commit 1dc00896840c9823c15089f379776f88a75e9663
parent 0791dd026e3c7e5c49282fe264a075dbcecb0607
Author: Quentin Carbonneaux <quentin.carbonneaux@yale.edu>
Date: Tue, 8 Dec 2015 13:39:17 -0500
wip on instruction selection
Diffstat:
5 files changed, 92 insertions(+), 48 deletions(-)
diff --git a/lisc/FP b/lisc/FP
@@ -8,12 +8,12 @@ Non FP: udiv, srem, urem, store(4), load(7),
-1 bit for s,d 16384
New conversion instructions:
- - Bitcast conversion between fp and ints. cast
+ - Bitcast conversion between fp and ints. bcast
- Convert fp to/from int. ftosi sitof (only signed at first, truncation)
- Convert single to/from double fconv
Summary:
- Add cast, ftosi sitof, fconv
+ Add bcast, ftosi sitof, fconv
Add a 'floating point' bit to instructions
Conversions from/to unsigned types are hard!
I need to find a reference for that.
diff --git a/lisc/Makefile b/lisc/Makefile
@@ -1,6 +1,6 @@
BIN = lisc
# OBJ = main.o util.o parse.o mem.o ssa.o copy.o live.o isel.o spill.o rega.o emit.o
-OBJ = main.o util.o parse.o ssa.o live.o
+OBJ = main.o util.o parse.o ssa.o live.o isel.o
CFLAGS = -Wall -Wextra -std=c99 -g -pedantic
diff --git a/lisc/isel.c b/lisc/isel.c
@@ -46,8 +46,8 @@ noimm(Ref r, Fn *fn)
* address data with 32bits
*/
return 0;
- case CNum:
- val = fn->con[r.val].val;
+ case CBits:
+ val = fn->con[r.val].bits.i;
return (val < INT32_MIN || val > INT32_MAX);
}
}
@@ -60,13 +60,38 @@ rslot(Ref r, Fn *fn)
return fn->tmp[r.val].slot;
}
+static int
+argcls(Ins *i)
+{
+ /* fixme, not correct for some instructions (bcast) */
+ return i->cls;
+}
+
static void
fixargs(Ins *i, Fn *fn)
{
+ Addr a;
Ref r0, r1;
- int s, n;
+ int s, n, k;
+ k = argcls(i);
for (n=0; n<2; n++) {
+ r0 = i->arg[n];
+ /* load floating points from memory
+ * slots, they can't be used as
+ * immediates
+ */
+ if (KBASE(k) == 1 && rtype(r0) == RCon) {
+ r1 = MEM(fn->nmem);
+ i->arg[n] = r1;
+ vgrow(&fn->mem, ++fn->nmem);
+ memset(&a, 0, sizeof a);
+ a.offset.type = CAddr;
+ sprintf(a.offset.label, ".fp%d", r0.val);
+ fn->con[r0.val].emit = 1;
+ fn->mem[fn->nmem-1] = a;
+ continue;
+ }
/* load constants that do not fit in
* a 32bit signed integer into a
* long temporary
@@ -75,7 +100,8 @@ fixargs(Ins *i, Fn *fn)
if (rtype(r0) == RCon && noimm(r0, fn)) {
r1 = newtmp("isel", fn);
i->arg[n] = r1;
- emit(OCopy, 1, r1, r0, R);
+ emit(OCopy, Kl, r1, r0, R);
+ continue;
}
