrowing of modulo operator. */
TRef lj_opt_narrow_mod(jit_State *J, TRef rb, TRef rc)
{
TRef tmp;
if ((J->flags & JIT_F_OPT_NARROW) &&
tref_isk(rc) && tref_isint(rc)) { /* Optimize x % k. */
int32_t k = IR(tref_ref(rc))->i;
if (k > 0 && (k & (k-1)) == 0) { /* i % 2^k ==> band(i, 2^k-1) */
if (tref_isinteger(rb))
return emitir(IRTI(IR_BAND), rb, lj_ir_kint(J, k-1));
}
}
/* b % c ==> b - floor(b/c)*c */
rb = lj_ir_tonum(J, rb);
rc = lj_ir_tonum(J, rc);
tmp = emitir(IRTN(IR_DIV), rb, rc);
tmp = emitir(IRTN(IR_FPMATH), tmp, IRFPM_FLOOR);
tmp = emitir(IRTN(IR_MUL), tmp, rc);
return emitir(IRTN(IR_SUB), rb, tmp);
}
static void LJ_FASTCALL recff_io_write(jit_State *J, RecordFFData *rd)
{
TRef fp = recff_io_fp(J, rd->data);
TRef zero = lj_ir_kint(J, 0);
TRef one = lj_ir_kint(J, 1);
ptrdiff_t i = rd->data == 0 ? 1 : 0;
for (; J->base[i]; i++) {
TRef str = lj_ir_tostr(J, J->base[i]);
TRef buf = emitir(IRT(IR_STRREF, IRT_P32), str, zero);
TRef len = emitir(IRTI(IR_FLOAD), str, IRFL_STR_LEN);
if (tref_isk(len) && IR(tref_ref(len))->i == 1) {
TRef tr = emitir(IRT(IR_XLOAD, IRT_U8), buf, IRXLOAD_READONLY);
tr = lj_ir_call(J, IRCALL_fputc, tr, fp);
if (results_wanted(J) != 0) /* Check result only if not ignored. */
emitir(IRTGI(IR_NE), tr, lj_ir_kint(J, -1));
} else {
TRef tr = lj_ir_call(J, IRCALL_fwrite, buf, one, len, fp);
if (results_wanted(J) != 0) /* Check result only if not ignored. */
emitir(IRTGI(IR_EQ), tr, len);
}
}
J->base[0] = TREF_TRUE;
}
/* Unroll loop. */
static void loop_unroll(jit_State *J)
{
IRRef1 phi[LJ_MAX_PHI];
uint32_t nphi = 0;
IRRef1 *subst;
SnapNo onsnap;
SnapShot *osnap, *loopsnap;
SnapEntry *loopmap, *psentinel;
IRRef ins, invar;
/* Use temp buffer for substitution table.
** Only non-constant refs in [REF_BIAS,invar) are valid indexes.
** Caveat: don't call into the VM or run the GC or the buffer may be gone.
*/
invar = J->cur.nins;
subst = (IRRef1 *)lj_str_needbuf(J->L, &G(J->L)->tmpbuf,
(invar-REF_BIAS)*sizeof(IRRef1)) - REF_BIAS;
subst[REF_BASE] = REF_BASE;
/* LOOP separates the pre-roll from the loop body. */
emitir_raw(IRTG(IR_LOOP, IRT_NIL), 0, 0);
/* Grow snapshot buffer and map for copy-substituted snapshots.
