5l, 6l, 8l: omit symbols for type, string, go.string

Much of the bulk of Go binaries is the symbol tables,
which give a name to every C string, Go string,
and reflection type symbol.  These names are not worth
much other than seeing what's where in a binary.

This CL deletes all those names from the symbol table,
instead aggregating the symbols into contiguous blocks
and giving them the names "string.*", "go.string.*", and "type.*".

Before:
$ 6nm $(which godoc.old) | sort | grep ' string\.' | tail -10
  59eda4 D string."aa87ca22be8b05378eb1c71...
  59ee08 D string."b3312fa7e23ee7e4988e056...
  59ee6c D string."func(*token.FileSet, st...
  59eed0 D string."func(io.Writer, []uint8...
  59ef34 D string."func(*tls.Config, *tls....
  59ef98 D string."func(*bool, **template....
  59effc D string."method(p *printer.print...
  59f060 D string."method(S *scanner.Scann...
  59f12c D string."func(*struct { begin in...
  59f194 D string."method(ka *tls.ecdheRSA...
$

After:
$ 6nm $(which godoc) | sort | grep ' string\.' | tail -10
  5e6a30 D string.*
$

Those names in the "Before" are truncated for the CL.
In the real binary they are the complete string, up to
a certain length, or else a unique identifier.
The same applies to the type and go.string symbols.

Removing the names cuts godoc by more than half:

-rwxr-xr-x 1 rsc rsc 9153405 2011-03-07 23:19 godoc.old
-rwxr-xr-x 1 rsc rsc 4290071 2011-03-07 23:19 godoc

For what it's worth, only 80% of what's left gets loaded
into memory; the other 20% is dwarf debugging information
only ever accessed by gdb:

-rwxr-xr-x 1 rsc rsc 3397787 2011-03-07 23:19 godoc.nodwarf

R=r, cw
CC=golang-dev
https://golang.org/cl/4245072

src/cmd/5l/asm.c

@@ -1978,6 +1978,8 @@ genasmsym(void (*put)(Sym*, char*, int, vlong, vlong, int, Sym*))
 
 	for(h=0; h<NHASH; h++) {
 		for(s=hash[h]; s!=S; s=s->hash) {
+			if(s->hide)
+				continue;
 			switch(s->type) {
 			case SCONST:
 			case SRODATA:

src/cmd/5l/l.h

@@ -136,6 +136,7 @@ struct	Sym
 	uchar	dynexport;
 	uchar	leaf;
 	uchar	stkcheck;
+	uchar	hide;
 	int32	dynid;
 	int32	plt;
 	int32	got;
@@ -202,22 +203,6 @@ struct	Count
 
 enum
 {
-	Sxxx,
-	
-	/* order here is order in output file */
-	STEXT		= 1,
-	SRODATA,
-	SELFDATA,
-	SDATA,
-	SBSS,
-
-	SXREF,
-	SFILE,
-	SCONST,
-	SDYNIMPORT,
-
-	SSUB	= 1<<8,
-
 	LFROM		= 1<<0,
 	LTO		= 1<<1,
 	LPOOL		= 1<<2,

src/cmd/6l/asm.c

@@ -1105,12 +1105,17 @@ genasmsym(void (*put)(Sym*, char*, int, vlong, vlong, int, Sym*))
 
 	for(h=0; h<NHASH; h++) {
 		for(s=hash[h]; s!=S; s=s->hash) {
+			if(s->hide)
+				continue;
 			switch(s->type&~SSUB) {
 			case SCONST:
 			case SRODATA:
 			case SDATA:
 			case SELFDATA:
 			case SMACHOGOT:
+			case STYPE:
+			case SSTRING:
+			case SGOSTRING:
 			case SWINDOWS:
 				if(!s->reachable)
 					continue;

src/cmd/6l/l.h

@@ -132,6 +132,7 @@ struct	Sym
 	uchar	dynexport;
 	uchar	special;
 	uchar	stkcheck;
+	uchar	hide;
 	int32	dynid;
 	int32	sig;
 	int32	plt;
@@ -177,28 +178,6 @@ struct	Movtab
 
 enum
 {
-	Sxxx,
-	
-	/* order here is order in output file */
-	STEXT		= 1,
-	SELFDATA,
-	SMACHOPLT,
-	SRODATA,
-	SDATA,
-	SMACHOGOT,
-	SWINDOWS,
-	SBSS,
-
-	SXREF,
-	SMACHODYNSTR,
-	SMACHODYNSYM,
-	SMACHOINDIRECTPLT,
-	SMACHOINDIRECTGOT,
-	SFILE,
-	SCONST,
-	SDYNIMPORT,
-	SSUB	= 1<<8,
-
 	NHASH		= 10007,
 	MINSIZ		= 8,
 	STRINGSZ	= 200,

src/cmd/8l/asm.c

@@ -1158,6 +1158,8 @@ genasmsym(void (*put)(Sym*, char*, int, vlong, vlong, int, Sym*))
 
 	for(h=0; h<NHASH; h++) {
 		for(s=hash[h]; s!=S; s=s->hash) {
+			if(s->hide)
+				continue;
 			switch(s->type&~SSUB) {
 			case SCONST:
 			case SRODATA:

src/cmd/8l/l.h

@@ -131,6 +131,7 @@ struct	Sym
 	uchar	dynexport;
 	uchar	special;
 	uchar	stkcheck;
+	uchar	hide;
 	int32	value;
 	int32	size;
 	int32	sig;
@@ -168,30 +169,6 @@ struct	Optab
 
 enum
 {
-	Sxxx,
-	
-	/* order here is order in output file */
-	STEXT,
-	SELFDATA,
-	SMACHOPLT,
-	SRODATA,
-	SDATA,
-	SMACHO,	/* Mach-O __nl_symbol_ptr */
-	SMACHOGOT,
-	SWINDOWS,
-	SBSS,
-
-	SXREF,
-	SMACHODYNSTR,
-	SMACHODYNSYM,
-	SMACHOINDIRECTPLT,
-	SMACHOINDIRECTGOT,
-	SFILE,
-	SCONST,
-	SDYNIMPORT,
-
-	SSUB = 1<<8,	/* sub-symbol, linked from parent via ->sub list */
-
 	NHASH		= 10007,
 	MINSIZ		= 4,
 	STRINGSZ	= 200,

src/cmd/ld/data.c

@@ -732,6 +732,7 @@ dodata(void)
 
 	last = nil;
 	datap = nil;
+
 	for(h=0; h<NHASH; h++) {
 		for(s=hash[h]; s!=S; s=s->hash){
 			if(!s->reachable || s->special)
@@ -786,7 +787,7 @@ dodata(void)
 	s = datap;
 	for(; s != nil && s->type < SDATA; s = s->next) {
 		s->type = SRODATA;
-		t = rnd(s->size, 4);
+		t = rnd(s->size, PtrSize);
 		s->size = t;
 		s->value = datsize;
 		datsize += t;

src/cmd/ld/dwarf.c

@@ -1467,7 +1467,7 @@ defdwsymb(Sym* sym, char *s, int t, vlong v, vlong size, int ver, Sym *gotype)
 	if (strncmp(s, "type._.", 7) == 0)
 		return;
 
-	if (strncmp(s, "type.", 5) == 0) {
+	if (strncmp(s, "type.", 5) == 0 && strcmp(s, "type.*") != 0) {
 		defgotype(sym);
 		return;
 	}

src/cmd/ld/go.c

@@ -662,6 +662,7 @@ deadcode(void)
 		if(strncmp(s->name, "weak.", 5) == 0) {
 			s->special = 1;  // do not lay out in data segment
 			s->reachable = 1;
+			s->hide = 1;
 		}
 }
 

src/cmd/ld/lib.h

@@ -28,8 +28,35 @@
 // OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 // THE SOFTWARE.
 
-// Where symbol table data gets mapped into memory.
+enum
-#define SYMDATVA 0x99LL<<24
+{
+	Sxxx,
+	
+	/* order here is order in output file */
+	STEXT,
+	SELFDATA,
+	SMACHOPLT,
+	STYPE,
+	SSTRING,
+	SGOSTRING,
+	SRODATA,
+	SDATA,
+	SMACHO,	/* Mach-O __nl_symbol_ptr */
+	SMACHOGOT,
+	SWINDOWS,
+	SBSS,
+
+	SXREF,
+	SMACHODYNSTR,
+	SMACHODYNSYM,
+	SMACHOINDIRECTPLT,
+	SMACHOINDIRECTGOT,
+	SFILE,
+	SCONST,
+	SDYNIMPORT,
+
+	SSUB = 1<<8,	/* sub-symbol, linked from parent via ->sub list */
+};
 
 typedef struct Library Library;
 struct Library

src/cmd/ld/symtab.c

@@ -340,6 +340,9 @@ putsymb(Sym *s, char *name, int t, vlong v, vlong size, int ver, Sym *typ)
 void
 symtab(void)
 {
+	int32 h;
+	Sym *s;
+
 	// Define these so that they'll get put into the symbol table.
 	// data.c:/^address will provide the actual values.
 	xdefine("text", STEXT, 0);
@@ -351,11 +354,50 @@ 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;
+	s->size = 0;
+	s->reachable = 1;
+
+	s = lookup("string.*", 0);
+	s->type = SSTRING;
+	s->size = 0;
+	s->reachable = 1;
+
+	s = lookup("go.string.*", 0);
+	s->type = SGOSTRING;
+	s->size = 0;
+	s->reachable = 1;
 
 	symt = lookup("symtab", 0);
 	symt->type = SRODATA;
 	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.
+	// hide the specific symbols.
+	for(h=0; h<NHASH; h++) {
+		for(s=hash[h]; s!=S; s=s->hash){
+			if(!s->reachable || s->special || s->type != SRODATA)
+				continue;
+			if(strncmp(s->name, "type.", 5) == 0) {
+				s->type = STYPE;
+				s->hide = 1;
+			}
+			if(strncmp(s->name, "string.", 7) == 0) {
+				s->type = SSTRING;
+				s->hide = 1;
+			}
+			if(strncmp(s->name, "go.string.", 10) == 0) {
+				s->type = SGOSTRING;
+				s->hide = 1;
+			}
+		}
+	}
 
 	genasmsym(putsymb);
 }