/* load fast locals' addresses into
* temporaries right before the
@@ -85,7 +111,8 @@ fixargs(Ins *i, Fn *fn)
if (s != -1) {
r1 = newtmp("isel", fn);
i->arg[n] = r1;
- emit(OAddr, 1, r1, SLOT(s), R);
+ emit(OAddr, Kl, r1, SLOT(s), R);
+ continue;
}
}
}
@@ -123,7 +150,7 @@ seladdr(Ref *r, ANum *an, Fn *fn)
}
static void
-selcmp(Ref arg[2], int w, Fn *fn)
+selcmp(Ref arg[2], int k, Fn *fn)
{
Ref r;
@@ -133,7 +160,7 @@ selcmp(Ref arg[2], int w, Fn *fn)
arg[0] = r;
}
assert(rtype(arg[0]) != RCon);
- emit(OXCmp, w, R, arg[1], arg[0]);
+ emit(OXCmp, k, R, arg[1], arg[0]);
fixargs(curi, fn);
}
@@ -141,7 +168,7 @@ static void
sel(Ins i, ANum *an, Fn *fn)
{
Ref r0, r1;
- int x, w;
+ int x, k;
int64_t val;
Ins *i0;
@@ -153,7 +180,7 @@ sel(Ins i, ANum *an, Fn *fn)
return;
}
i0 = curi;
- w = i.wide;
+ k = i.cls;
switch (i.op) {
case ODiv:
case ORem:
@@ -161,8 +188,8 @@ sel(Ins i, ANum *an, Fn *fn)
r0 = TMP(RAX), r1 = TMP(RDX);
else
r0 = TMP(RDX), r1 = TMP(RAX);
- emit(OCopy, w, i.to, r0, R);
- emit(OCopy, w, R, r1, R);
+ emit(OCopy, k, i.to, r0, R);
+ emit(OCopy, k, R, r1, R);
if (rtype(i.arg[1]) == RCon) {
/* immediates not allowed for
* divisions in x86
@@ -170,11 +197,11 @@ sel(Ins i, ANum *an, Fn *fn)
r0 = newtmp("isel", fn);
} else
r0 = i.arg[1];
- emit(OXDiv, w, R, r0, R);
- emit(OSign, w, TMP(RDX), TMP(RAX), R);
- emit(OCopy, w, TMP(RAX), i.arg[0], R);
+ emit(OXDiv, k, R, r0, R);
+ emit(OSign, k, TMP(RDX), TMP(RAX), R);
+ emit(OCopy, k, TMP(RAX), i.arg[0], R);
if (rtype(i.arg[1]) == RCon)
- emit(OCopy, w, r0, i.arg[1], R);
+ emit(OCopy, k, r0, i.arg[1], R);
break;
case ONop:
break;
@@ -209,16 +236,17 @@ Emit:
* (rsp = 0) mod 16
*/
if (rtype(i.arg[0]) == RCon) {
- val = fn->con[i.arg[0].val].val;
+ assert(fn->con[i.arg[0].val].type == CBits);
+ val = fn->con[i.arg[0].val].bits.i;
val = (val + 15) & ~INT64_C(15);
if (val < 0 || val > INT32_MAX)
diag("isel: alloc too large");
- emit(OAlloc, 0, i.to, getcon(val, fn), R);
+ emit(OAlloc, Kl, i.to, getcon(val, fn), R);
} else {
/* r0 = (i.arg[0] + 15) & -16 */
r0 = newtmp("isel", fn);
r1 = newtmp("isel", fn);
- emit(OSAlloc, 0, i.to, r0, R);
+ emit(OSAlloc, Kl, i.to, r0, R);
emit(OAnd, 1, r0, r1, getcon(-16, fn));
emit(OAdd, 1, r1, i.arg[0], getcon(15, fn));
}
@@ -232,8 +260,8 @@ Emit:
x = i.op - OCmp;
if (rtype(i.arg[0]) == RCon)
x = COP(x);
- emit(OXSet+x, 0, i.to, R, R);
