runtime: use per-thread profiler for SetCgoTraceback platforms

Updates #35057

Change-Id: I61d772a2cbfb27540fb70c14676c68593076ca94
Reviewed-on: https://go-review.googlesource.com/c/go/+/342054
Run-TryBot: Rhys Hiltner <rhys@justin.tv>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Trust: Michael Knyszek <mknyszek@google.com>

6 files changed, 81 insertions, 18 deletions

diff --git a/src/runtime/os_linux.go b/src/runtime/os_linux.go
index a4646577cb..06773c2193 100644
--- a/src/runtime/os_linux.go
+++ b/src/runtime/os_linux.go
@@ -529,10 +529,8 @@ func signalM(mp *m, sig int) {
 	tgkill(getpid(), int(mp.procid), sig)
 }
 
-// go118UseTimerCreateProfiler enables the per-thread CPU profiler. Platforms
-// with support for SetCgoTraceback do some signal handling in assembly; do not
-// enable it for them until the changes to those code paths are in place.
-const go118UseTimerCreateProfiler = GOARCH != "amd64" && GOARCH != "ppc64le"
+// go118UseTimerCreateProfiler enables the per-thread CPU profiler.
+const go118UseTimerCreateProfiler = true
 
 // validSIGPROF compares this signal delivery's code against the signal sources
 // that the profiler uses, returning whether the delivery should be processed.
diff --git a/src/runtime/signal_unix.go b/src/runtime/signal_unix.go
index 0b32598d54..07f371cefe 100644
--- a/src/runtime/signal_unix.go
+++ b/src/runtime/signal_unix.go
@@ -480,15 +480,26 @@ var sigprofCallersUse uint32
 // and the signal handler collected a stack trace in sigprofCallers.
 // When this is called, sigprofCallersUse will be non-zero.
 // g is nil, and what we can do is very limited.
+//
+// It is called from the signal handling functions written in assembly code that
+// are active for cgo programs, cgoSigtramp and sigprofNonGoWrapper, which have
+// not verified that the SIGPROF delivery corresponds to the best available
+// profiling source for this thread.
+//
 //go:nosplit
 //go:nowritebarrierrec
-func sigprofNonGo() {
+func sigprofNonGo(sig uint32, info *siginfo, ctx unsafe.Pointer) {
 	if prof.hz != 0 {
-		n := 0
-		for n < len(sigprofCallers) && sigprofCallers[n] != 0 {
-			n++
+		c := &sigctxt{info, ctx}
+		// Some platforms (Linux) have per-thread timers, which we use in
+		// combination with the process-wide timer. Avoid double-counting.
+		if validSIGPROF(nil, c) {
+			n := 0
+			for n < len(sigprofCallers) && sigprofCallers[n] != 0 {
+				n++
+			}
+			cpuprof.addNonGo(sigprofCallers[:n])
 		}
-		cpuprof.addNonGo(sigprofCallers[:n])
 	}
 
 	atomic.Store(&sigprofCallersUse, 0)
diff --git a/src/runtime/sys_darwin_amd64.s b/src/runtime/sys_darwin_amd64.s
index 3bd027f982..5d89cda8e6 100644
--- a/src/runtime/sys_darwin_amd64.s
+++ b/src/runtime/sys_darwin_amd64.s
@@ -230,6 +230,23 @@ TEXT runtime·sigtramp(SB),NOSPLIT,$0
 	POP_REGS_HOST_TO_ABI0()
 	RET
 
+// Called using C ABI.
+TEXT runtime·sigprofNonGoWrapper<>(SB),NOSPLIT,$0
+	// Transition from C ABI to Go ABI.
+	PUSH_REGS_HOST_TO_ABI0()
+
+	// Call into the Go signal handler
+	NOP	SP		// disable vet stack checking
+	ADJSP	$24
+	MOVL	DI, 0(SP)	// sig
+	MOVQ	SI, 8(SP)	// info
+	MOVQ	DX, 16(SP)	// ctx
+	CALL	·sigprofNonGo(SB)
+	ADJSP	$-24
+
+	POP_REGS_HOST_TO_ABI0()
+	RET
+
 // Used instead of sigtramp in programs that use cgo.
 // Arguments from kernel are in DI, SI, DX.
 TEXT runtime·cgoSigtramp(SB),NOSPLIT,$0
@@ -297,12 +314,12 @@ sigtrampnog:
 	JNZ	sigtramp  // Skip stack trace if already locked.
 
