runtime: set up TLS without cgo on darwin/arm64

Currently, on darwin/arm64 we set up TLS using cgo. TLS is not
set for pure Go programs. As we use libc for syscalls on darwin,
we need to save the G register before the libc call. Otherwise it
is not signal-safe, as a signal may land during the execution of
a libc function, where the G register may be clobbered.

This CL initializes TLS in Go, by calling the pthread functions
directly without cgo. This makes it possible to save the G
register to TLS in pure Go programs (done in a later CL).

Inspired by Elias's CL 209197. Write the logic in Go instead of
assembly.

Updates #38485, #35853.

Change-Id: I257ba2a411ad387b2f4d50d10129d37fec7a226e
Reviewed-on: https://go-review.googlesource.com/c/go/+/265118
Trust: Cherry Zhang <cherryyz@google.com>
Trust: Elias Naur <mail@eliasnaur.com>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

5 files changed, 93 insertions, 36 deletions

diff --git a/src/runtime/asm_arm64.s b/src/runtime/asm_arm64.s
index a45e342478..6257c1a183 100644
--- a/src/runtime/asm_arm64.s
+++ b/src/runtime/asm_arm64.s
@@ -15,6 +15,19 @@ TEXT runtime·rt0_go(SB),NOSPLIT,$0
 	MOVW	R0, 8(RSP) // argc
 	MOVD	R1, 16(RSP) // argv
 
+#ifdef TLS_darwin
+	// Initialize TLS.
+	MOVD	ZR, g // clear g, make sure it's not junk.
+	SUB	$32, RSP
+	MRS_TPIDR_R0
+	AND	$~7, R0
+	MOVD	R0, 16(RSP)             // arg2: TLS base
+	MOVD	$runtime·tls_g(SB), R2
+	MOVD	R2, 8(RSP)              // arg1: &tlsg
+	BL	·tlsinit(SB)
+	ADD	$32, RSP
+#endif
+
 	// create istack out of the given (operating system) stack.
 	// _cgo_init may update stackguard.
 	MOVD	$runtime·g0(SB), g
@@ -29,9 +42,9 @@ TEXT runtime·rt0_go(SB),NOSPLIT,$0
 	MOVD	_cgo_init(SB), R12
 	CBZ	R12, nocgo
 
+#ifdef GOOS_android
 	MRS_TPIDR_R0			// load TLS base pointer
 	MOVD	R0, R3			// arg 3: TLS base pointer
-#ifdef TLSG_IS_VARIABLE
 	MOVD	$runtime·tls_g(SB), R2 	// arg 2: &tls_g
 #else
 	MOVD	$0, R2		        // arg 2: not used when using platform's TLS
diff --git a/src/runtime/cgo/gcc_darwin_arm64.c b/src/runtime/cgo/gcc_darwin_arm64.c
index dbe848b4ee..a5f07f1f1b 100644
--- a/src/runtime/cgo/gcc_darwin_arm64.c
+++ b/src/runtime/cgo/gcc_darwin_arm64.c
@@ -20,37 +20,6 @@
 #include <CoreFoundation/CFString.h>
 #endif
 
-#define magic (0xc476c475c47957UL)
-
-// inittls allocates a thread-local storage slot for g.
-//
-// It finds the first available slot using pthread_key_create and uses
-// it as the offset value for runtime.tlsg.
-static void
-inittls(void **tlsg, void **tlsbase)
-{
-	pthread_key_t k;
-	int i, err;
-
-	err = pthread_key_create(&k, nil);
-	if(err != 0) {
-		fprintf(stderr, "runtime/cgo: pthread_key_create failed: %d\n", err);
-		abort();
-	}
-	//fprintf(stderr, "runtime/cgo: k = %d, tlsbase = %p\n", (int)k, tlsbase); // debug
-	pthread_setspecific(k, (void*)magic);
-	// The first key should be at 257.
-	for (i=0; i<PTHREAD_KEYS_MAX; i++) {
-		if (*(tlsbase+i) == (void*)magic) {
-			*tlsg = (void*)(i*sizeof(void *));
-			pthread_setspecific(k, 0);
-			return;
-		}
-	}
-	fprintf(stderr, "runtime/cgo: could not find pthread key.\n");
-	abort();
-}
-
 static void *threadentry(void*);
 static void (*setg_gcc)(void*);
 
@@ -156,7 +125,7 @@ init_working_dir()
 #endif // TARGET_OS_IPHONE
 
 void
-x_cgo_init(G *g, void (*setg)(void*), void **tlsg, void **tlsbase)
+x_cgo_init(G *g, void (*setg)(void*))
 {
 	pthread_attr_t attr;
 	size_t size;
@@ -168,9 +137,6 @@ x_cgo_init(G *g, void (*setg)(void*), void **tlsg, void **tlsbase)
 	g->stacklo = (uintptr)&attr - size + 4096;
 	pthread_attr_destroy(&attr);
 
