[release-branch.go1.16] cmd/compile: mark R16, R17 clobbered for non-standard calls on ARM64

On ARM64, (external) linker generated trampoline may clobber R16
and R17. In CL 183842 we change Duff's devices not to use those
registers. However, this is not enough. The register allocator
also needs to know that these registers may be clobbered in any
calls that don't follow the standard Go calling convention. This
include Duff's devices and the write barrier.

Fixes #46928.
Updates #32773.

Change-Id: Ia52a891d9bbb8515c927617dd53aee5af5bd9aa4
Reviewed-on: https://go-review.googlesource.com/c/go/+/184437
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Meng Zhuo <mzh@golangcn.org>
Reviewed-by: Keith Randall <khr@golang.org>
Trust: Meng Zhuo <mzh@golangcn.org>
(cherry picked from commit 11b4aee05bfe83513cf08f83091e5aef8b33e766)
Reviewed-on: https://go-review.googlesource.com/c/go/+/331029
Trust: Cherry Mui <cherryyz@google.com>
Run-TryBot: Cherry Mui <cherryyz@google.com>

2 files changed, 9 insertions, 6 deletions

diff --git a/src/cmd/compile/internal/ssa/gen/ARM64Ops.go b/src/cmd/compile/internal/ssa/gen/ARM64Ops.go
index b0bc9c78ff..4d1d14e18b 100644
--- a/src/cmd/compile/internal/ssa/gen/ARM64Ops.go
+++ b/src/cmd/compile/internal/ssa/gen/ARM64Ops.go
@@ -502,13 +502,14 @@ func init() {
 		// auxint = offset into duffzero code to start executing
 		// returns mem
 		// R20 changed as side effect
+		// R16 and R17 may be clobbered by linker trampoline.
 		{
 			name:      "DUFFZERO",
 			aux:       "Int64",
 			argLength: 2,
 			reg: regInfo{
 				inputs:   []regMask{buildReg("R20")},
-				clobbers: buildReg("R20 R30"),
+				clobbers: buildReg("R16 R17 R20 R30"),
 			},
 			faultOnNilArg0: true,
 			unsafePoint:    true, // FP maintenance around DUFFZERO can be clobbered by interrupts
@@ -542,13 +543,14 @@ func init() {
 		// auxint = offset into duffcopy code to start executing
 		// returns mem
 		// R20, R21 changed as side effect
+		// R16 and R17 may be clobbered by linker trampoline.
 		{
 			name:      "DUFFCOPY",
 			aux:       "Int64",
 			argLength: 3,
 			reg: regInfo{
 				inputs:   []regMask{buildReg("R21"), buildReg("R20")},
-				clobbers: buildReg("R20 R21 R26 R30"),
+				clobbers: buildReg("R16 R17 R20 R21 R26 R30"),
 			},
 			faultOnNilArg0: true,
 			faultOnNilArg1: true,
@@ -707,7 +709,8 @@ func init() {
 		// LoweredWB invokes runtime.gcWriteBarrier. arg0=destptr, arg1=srcptr, arg2=mem, aux=runtime.gcWriteBarrier
 		// It saves all GP registers if necessary,
 		// but clobbers R30 (LR) because it's a call.
-		{name: "LoweredWB", argLength: 3, reg: regInfo{inputs: []regMask{buildReg("R2"), buildReg("R3")}, clobbers: (callerSave &^ gpg) | buildReg("R30")}, clobberFlags: true, aux: "Sym", symEffect: "None"},
+		// R16 and R17 may be clobbered by linker trampoline.
+		{name: "LoweredWB", argLength: 3, reg: regInfo{inputs: []regMask{buildReg("R2"), buildReg("R3")}, clobbers: (callerSave &^ gpg) | buildReg("R16 R17 R30")}, clobberFlags: true, aux: "Sym", symEffect: "None"},
 
 		// There are three of these functions so that they can have three different register inputs.
 		// When we check 0 <= c <= cap (A), then 0 <= b <= c (B), then 0 <= a <= b (C), we want the
diff --git a/src/cmd/compile/internal/ssa/opGen.go b/src/cmd/compile/internal/ssa/opGen.go
index e590f6ba5d..a3f9a22389 100644
--- a/src/cmd/compile/internal/ssa/opGen.go
+++ b/src/cmd/compile/internal/ssa/opGen.go
@@ -20848,7 +20848,7 @@ var opcodeTable = [...]opInfo{
 			inputs: []inputInfo{
 				{0, 1048576}, // R20
 			},
-			clobbers: 537919488, // R20 R30
+			clobbers: 538116096, // R16 R17 R20 R30
 		},
 	},
 	{
@@ -20876,7 +20876,7 @@ var opcodeTable = [...]opInfo{
 				{0, 2097152}, // R21
 				{1, 1048576}, // R20
 			},
-			clobbers: 607125504, // R20 R21 R26 R30
+			clobbers: 607322112, // R16 R17 R20 R21 R26 R30
 		},
 	},
 	{
@@ -21373,7 +21373,7 @@ var opcodeTable = [...]opInfo{
 				{0, 4}, // R2
 				{1, 8}, // R3
 			},
-			clobbers: 9223372035244163072, // R30 F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26 F27 F28 F29 F30 F31
+			clobbers: 9223372035244359680, // R16 R17 R30 F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 F13 F14 F15 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26 F27 F28 F29 F30 F31
 		},
 	},
 	{