runtime: make asmcgocall g0/gsignal checks consistent

In asmcgocall() we need to switch to the g0 stack if we're not already on
the g0 stack or the gsignal stack. The prefered way of doing this is to
check gsignal first, then g0, since if we are going to switch to g0 we will
need g0 handy (thus avoiding a second load).

Rewrite/reorder 386 and amd64 to check gsignal first - this shaves a few
assembly instructions off and makes the order consistent with arm, arm64,
mips64 and ppc64. Add missing gsignal checks to mips, riscv64 and s390x.

Change-Id: I1b027bf393c25e0c33e1d8eb80de67e4a0a3f561
Reviewed-on: https://go-review.googlesource.com/c/go/+/335869
Trust: Joel Sing <joel@sing.id.au>
Run-TryBot: Joel Sing <joel@sing.id.au>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>

9 files changed, 39 insertions, 29 deletions

diff --git a/src/runtime/asm_386.s b/src/runtime/asm_386.s
index 11c60309f4..571aa28a9e 100644
--- a/src/runtime/asm_386.s
+++ b/src/runtime/asm_386.s
@@ -633,18 +633,18 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-12
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	get_tls(CX)
-	MOVL	g(CX), BP
-	CMPL	BP, $0
-	JEQ	nosave	// Don't even have a G yet.
-	MOVL	g_m(BP), BP
-	MOVL	m_g0(BP), SI
 	MOVL	g(CX), DI
-	CMPL	SI, DI
-	JEQ	noswitch
+	CMPL	DI, $0
+	JEQ	nosave	// Don't even have a G yet.
+	MOVL	g_m(DI), BP
 	CMPL	DI, m_gsignal(BP)
 	JEQ	noswitch
+	MOVL	m_g0(BP), SI
+	CMPL	DI, SI
+	JEQ	noswitch
 	CALL	gosave_systemstack_switch<>(SB)
 	get_tls(CX)
 	MOVL	SI, g(CX)
diff --git a/src/runtime/asm_amd64.s b/src/runtime/asm_amd64.s
index 2083ecb53e..a6b321aa42 100644
--- a/src/runtime/asm_amd64.s
+++ b/src/runtime/asm_amd64.s
@@ -667,22 +667,21 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-20
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	get_tls(CX)
-	MOVQ	g(CX), R8
-	CMPQ	R8, $0
-	JEQ	nosave
-	MOVQ	g_m(R8), R8
-	MOVQ	m_g0(R8), SI
 	MOVQ	g(CX), DI
-	CMPQ	SI, DI
+	CMPQ	DI, $0
 	JEQ	nosave
+	MOVQ	g_m(DI), R8
 	MOVQ	m_gsignal(R8), SI
-	CMPQ	SI, DI
+	CMPQ	DI, SI
+	JEQ	nosave
+	MOVQ	m_g0(R8), SI
+	CMPQ	DI, SI
 	JEQ	nosave
 
 	// Switch to system stack.
-	MOVQ	m_g0(R8), SI
 	CALL	gosave_systemstack_switch<>(SB)
 	MOVQ	SI, g(CX)
 	MOVQ	(g_sched+gobuf_sp)(SI), SP
diff --git a/src/runtime/asm_arm.s b/src/runtime/asm_arm.s
index a1164781d2..b47184e36b 100644
--- a/src/runtime/asm_arm.s
+++ b/src/runtime/asm_arm.s
@@ -556,7 +556,8 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-12
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	MOVW	g_m(g), R8
 	MOVW	m_gsignal(R8), R3
 	CMP	R3, g
diff --git a/src/runtime/asm_arm64.s b/src/runtime/asm_arm64.s
index e51ce2f831..8cbd17fa75 100644
--- a/src/runtime/asm_arm64.s
+++ b/src/runtime/asm_arm64.s
@@ -1027,7 +1027,8 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-20
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	MOVD	g_m(g), R8
 	MOVD	m_gsignal(R8), R3
 	CMP	R3, g
diff --git a/src/runtime/asm_mips64x.s b/src/runtime/asm_mips64x.s
index 940a38a0d6..e0e5cbb704 100644
--- a/src/runtime/asm_mips64x.s
+++ b/src/runtime/asm_mips64x.s
@@ -424,7 +424,8 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-20
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	MOVV	g_m(g), R5
 	MOVV	m_gsignal(R5), R6
 	BEQ	R6, g, g0
diff --git a/src/runtime/asm_mipsx.s b/src/runtime/asm_mipsx.s
index 87a1344e8f..1b550719d1 100644
--- a/src/runtime/asm_mipsx.s
+++ b/src/runtime/asm_mipsx.s
@@ -413,8 +413,11 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-12
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	MOVW	g_m(g), R5
+	MOVW	m_gsignal(R5), R6
+	BEQ	R6, g, g0
 	MOVW	m_g0(R5), R6
 	BEQ	R6, g, g0
 
diff --git a/src/runtime/asm_ppc64x.s b/src/runtime/asm_ppc64x.s
index 5dc96c5947..7270abbdee 100644
--- a/src/runtime/asm_ppc64x.s
+++ b/src/runtime/asm_ppc64x.s
@@ -541,9 +541,8 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-20
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
-	// Moreover, if it's called inside the signal handler, it must not switch
-	// to g0 as it can be in use by another syscall.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	MOVD	g_m(g), R8
 	MOVD	m_gsignal(R8), R6
 	CMP	R6, g
diff --git a/src/runtime/asm_riscv64.s b/src/runtime/asm_riscv64.s
index 9927a817f7..d7ab90d1ab 100644
--- a/src/runtime/asm_riscv64.s
+++ b/src/runtime/asm_riscv64.s
@@ -310,8 +310,11 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-20
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	MOV	g_m(g), X6
+	MOV	m_gsignal(X6), X7
+	BEQ	X7, g, g0
 	MOV	m_g0(X6), X7
 	BEQ	X7, g, g0
 
diff --git a/src/runtime/asm_s390x.s b/src/runtime/asm_s390x.s
index d4110d563f..5894fe5783 100644
--- a/src/runtime/asm_s390x.s
+++ b/src/runtime/asm_s390x.s
@@ -513,12 +513,15 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-20
 
 	// Figure out if we need to switch to m->g0 stack.
 	// We get called to create new OS threads too, and those
-	// come in on the m->g0 stack already.
+	// come in on the m->g0 stack already. Or we might already
+	// be on the m->gsignal stack.
 	MOVD	g_m(g), R6
-	MOVD	m_g0(R6), R6
-	CMPBEQ	R6, g, g0
+	MOVD	m_gsignal(R6), R7
+	CMPBEQ	R7, g, g0
+	MOVD	m_g0(R6), R7
+	CMPBEQ	R7, g, g0
 	BL	gosave_systemstack_switch<>(SB)
-	MOVD	R6, g
+	MOVD	R7, g
 	BL	runtime·save_g(SB)
 	MOVD	(g_sched+gobuf_sp)(g), R15