cmd/ld: use native-endian symbol values in symbol table

The Plan 9 symbol table format defines big-endian symbol values
for portability, but we want to be able to generate an ELF object file
and let the host linker link it, as part of the solution to issue 4069.
The symbol table itself, since it is loaded into memory at run time,
must be filled in by the final host linker, using relocation directives
to set the symbol values. On a little-endian machine, the linker will
only fill in little-endian values during relocation, so we are forced
to use little-endian symbol values.

To preserve most of the original portability of the symbol table
format, we make the table itself say whether it uses big- or
little-endian values. If the table begins with the magic sequence
        fe ff ff ff 00 00
then the actual table begins after those six bytes and contains
little-endian symbol values. Otherwise, the table is in the original
format and contains big-endian symbol values. The magic sequence
looks like an "end of table" entry (the fifth byte is zero), so legacy
readers will see a little-endian table as an empty table.

All the gc architectures are little-endian today, so the practical
effect of this CL is to make all the generated tables little-endian,
but if a big-endian system comes along, ld will not generate
the magic sequence, and the various readers will fall back to the
original big-endian interpretation.

R=ken2
CC=golang-dev
https://golang.org/cl/7066043

1 files changed, 55 insertions, 10 deletions

diff --git a/src/cmd/ld/symtab.c b/src/cmd/ld/symtab.c
index a27b181edc..92627f5986 100644
--- a/src/cmd/ld/symtab.c
+++ b/src/cmd/ld/symtab.c
@@ -199,7 +199,7 @@ static void
 slputb(int32 v)
 {
 	uchar *p;
-	
+
 	symgrow(symt, symt->size+4);
 	p = symt->p + symt->size;
 	*p++ = v>>24;
@@ -209,6 +209,22 @@ slputb(int32 v)
 	symt->size += 4;
 }
 
+static void
+slputl(int32 v)
+{
+	uchar *p;
+
+	symgrow(symt, symt->size+4);
+	p = symt->p + symt->size;
+	*p++ = v;
+	*p++ = v>>8;
+	*p++ = v>>16;
+	*p = v>>24;
+	symt->size += 4;
+}
+
+static void (*slput)(int32);
+
 void
 wputl(ushort w)
 {
@@ -269,15 +285,24 @@ putsymb(Sym *s, char *name, int t, vlong v, vlong size, int ver, Sym *typ)
 //		l = 8;
 	if(s != nil) {
 		rel = addrel(symt);
-		rel->siz = l + Rbig;
+		rel->siz = l;
 		rel->sym = s;
 		rel->type = D_ADDR;
 		rel->off = symt->size;
 		v = 0;
-	}	
-	if(l == 8)
-		slputb(v>>32);
-	slputb(v);
+	}
+
+	if(l == 8) {
+		if(slput == slputl) {
+			slputl(v);
+			slputl(v>>32);
+		} else {
+			slputb(v>>32);
+			slputb(v);
+		}
+	} else
+		slput(v);
+
 	if(ver)
 		t += 'a' - 'A';
 	scput(t+0x80);			/* 0x80 is variable length */
@@ -306,8 +331,8 @@ putsymb(Sym *s, char *name, int t, vlong v, vlong size, int ver, Sym *typ)
 		rel->off = symt->size;
 	}
 	if(l == 8)
-		slputb(0);
-	slputb(0);
+		slput(0);
+	slput(0);
 
 	if(debug['n']) {
 		if(t == 'z' || t == 'Z') {
@@ -356,7 +381,7 @@ symtab(void)
 	xdefine("end", SBSS, 0);
 	xdefine("epclntab", SRODATA, 0);
 	xdefine("esymtab", SRODATA, 0);
-	
+
 	// pseudo-symbols to mark locations of type, string, and go string data.
 	s = lookup("type.*", 0);
 	s->type = STYPE;
@@ -372,7 +397,7 @@ symtab(void)
 	symt->type = SSYMTAB;
 	symt->size = 0;
 	symt->reachable = 1;
-	
+
 	// assign specific types so that they sort together.
 	// within a type they sort by size, so the .* symbols
 	// just defined above will be first.
@@ -396,5 +421,25 @@ symtab(void)
 
 	if(debug['s'])
 		return;
+
+	switch(thechar) {
+	default:
+		diag("unknown architecture %c", thechar);
+		errorexit();
+	case '5':
+	case '6':
+	case '8':
+		// magic entry to denote little-endian symbol table
+		slputl(0xfffffffe);
+		scput(0);
+		scput(0);
+		slput = slputl;
+		break;
+	case 'v':
+		// big-endian (in case one comes along)
+		slput = slputb;
+		break;
+	}
+
 	genasmsym(putsymb);
 }