cmd/ld: diagnose Go calling C

For example:
go build -ldflags -C cmd/go 2>&1 | awk '{print $NF}' | sort | uniq -c | sort -nr

LGTM=khr
R=khr, josharian
CC=golang-codereviews
https://codereview.appspot.com/135170044

6 files changed, 75 insertions, 6 deletions

diff --git a/src/cmd/cc/pgen.c b/src/cmd/cc/pgen.c
index 53410a11a..99128d400 100644
--- a/src/cmd/cc/pgen.c
+++ b/src/cmd/cc/pgen.c
@@ -131,6 +131,7 @@ codgen(Node *n, Node *nn)
 	nearln = nn->lineno;
 
 	p = gtext(n1->sym, stkoff);
+	p->from.sym->cfunc = 1;
 	sp = p;
 
 	/*
diff --git a/src/cmd/ld/lib.c b/src/cmd/ld/lib.c
index a68993715..51e10bb99 100644
--- a/src/cmd/ld/lib.c
+++ b/src/cmd/ld/lib.c
@@ -1554,3 +1554,56 @@ diag(char *fmt, ...)
 		errorexit();
 	}
 }
+
+void
+checkgo(void)
+{
+	LSym *s;
+	Reloc *r;
+	int i;
+	int changed;
+	
+	if(!debug['C'])
+		return;
+	
+	// TODO(rsc,khr): Eventually we want to get to no Go-called C functions at all,
+	// which would simplify this logic quite a bit.
+
+	// Mark every Go-called C function with cfunc=2, recursively.
+	do {
+		changed = 0;
+		for(s = ctxt->textp; s != nil; s = s->next) {
+			if(s->cfunc == 0 || (s->cfunc == 2 && s->nosplit)) {
+				for(i=0; i<s->nr; i++) {
+					r = &s->r[i];
+					if(r->sym == nil)
+						continue;
+					if((r->type == R_CALL || r->type == R_CALLARM) && r->sym->type == STEXT) {
+						if(r->sym->cfunc == 1) {
+							changed = 1;
+							r->sym->cfunc = 2;
+						}
+					}
+				}
+			}
+		}
+	}while(changed);
+
+	// Complain about Go-called C functions that can split the stack
+	// (that can be preempted for garbage collection or trigger a stack copy).
+	for(s = ctxt->textp; s != nil; s = s->next) {
+		if(s->cfunc == 0 || (s->cfunc == 2 && s->nosplit)) {
+			for(i=0; i<s->nr; i++) {
+				r = &s->r[i];
+				if(r->sym == nil)
+					continue;
+				if((r->type == R_CALL || r->type == R_CALLARM) && r->sym->type == STEXT) {
+					if(s->cfunc == 0 && r->sym->cfunc == 2 && !r->sym->nosplit)
+						print("Go %s calls C %s\n", s->name, r->sym->name);
+					else if(s->cfunc == 2 && s->nosplit && !r->sym->nosplit)
+						print("Go calls C %s calls %s\n", s->name, r->sym->name);
+				}
+			}
+		}
+	}
+}
diff --git a/src/cmd/ld/lib.h b/src/cmd/ld/lib.h
index dd2399023..067ffa0bc 100644
--- a/src/cmd/ld/lib.h
+++ b/src/cmd/ld/lib.h
@@ -183,6 +183,7 @@ uint16	be16(uchar *b);
 uint32	be32(uchar *b);
 uint64	be64(uchar *b);
 void	callgraph(void);
+void	checkgo(void);
 void	cflush(void);
 void	codeblk(int64 addr, int64 size);
 vlong	cpos(void);
diff --git a/src/cmd/ld/pobj.c b/src/cmd/ld/pobj.c
index d78dacd36..54c5ef247 100644
--- a/src/cmd/ld/pobj.c
+++ b/src/cmd/ld/pobj.c
@@ -71,6 +71,7 @@ main(int argc, char *argv[])
 	if(thechar == '6')
 		flagcount("8", "assume 64-bit addresses", &debug['8']);
 	flagfn1("B", "info: define ELF NT_GNU_BUILD_ID note", addbuildinfo);
+	flagcount("C", "check Go calls to C code", &debug['C']);
 	flagint64("D", "addr: data address", &INITDAT);
 	flagstr("E", "sym: entry symbol", &INITENTRY);
 	if(thechar == '5')
@@ -162,6 +163,7 @@ main(int argc, char *argv[])
 		mark(linklookup(ctxt, "runtime.read_tls_fallback", 0));
 	}
 