 	// Jump to the traceback function in runtime/cgo.
-	// It will call back to sigprofNonGo, which will ignore the
-	// arguments passed in registers.
+	// It will call back to sigprofNonGo, via sigprofNonGoWrapper, to convert
+	// the arguments to the Go calling convention.
 	// First three arguments to traceback function are in registers already.
 	MOVQ	runtime·cgoTraceback(SB), CX
 	MOVQ	$runtime·sigprofCallers(SB), R8
-	MOVQ	$runtime·sigprofNonGo(SB), R9
+	MOVQ	$runtime·sigprofNonGoWrapper<>(SB), R9
 	MOVQ	_cgo_callers(SB), AX
 	JMP	AX
 
diff --git a/src/runtime/sys_freebsd_amd64.s b/src/runtime/sys_freebsd_amd64.s
index 71a60cae65..165e97c60d 100644
--- a/src/runtime/sys_freebsd_amd64.s
+++ b/src/runtime/sys_freebsd_amd64.s
@@ -255,6 +255,23 @@ TEXT runtime·sigtramp(SB),NOSPLIT,$0
         POP_REGS_HOST_TO_ABI0()
 	RET
 
+// Called using C ABI.
+TEXT runtime·sigprofNonGoWrapper<>(SB),NOSPLIT,$0
+	// Transition from C ABI to Go ABI.
+	PUSH_REGS_HOST_TO_ABI0()
+
+	// Call into the Go signal handler
+	NOP	SP		// disable vet stack checking
+	ADJSP	$24
+	MOVL	DI, 0(SP)	// sig
+	MOVQ	SI, 8(SP)	// info
+	MOVQ	DX, 16(SP)	// ctx
+	CALL	·sigprofNonGo(SB)
+	ADJSP	$-24
+
+	POP_REGS_HOST_TO_ABI0()
+	RET
+
 // Used instead of sigtramp in programs that use cgo.
 // Arguments from kernel are in DI, SI, DX.
 TEXT runtime·cgoSigtramp(SB),NOSPLIT,$0
@@ -322,12 +339,12 @@ sigtrampnog:
 	JNZ	sigtramp  // Skip stack trace if already locked.
 
 	// Jump to the traceback function in runtime/cgo.
-	// It will call back to sigprofNonGo, which will ignore the
-	// arguments passed in registers.
+	// It will call back to sigprofNonGo, via sigprofNonGoWrapper, to convert
+	// the arguments to the Go calling convention.
 	// First three arguments to traceback function are in registers already.
 	MOVQ	runtime·cgoTraceback(SB), CX
 	MOVQ	$runtime·sigprofCallers(SB), R8
-	MOVQ	$runtime·sigprofNonGo(SB), R9
+	MOVQ	$runtime·sigprofNonGoWrapper<>(SB), R9
 	MOVQ	_cgo_callers(SB), AX
 	JMP	AX
 
diff --git a/src/runtime/sys_linux_amd64.s b/src/runtime/sys_linux_amd64.s
index 345dc90eb0..f0e58e11db 100644
--- a/src/runtime/sys_linux_amd64.s
+++ b/src/runtime/sys_linux_amd64.s
@@ -364,6 +364,23 @@ TEXT runtime·sigtramp(SB),NOSPLIT,$0
         POP_REGS_HOST_TO_ABI0()
 	RET
 
+// Called using C ABI.
+TEXT runtime·sigprofNonGoWrapper<>(SB),NOSPLIT,$0
+	// Transition from C ABI to Go ABI.
+	PUSH_REGS_HOST_TO_ABI0()
+
+	// Call into the Go signal handler
+	NOP	SP		// disable vet stack checking
+	ADJSP	$24
+	MOVL	DI, 0(SP)	// sig
+	MOVQ	SI, 8(SP)	// info
+	MOVQ	DX, 16(SP)	// ctx
+	CALL	·sigprofNonGo(SB)
+	ADJSP	$-24
+
+	POP_REGS_HOST_TO_ABI0()
+	RET
+
 // Used instead of sigtramp in programs that use cgo.
 // Arguments from kernel are in DI, SI, DX.
 TEXT runtime·cgoSigtramp(SB),NOSPLIT,$0
@@ -431,12 +448,12 @@ sigtrampnog:
 	JNZ	sigtramp  // Skip stack trace if already locked.
 
 	// Jump to the traceback function in runtime/cgo.
-	// It will call back to sigprofNonGo, which will ignore the
-	// arguments passed in registers.
+	// It will call back to sigprofNonGo, via sigprofNonGoWrapper, to convert
+	// the arguments to the Go calling convention.
 	// First three arguments to traceback function are in registers already.
 	MOVQ	runtime·cgoTraceback(SB), CX
 	MOVQ	$runtime·sigprofCallers(SB), R8
-	MOVQ	$runtime·sigprofNonGo(SB), R9
+	MOVQ	$runtime·sigprofNonGoWrapper<>(SB), R9
 	MOVQ	_cgo_callers(SB), AX
 	JMP	AX
 
diff --git a/src/runtime/sys_linux_ppc64x.s b/src/runtime/sys_linux_ppc64x.s
index 56d600b6ea..9347afaf19 100644
--- a/src/runtime/sys_linux_ppc64x.s
+++ b/src/runtime/sys_linux_ppc64x.s
@@ -743,6 +743,9 @@ TEXT cgoSigtramp<>(SB),NOSPLIT,$0
 TEXT runtime·sigprofNonGoWrapper<>(SB),NOSPLIT,$0
 	// We're coming from C code, set up essential register, then call sigprofNonGo.
 	CALL	runtime·reginit(SB)
+	MOVW	R3, FIXED_FRAME+0(R1)	// sig
+	MOVD	R4, FIXED_FRAME+8(R1)	// info
+	MOVD	R5, FIXED_FRAME+16(R1)	// ctx
 	CALL	runtime·sigprofNonGo(SB)
 	RET