- selcmp(i.arg, w, fn);
+ emit(OXSet+x, Kw, i.to, R, R);
+ selcmp(i.arg, k, fn);
break;
}
diag("isel: non-exhaustive implementation");
@@ -289,7 +317,7 @@ seljmp(Blk *b, Fn *fn)
b->jmp.type = JRet0;
r = b->jmp.arg;
b->jmp.arg = R;
- emit(OCopy, w, TMP(RAX), r, R);
+ emit(OCopy, w ? Kl : Kw, TMP(RAX), r, R);
return;
case JJnz:;
}
@@ -312,7 +340,7 @@ seljmp(Blk *b, Fn *fn)
b->jmp.type = JXJc + c;
if (fn->tmp[r.val].nuse == 1) {
assert(fn->tmp[r.val].ndef == 1);
- selcmp(fi->arg, fi->wide, fn);
+ selcmp(fi->arg, fi->cls, fn);
*fi = (Ins){.op = ONop};
}
return;
@@ -512,9 +540,9 @@ selcall(Fn *fn, Ins *i0, Ins *i1)
diag("struct-returning function not implemented");
if (stk) {
r = getcon(-(int64_t)stk, fn);
- emit(OSAlloc, 0, R, r, R);
+ emit(OSAlloc, Kl, R, r, R);
}
- emit(OCopy, i1->wide, i1->to, TMP(RAX), R);
+ emit(OCopy, i1->cls, i1->to, TMP(RAX), R);
emit(OCall, 0, R, i1->arg[0], CALL(1 + ci));
for (i=i0, a=ac, ni=0; i<i1; i++, a++) {
@@ -526,17 +554,17 @@ selcall(Fn *fn, Ins *i0, Ins *i1)
if (a->size > 8) {
r = TMP(rsave[ni+1]);
r1 = newtmp("isel", fn);
- emit(OLoad, 1, r, r1, R);
+ emit(OLoadl, Kl, r, r1, R);
r = getcon(8, fn);
- emit(OAdd, 1, r1, i->arg[1], r);
+ emit(OAdd, Kl, r1, i->arg[1], r);
r = TMP(rsave[ni]);
ni += 2;
} else
r = TMP(rsave[ni++]);
- emit(OLoad, 1, r, i->arg[1], R);
+ emit(OLoadl, Kl, r, i->arg[1], R);
} else {
r = TMP(rsave[ni++]);
- emit(OCopy, i->wide, r, i->arg[0], R);
+ emit(OCopy, i->cls, r, i->arg[0], R);
}
}
for (i=i0, a=ac; i<i1; i++, a++) {
@@ -549,12 +577,12 @@ selcall(Fn *fn, Ins *i0, Ins *i1)
if (i->op == OArgc) {
assert(!"argc todo 1");
} else {
- emit(OXPush, 1, R, i->arg[0], R);
+ emit(OXPush, Kl, R, i->arg[0], R);
}
}
if (stk) {
assert(stk == 8);
- emit(OXPush, 1, R, CON_Z, R);
+ emit(OXPush, Kl, R, CON_Z, R);
}
}
@@ -580,7 +608,7 @@ selpar(Fn *fn, Ins *i0, Ins *i1)
continue;
case 2:
stk -= 2;
- *curi++ = (Ins){OLoad, i->wide, i->to, {SLOT(stk)}};
+ *curi++ = (Ins){OLoad, i->to, {SLOT(stk)}, i->cls};
continue;
}
r = TMP(rsave[ni++]);
@@ -588,16 +616,16 @@ selpar(Fn *fn, Ins *i0, Ins *i1)
if (a->rty[0] == RSse || a->rty[1] == RSse)
diag("isel: unsupported float struct");
r1 = newtmp("isel", fn);
- *curi++ = (Ins){OCopy, 1, r1, {r}};
+ *curi++ = (Ins){OCopy, r1, {r}, Kl};
a->rty[0] = r1.val;
if (a->size > 8) {
r = TMP(rsave[ni++]);
r1 = newtmp("isel", fn);
- *curi++ = (Ins){OCopy, 1, r1, {r}};
