runtime: use custom thunks for race calls instead of cgo

Implement custom assembly thunks for hot race calls (memory accesses and function entry/exit).
The thunks extract caller pc, verify that the address is in heap or global and switch to g0 stack.

Before:
ok  	regexp	3.692s
ok  	compress/bzip2	9.461s
ok  	encoding/json	6.380s
After:
ok  	regexp	2.229s (-40%)
ok  	compress/bzip2	4.703s (-50%)
ok  	encoding/json	3.629s (-43%)

For comparison, normal non-race build:
ok  	regexp	0.348s
ok  	compress/bzip2	0.304s
ok  	encoding/json	0.661s
Race build:
ok  	regexp	2.229s (+540%)
ok  	compress/bzip2	4.703s (+1447%)
ok  	encoding/json	3.629s (+449%)

Also removes some race-related special cases from cgocall and scheduler.
In long-term it will allow to remove cyclic runtime/race dependency on cmd/cgo.

Fixes #4249.
Fixes #7460.
Update #6508
Update #6688

R=iant, rsc, bradfitz
CC=golang-codereviews
https://golang.org/cl/55100044

Author: dvyukov

src/pkg/runtime/cgocall.c

@@ -100,11 +100,6 @@ runtime·cgocall(void (*fn)(void*), void *arg)
 	Defer d;
 	SEHUnwind sehunwind;
 
-	if(m->racecall) {
-		runtime·asmcgocall(fn, arg);
-		return;
-	}
-
 	if(!runtime·iscgo && !Solaris && !Windows)
 		runtime·throw("cgocall unavailable");
 
@@ -256,21 +251,9 @@ runtime·cgocallbackg(void)
 		runtime·exit(2);
 	}
 
-	if(m->racecall) {
-		// We were not in syscall, so no need to call runtime·exitsyscall.
-		// However we must set m->locks for the following reason.
-		// Race detector runtime makes __tsan_symbolize cgo callback
-		// holding internal mutexes. The mutexes are not cooperative with Go scheduler.
-		// So if we deschedule a goroutine that holds race detector internal mutex
-		// (e.g. preempt it), another goroutine will deadlock trying to acquire the same mutex.
-		m->locks++;
-		runtime·cgocallbackg1();
-		m->locks--;
-	} else {
-		runtime·exitsyscall();	// coming out of cgo call
-		runtime·cgocallbackg1();
-		runtime·entersyscall();	// going back to cgo call
-	}
+	runtime·exitsyscall();	// coming out of cgo call
+	runtime·cgocallbackg1();
+	runtime·entersyscall();	// going back to cgo call
 }
 
 void
@@ -292,14 +275,14 @@ runtime·cgocallbackg1(void)
 	d.special = true;
 	g->defer = &d;
 
-	if(raceenabled && !m->racecall)
+	if(raceenabled)
 		runtime·raceacquire(&cgosync);
 
 	// Invoke callback.
 	cb = CBARGS;
 	runtime·newstackcall(cb->fn, cb->arg, cb->argsize);
 
-	if(raceenabled && !m->racecall)
+	if(raceenabled)
 		runtime·racereleasemerge(&cgosync);
 
 	// Pop defer.

src/pkg/runtime/malloc.goc

@@ -295,9 +295,6 @@ runtime·free(void *v)
 	if(size < TinySize)
 		runtime·throw("freeing too small block");
 
-	if(raceenabled)
-		runtime·racefree(v);
-
 	// Ensure that the span is swept.
 	// If we free into an unswept span, we will corrupt GC bitmaps.
 	runtime·MSpan_EnsureSwept(s);

src/pkg/runtime/proc.c

@@ -2235,11 +2235,6 @@ runtime·sigprof(uint8 *pc, uint8 *sp, uint8 *lr, G *gp, M *mp)
 	   ((uint8*)runtime·gogo <= pc && pc < (uint8*)runtime·gogo + RuntimeGogoBytes))
 		traceback = false;
 
-	// Race detector calls asmcgocall w/o entersyscall/exitsyscall,
-	// we can not currently unwind through asmcgocall.
-	if(mp != nil && mp->racecall)
-		traceback = false;
-
 	runtime·lock(&prof);
 	if(prof.fn == nil) {
 		runtime·unlock(&prof);