-	// yes, tlsbase from mrs might not be correctly aligned.
-	inittls(tlsg, (void**)((uintptr)tlsbase & ~7));
-
 #if TARGET_OS_IPHONE
 	darwin_arm_init_mach_exception_handler();
 	darwin_arm_init_thread_exception_port();
diff --git a/src/runtime/defs_darwin_arm64.go b/src/runtime/defs_darwin_arm64.go
index 2f466045d4..9076e8bd54 100644
--- a/src/runtime/defs_darwin_arm64.go
+++ b/src/runtime/defs_darwin_arm64.go
@@ -94,6 +94,8 @@ const (
 
 	_PTHREAD_CREATE_DETACHED = 0x2
 
+	_PTHREAD_KEYS_MAX = 512
+
 	_F_SETFD    = 0x2
 	_F_GETFL    = 0x3
 	_F_SETFL    = 0x4
@@ -233,3 +235,5 @@ type machTimebaseInfo struct {
 	numer uint32
 	denom uint32
 }
+
+type pthreadkey uint64
diff --git a/src/runtime/sys_darwin_arm64.go b/src/runtime/sys_darwin_arm64.go
new file mode 100644
index 0000000000..9c14f33a1c
--- /dev/null
+++ b/src/runtime/sys_darwin_arm64.go
@@ -0,0 +1,62 @@
+// Copyright 2020 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package runtime
+
+import (
+	"runtime/internal/sys"
+	"unsafe"
+)
+
+// libc function wrappers. Must run on system stack.
+
+//go:nosplit
+//go:cgo_unsafe_args
+func g0_pthread_key_create(k *pthreadkey, destructor uintptr) int32 {
+	return asmcgocall(unsafe.Pointer(funcPC(pthread_key_create_trampoline)), unsafe.Pointer(&k))
+}
+func pthread_key_create_trampoline()
+
+//go:nosplit
+//go:cgo_unsafe_args
+func g0_pthread_setspecific(k pthreadkey, value uintptr) int32 {
+	return asmcgocall(unsafe.Pointer(funcPC(pthread_setspecific_trampoline)), unsafe.Pointer(&k))
+}
+func pthread_setspecific_trampoline()
+
+//go:cgo_import_dynamic libc_pthread_key_create pthread_key_create "/usr/lib/libSystem.B.dylib"
+//go:cgo_import_dynamic libc_pthread_setspecific pthread_setspecific "/usr/lib/libSystem.B.dylib"
+
+// tlsinit allocates a thread-local storage slot for g.
+//
+// It finds the first available slot using pthread_key_create and uses
+// it as the offset value for runtime.tlsg.
+//
+// This runs at startup on g0 stack, but before g is set, so it must
+// not split stack (transitively). g is expected to be nil, so things
+// (e.g. asmcgocall) will skip saving or reading g.
+//
+//go:nosplit
+func tlsinit(tlsg *uintptr, tlsbase *[_PTHREAD_KEYS_MAX]uintptr) {
+	var k pthreadkey
+	err := g0_pthread_key_create(&k, 0)
+	if err != 0 {
+		abort()
+	}
+
+	const magic = 0xc476c475c47957
+	err = g0_pthread_setspecific(k, magic)
+	if err != 0 {
+		abort()
+	}
+
+	for i, x := range tlsbase {
+		if x == magic {
+			*tlsg = uintptr(i * sys.PtrSize)
+			g0_pthread_setspecific(k, 0)
+			return
+		}
+	}
+	abort()
+}
diff --git a/src/runtime/sys_darwin_arm64.s b/src/runtime/sys_darwin_arm64.s
index f8d6f28dc7..31b997df13 100644
--- a/src/runtime/sys_darwin_arm64.s
+++ b/src/runtime/sys_darwin_arm64.s
@@ -497,6 +497,18 @@ TEXT runtime·pthread_kill_trampoline(SB),NOSPLIT,$0
 	BL	libc_pthread_kill(SB)
 	RET
 
+TEXT runtime·pthread_key_create_trampoline(SB),NOSPLIT,$0
+	MOVD	8(R0), R1	// arg 2 destructor
+	MOVD	0(R0), R0	// arg 1 *key
+	BL	libc_pthread_key_create(SB)
+	RET
+
+TEXT runtime·pthread_setspecific_trampoline(SB),NOSPLIT,$0
+	MOVD	8(R0), R1	// arg 2 value
+	MOVD	0(R0), R0	// arg 1 key
+	BL	libc_pthread_setspecific(SB)
+	RET
+
 // syscall calls a function in libc on behalf of the syscall package.
 // syscall takes a pointer to a struct like:
 // struct {