+ *curi++ = (Ins){OCopy, r1, {r}, Kl};
a->rty[1] = r1.val;
}
} else
- *curi++ = (Ins){OCopy, i->wide, i->to, {r}};
+ *curi++ = (Ins){OCopy, i->to, {r}, i->cls};
}
for (i=i0, a=ac; i<i1; i++, a++) {
if (i->op != OParc || a->inmem)
@@ -608,14 +636,14 @@ selpar(Fn *fn, Ins *i0, Ins *i1)
r1 = i->to;
r = TMP(a->rty[0]);
r2 = getcon(a->size, fn);
- *curi++ = (Ins){OAlloc+al, 1, r1, {r2}};
- *curi++ = (Ins){OStorel, 0, R, {r, r1}};
+ *curi++ = (Ins){OAlloc+al, r1, {r2}, Kl};
+ *curi++ = (Ins){OStorel, R, {r, r1}, 0};
if (a->size > 8) {
r = newtmp("isel", fn);
r2 = getcon(8, fn);
- *curi++ = (Ins){OAdd, 1, r, {r1, r2}};
+ *curi++ = (Ins){OAdd, r, {r1, r2}, Kl};
r1 = TMP(a->rty[1]);
- *curi++ = (Ins){OStorel, 0, R, {r1, r}};
+ *curi++ = (Ins){OStorel, R, {r1, r}, 0};
}
}
}
@@ -640,9 +668,9 @@ ascale(Ref r, Con *con)
if (rtype(r) != RCon)
return 0;
- if (con[r.val].type != CNum)
+ if (con[r.val].type != CBits)
return 0;
- n = con[r.val].val;
+ n = con[r.val].bits.i;
return n == 1 || n == 2 || n == 4 || n == 8;
}
@@ -737,7 +765,7 @@ amatch(Addr *a, Ref r, ANum *ai, Fn *fn, int top)
case 3: /* s * i */
if (!top) {
a->index = al;
- a->scale = fn->con[ar.val].val;
+ a->scale = fn->con[ar.val].bits.i;
} else
a->base = r;
break;
@@ -834,7 +862,7 @@ isel(Fn *fn)
if (i->op == al) {
if (rtype(i->arg[0]) != RCon)
break;
- sz = fn->con[i->arg[0].val].val;
+ sz = fn->con[i->arg[0].val].bits.i;
if (sz < 0 || sz >= INT_MAX-3)
diag("isel: invalid alloc size");
sz = (sz + n-1) & -n;
@@ -855,7 +883,7 @@ isel(Fn *fn)
s = rslot(p->arg[a], fn);
if (s != -1) {
p->arg[a] = newtmp("isel", fn);
- emit(OAddr, 1, p->arg[a], SLOT(s), R);
+ emit(OAddr, Kl, p->arg[a], SLOT(s), R);
}
}
for (m=0; m<n; m++)
diff --git a/lisc/lisc.h b/lisc/lisc.h
@@ -181,18 +181,32 @@ enum Op {
OAnd,
OCmp,
OCmp1 = OCmp + NCmp-1,
+
OStored,
OStores,
OStorel,
OStorew,
OStoreh,
OStoreb,
- OLoad,
- OLoad1 = OLoad + NTy-1,
+
+ OLoadl,
+ OLoadsw,
+ OLoaduw,
+ OLoadsh,
+ OLoaduh,
+ OLoadsb,
+ OLoadub,
+ OLoadd,
+ OLoads,
+ OLoad = OLoadl,
+ OLoad1 = OLoads,
+
OExt,
OExt1 = OExt + NTy-1,
+
OAlloc,
OAlloc1 = OAlloc + NAlign-1,
+
OCopy,
NPubOp,
@@ -321,6 +335,7 @@ struct Con {
double f;
} bits;
char flt;
+ char emit;
};
typedef struct Addr Addr;
diff --git a/lisc/main.c b/lisc/main.c
@@ -56,6 +56,7 @@ func(Fn *fn)
fillpreds(fn);
filluse(fn);
+ isel(fn);
filllive(fn);
#if 0
memopt(fn);
