prepare for multi-target

This big diff does multiple changes to allow
the addition of new targets to qbe.  The
changes are listed below in decreasing order
of impact.

1. Add a new Target structure.

To add support for a given target, one has to
implement all the members of the Target
structure.  All the source files where changed
to use this interface where needed.

2. Single out amd64-specific code.

In this commit, the amd64 target T_amd64_sysv
is the only target available, it is implemented
in the amd64/ directory.  All the non-static
items in this directory are prefixed with either
amd64_ or amd64_sysv (for items that are
specific to the System V ABI).

3. Centralize Ops information.

There is now a file 'ops.h' that must be used to
store all the available operations together with
their metadata.  The various targets will only
select what they need; but it is beneficial that
there is only *one* place to change to add a new
instruction.

One good side effect of this change is that any
operation 'xyz' in the IL now as a corresponding
'Oxyz' in the code.

4. Misc fixes.

One notable change is that instruction selection
now generates generic comparison operations and
the lowering to the target's comparisons is done
in the emitter.

GAS directives for data are the same for many
targets, so data emission was extracted in a
file 'gas.c'.

5. Modularize the Makefile.

The Makefile now has a list of C files that
are target-independent (SRC), and one list
of C files per target.  Each target can also
use its own 'all.h' header (for example to
define registers).

1 files changed, 31 insertions, 31 deletions

diff --git a/fold.c b/fold.c
index 6129421..55672dd 100644
--- a/fold.c
+++ b/fold.c
@@ -100,7 +100,7 @@ visitins(Ins *i, Fn *fn)
 
 	if (rtype(i->to) != RTmp)
 		return;
-	if (opdesc[i->op].cfold) {
+	if (optab[i->op].canfold) {
 		l = latval(i->arg[0]);
 		if (!req(i->arg[1], R))
 			r = latval(i->arg[1]);
@@ -114,7 +114,7 @@ visitins(Ins *i, Fn *fn)
 			v = opfold(i->op, i->cls, &fn->con[l], &fn->con[r], fn);
 	} else
 		v = Bot;
-	/* fprintf(stderr, "\nvisiting %s (%p)", opdesc[i->op].name, (void *)i); */
+	/* fprintf(stderr, "\nvisiting %s (%p)", optab[i->op].name, (void *)i); */
 	update(i->to.val, v, fn);
 }
 
@@ -360,7 +360,7 @@ foldint(Con *res, int op, int w, Con *cl, Con *cr)
 	else if (cl->type == CAddr || cr->type == CAddr) {
 		if (Ocmpl <= op && op <= Ocmpl1)
 			return 1;
-		err("invalid address operand for '%s'", opdesc[op].name);
+		err("invalid address operand for '%s'", optab[op].name);
 	}
 	switch (op) {
 	case Oadd:  x = l.u + r.u; break;
@@ -397,42 +397,42 @@ foldint(Con *res, int op, int w, Con *cl, Con *cr)
 			} else
 				op -= Ocmpl - Ocmpw;
 			switch (op - Ocmpw) {
-			case ICule: x = l.u <= r.u; break;
-			case ICult: x = l.u < r.u;  break;
-			case ICsle: x = l.s <= r.s; break;
-			case ICslt: x = l.s < r.s;  break;
-			case ICsgt: x = l.s > r.s;  break;
-			case ICsge: x = l.s >= r.s; break;
-			case ICugt: x = l.u > r.u;  break;
-			case ICuge: x = l.u >= r.u; break;
-			case ICeq:  x = l.u == r.u; break;
-			case ICne:  x = l.u != r.u; break;
+			case Ciule: x = l.u <= r.u; break;
+			case Ciult: x = l.u < r.u;  break;
+			case Cisle: x = l.s <= r.s; break;
+			case Cislt: x = l.s < r.s;  break;
+			case Cisgt: x = l.s > r.s;  break;
+			case Cisge: x = l.s >= r.s; break;
+			case Ciugt: x = l.u > r.u;  break;
+			case Ciuge: x = l.u >= r.u; break;
+			case Cieq:  x = l.u == r.u; break;
+			case Cine:  x = l.u != r.u; break;
 			default: die("unreachable");
 			}
 		}
 		else if (Ocmps <= op && op <= Ocmps1) {
 			switch (op - Ocmps) {
-			case FCle: x = l.fs <= r.fs; break;
-			case FClt: x = l.fs < r.fs;  break;
-			case FCgt: x = l.fs > r.fs;  break;
-			case FCge: x = l.fs >= r.fs; break;
-			case FCne: x = l.fs != r.fs; break;
-			case FCeq: x = l.fs == r.fs; break;
-			case FCo: x = l.fs < r.fs || l.fs >= r.fs; break;
-			case FCuo: x = !(l.fs < r.fs || l.fs >= r.fs); break;
+			case Cfle: x = l.fs <= r.fs; break;
+			case Cflt: x = l.fs < r.fs;  break;
+			case Cfgt: x = l.fs > r.fs;  break;
+			case Cfge: x = l.fs >= r.fs; break;
+			case Cfne: x = l.fs != r.fs; break;
+			case Cfeq: x = l.fs == r.fs; break;
+			case Cfo: x = l.fs < r.fs || l.fs >= r.fs; break;
+			case Cfuo: x = !(l.fs < r.fs || l.fs >= r.fs); break;
 			default: die("unreachable");
 			}
 		}
 		else if (Ocmpd <= op && op <= Ocmpd1) {
 			switch (op - Ocmpd) {
-			case FCle: x = l.fd <= r.fd; break;
-			case FClt: x = l.fd < r.fd;  break;
-			case FCgt: x = l.fd > r.fd;  break;
-			case FCge: x = l.fd >= r.fd; break;
-			case FCne: x = l.fd != r.fd; break;
-			case FCeq: x = l.fd == r.fd; break;
-			case FCo: x = l.fd < r.fd || l.fd >= r.fd; break;
-			case FCuo: x = !(l.fd < r.fd || l.fd >= r.fd); break;
+			case Cfle: x = l.fd <= r.fd; break;
+			case Cflt: x = l.fd < r.fd;  break;
+			case Cfgt: x = l.fd > r.fd;  break;
+			case Cfge: x = l.fd >= r.fd; break;
+			case Cfne: x = l.fd != r.fd; break;
+			case Cfeq: x = l.fd == r.fd; break;
+			case Cfo: x = l.fd < r.fd || l.fd >= r.fd; break;
+			case Cfuo: x = !(l.fd < r.fd || l.fd >= r.fd); break;
 			default: die("unreachable");
 			}
 		}
@@ -453,7 +453,7 @@ foldflt(Con *res, int op, int w, Con *cl, Con *cr)
 	double xd, ld, rd;
 
 	if (cl->type != CBits || cr->type != CBits)
-		err("invalid address operand for '%s'", opdesc[op].name);
+		err("invalid address operand for '%s'", optab[op].name);
 	if (w)  {
 		ld = cl->bits.d;
 		rd = cr->bits.d;
@@ -495,7 +495,7 @@ opfold(int op, int cls, Con *cl, Con *cr, Fn *fn)
 
 	if ((op == Odiv || op == Oudiv
 	|| op == Orem || op == Ourem) && czero(cr, KWIDE(cls)))
-		err("null divisor in '%s'", opdesc[op].name);
+		err("null divisor in '%s'", optab[op].name);
 	if (cls == Kw || cls == Kl) {
 		if (foldint(&c, op, cls == Kl, cl, cr))
 			return Bot;