** Need up to twice the number of snapshots minus #0 and loop snapshot.
** Need up to twice the number of entries plus fallback substitutions
** from the loop snapshot entries for each new snapshot.
** Caveat: both calls may reallocate J->cur.snap and J->cur.snapmap!
*/
onsnap = J->cur.nsnap;
lj_snap_grow_buf(J, 2*onsnap-2);
lj_snap_grow_map(J, J->cur.nsnapmap*2+(onsnap-2)*J->cur.snap[onsnap-1].nent);
/* The loop snapshot is used for fallback substitutions. */
loopsnap = &J->cur.snap[onsnap-1];
loopmap = &J->cur.snapmap[loopsnap->mapofs];
/* The PC of snapshot #0 and the loop snapshot must match. */
psentinel = &loopmap[loopsnap->nent];
lua_assert(*psentinel == J->cur.snapmap[J->cur.snap[0].nent]);
*psentinel = SNAP(255, 0, 0); /* Replace PC with temporary sentinel. */
/* Start substitution with snapshot #1 (#0 is empty for root traces). */
osnap = &J->cur.snap[1];
/* Copy and substitute all recorded instructions and snapshots. */
for (ins = REF_FIRST; ins < invar; ins++) {
IRIns *ir;
IRRef op1, op2;
if (ins >= osnap->ref) /* Instruction belongs to next snapshot? */
loop_subst_snap(J, osnap++, loopmap, subst); /* Copy-substitute it. */
/* Substitute instruction operands. */
ir = IR(ins);
op1 = ir->op1;
if (!irref_isk(op1)) op1 = subst[op1];
op2 = ir->op2;
if (!irref_isk(op2)) op2 = subst[op2];
if (irm_kind(lj_ir_mode[ir->o]) == IRM_N &&
op1 == ir->op1 && op2 == ir->op2) { /* Regular invariant ins? */
subst[ins] = (IRRef1)ins; /* Shortcut. */
} else {
/* Re-emit substituted instruction to the FOLD/CSE/etc. pipeline. */
IRType1 t = ir->t; /* Get this first, since emitir may invalidate ir. */
IRRef ref = tref_ref(emitir(ir->ot & ~IRT_ISPHI, op1, op2));
subst[ins] = (IRRef1)ref;
if (ref != ins) {
IRIns *irr = IR(ref);
if (ref < invar) { /* Loop-carried dependency? */
/* Potential PHI? */
if (!irref_isk(ref) && !irt_isphi(irr->t) && !irt_ispri(irr->t)) {
irt_setphi(irr->t);
if (nphi >= LJ_MAX_PHI)
lj_trace_err(J, LJ_TRERR_PHIOV);
phi[nphi++] = (IRRef1)ref;
}
/* Check all loop-carried dependencies for type instability. */
if (!irt_sametype(t, irr->t)) {
if (irt_isinteger(t) && irt_isinteger(irr->t))
continue;
else if (irt_isnum(t) && irt_isinteger(irr->t)) /* Fix int->num. */
ref = tref_ref(emitir(IRTN(IR_CONV), ref, IRCONV_NUM_INT));
else if (irt_isnum(irr->t) && irt_isinteger(t)) /* Fix num->int. */
ref = tref_ref(emitir(IRTGI(IR_CONV), ref,
IRCONV_INT_NUM|IRCONV_CHECK));
else
lj_trace_err(J, LJ_TRERR_TYPEINS);
subst[ins] = (IRRef1)ref;
irr = IR(ref);
goto phiconv;
}
} else if (ref != REF_DROP && irr->o == IR_CONV &&
ref > invar && irr->op1 < invar) {
/* May need an extra PHI for a CONV. */
ref = irr->op1;
irr = IR(ref);
phiconv:
if (ref < invar && !irref_isk(ref) && !irt_isphi(irr->t)) {
irt_setphi(irr->t);
if (nphi >= LJ_MAX_PHI)
lj_trace_err(J, LJ_TRERR_PHIOV);
phi[nphi++] = (IRRef1)ref;
}
}
}
}
}
if (!irt_isguard(J->guardemit)) /* Drop redundant snapshot. */