+	checkgo();
 	deadcode();
 	callgraph();
 	paramspace = "SP";	/* (FP) now (SP) on output */
diff --git a/src/liblink/objfile.c b/src/liblink/objfile.c
index 22c9199d5..dc463d474 100644
--- a/src/liblink/objfile.c
+++ b/src/liblink/objfile.c
@@ -38,7 +38,8 @@
 //	- type [int]
 //	- name [string]
 //	- version [int]
-//	- dupok [int]
+//	- flags [int]
+//		1 dupok
 //	- size [int]
 //	- gotype [symbol reference]
 //	- p [data block]
@@ -50,7 +51,9 @@
 //	- args [int]
 //	- locals [int]
 //	- nosplit [int]
-//	- leaf [int]
+//	- flags [int]
+//		1 leaf
+//		2 C function
 //	- nlocal [int]
 //	- local [nlocal automatics]
 //	- pcln [pcln table]
@@ -289,6 +292,8 @@ writesym(Link *ctxt, Biobuf *b, LSym *s)
 			Bprint(ctxt->bso, "t=%d ", s->type);
 		if(s->dupok)
 			Bprint(ctxt->bso, "dupok ");
+		if(s->cfunc)
+			Bprint(ctxt->bso, "cfunc ");
 		if(s->nosplit)
 			Bprint(ctxt->bso, "nosplit ");
 		Bprint(ctxt->bso, "size=%lld value=%lld", (vlong)s->size, (vlong)s->value);
@@ -351,7 +356,7 @@ writesym(Link *ctxt, Biobuf *b, LSym *s)
 		wrint(b, s->args);
 		wrint(b, s->locals);
 		wrint(b, s->nosplit);
-		wrint(b, s->leaf);
+		wrint(b, s->leaf | s->cfunc<<1);
 		n = 0;
 		for(a = s->autom; a != nil; a = a->link)
 			n++;
@@ -519,6 +524,7 @@ readsym(Link *ctxt, Biobuf *f, char *pkg, char *pn)
 	if(v != 0 && v != 1)
 		sysfatal("invalid symbol version %d", v);
 	dupok = rdint(f);
+	dupok &= 1;
 	size = rdint(f);
 	
 	if(v != 0)
@@ -573,7 +579,9 @@ readsym(Link *ctxt, Biobuf *f, char *pkg, char *pn)
 		s->args = rdint(f);
 		s->locals = rdint(f);
 		s->nosplit = rdint(f);
-		s->leaf = rdint(f);
+		v = rdint(f);
+		s->leaf = v&1;
+		s->cfunc = v&2;
 		n = rdint(f);
 		for(i=0; i<n; i++) {
 			a = emallocz(sizeof *a);
@@ -629,6 +637,8 @@ readsym(Link *ctxt, Biobuf *f, char *pkg, char *pn)
 			Bprint(ctxt->bso, "t=%d ", s->type);
 		if(s->dupok)
 			Bprint(ctxt->bso, "dupok ");
+		if(s->cfunc)
+			Bprint(ctxt->bso, "cfunc ");
 		if(s->nosplit)
 			Bprint(ctxt->bso, "nosplit ");
 		Bprint(ctxt->bso, "size=%lld value=%lld", (vlong)s->size, (vlong)s->value);
diff --git a/src/pkg/debug/goobj/read.go b/src/pkg/debug/goobj/read.go
index c95fe1e47..79a83e59a 100644
--- a/src/pkg/debug/goobj/read.go
+++ b/src/pkg/debug/goobj/read.go
@@ -602,7 +602,8 @@ func (r *objReader) parseObject(prefix []byte) error {
 		s := &Sym{SymID: r.readSymID()}
 		r.p.Syms = append(r.p.Syms, s)
 		s.Kind = SymKind(typ)
-		s.DupOK = r.readInt() != 0
+		flags := r.readInt()
+		s.DupOK = flags&1 != 0
 		s.Size = r.readInt()
 		s.Type = r.readSymID()
 		s.Data = r.readData()
@@ -623,7 +624,8 @@ func (r *objReader) parseObject(prefix []byte) error {
 			s.Func = f
 			f.Args = r.readInt()
 			f.Frame = r.readInt()
-			f.Leaf = r.readInt() != 0
+			flags := r.readInt()
+			f.Leaf = flags&1 != 0
 			f.NoSplit = r.readInt() != 0
 			f.Var = make([]Var, r.readInt())
 			for i := range f.Var {