J->cur.nsnapmap = (uint16_t)J->cur.snap[--J->cur.nsnap].mapofs;
lua_assert(J->cur.nsnapmap <= J->sizesnapmap);
*psentinel = J->cur.snapmap[J->cur.snap[0].nent]; /* Restore PC. */
loop_emit_phi(J, subst, phi, nphi, onsnap);
}
/* Emit or eliminate collected PHIs. */
static void loop_emit_phi(jit_State *J, IRRef1 *subst, IRRef1 *phi, IRRef nphi,
SnapNo onsnap)
{
int passx = 0;
IRRef i, j, nslots;
IRRef invar = J->chain[IR_LOOP];
/* Pass #1: mark redundant and potentially redundant PHIs. */
for (i = 0, j = 0; i < nphi; i++) {
IRRef lref = phi[i];
IRRef rref = subst[lref];
if (lref == rref || rref == REF_DROP) { /* Invariants are redundant. */
irt_clearphi(IR(lref)->t);
} else {
phi[j++] = (IRRef1)lref;
if (!(IR(rref)->op1 == lref || IR(rref)->op2 == lref)) {
/* Quick check for simple recurrences failed, need pass2. */
irt_setmark(IR(lref)->t);
passx = 1;
}
}
}
nphi = j;
/* Pass #2: traverse variant part and clear marks of non-redundant PHIs. */
if (passx) {
SnapNo s;
for (i = J->cur.nins-1; i > invar; i--) {
IRIns *ir = IR(i);
if (!irref_isk(ir->op2)) irt_clearmark(IR(ir->op2)->t);
if (!irref_isk(ir->op1)) {
irt_clearmark(IR(ir->op1)->t);
if (ir->op1 < invar &&
ir->o >= IR_CALLN && ir->o <= IR_CARG) { /* ORDER IR */
ir = IR(ir->op1);
while (ir->o == IR_CARG) {
if (!irref_isk(ir->op2)) irt_clearmark(IR(ir->op2)->t);
if (irref_isk(ir->op1)) break;
ir = IR(ir->op1);
irt_clearmark(ir->t);
}
}
}
}
for (s = J->cur.nsnap-1; s >= onsnap; s--) {
SnapShot *snap = &J->cur.snap[s];
SnapEntry *map = &J->cur.snapmap[snap->mapofs];
MSize n, nent = snap->nent;
for (n = 0; n < nent; n++) {
IRRef ref = snap_ref(map[n]);
if (!irref_isk(ref)) irt_clearmark(IR(ref)->t);
}
}
}
/* Pass #3: add PHIs for variant slots without a corresponding SLOAD. */
nslots = J->baseslot+J->maxslot;
for (i = 1; i < nslots; i++) {
IRRef ref = tref_ref(J->slot[i]);
while (!irref_isk(ref) && ref != subst[ref]) {
IRIns *ir = IR(ref);
irt_clearmark(ir->t); /* Unmark potential uses, too. */
if (irt_isphi(ir->t) || irt_ispri(ir->t))
break;
irt_setphi(ir->t);
if (nphi >= LJ_MAX_PHI)
lj_trace_err(J, LJ_TRERR_PHIOV);
phi[nphi++] = (IRRef1)ref;
ref = subst[ref];
if (ref > invar)
break;
}
}
/* Pass #4: propagate non-redundant PHIs. */
while (passx) {
passx = 0;
for (i = 0; i < nphi; i++) {
IRRef lref = phi[i];
IRIns *ir = IR(lref);
if (!irt_ismarked(ir->t)) { /* Propagate only from unmarked PHIs. */
IRIns *irr = IR(subst[lref]);
if (irt_ismarked(irr->t)) { /* Right ref points to other PHI? */
irt_clearmark(irr->t); /* Mark that PHI as non-redundant. */
passx = 1; /* Retry. */
}
}
}
}
/* Pass #5: emit PHI instructions or eliminate PHIs. */
for (i = 0; i < nphi; i++) {
IRRef lref = phi[i];
IRIns *ir = IR(lref);
if (!irt_ismarked(ir->t)) { /* Emit PHI if not marked. */
IRRef rref = subst[lref];
if (rref > invar)
irt_setphi(IR(rref)->t);
emitir_raw(IRT(IR_PHI, irt_type(ir->t)), lref, rref);
} else { /* Otherwise eliminate PHI. */
irt_clearmark(ir->t);
irt_clearphi(ir->t);
}
}
}