From 30c18878730434027dbefd343aad74963a1fdc48 Mon Sep 17 00:00:00 2001 From: Austin Clements Date: Thu, 1 Oct 2020 17:22:38 -0400 Subject: runtime,cmd/cgo: simplify C -> Go call path This redesigns the way calls work from C to exported Go functions. It removes several steps from the call path, makes cmd/cgo no longer sensitive to the Go calling convention, and eliminates the use of reflectcall from cgo. In order to avoid generating a large amount of FFI glue between the C and Go ABIs, the cgo tool has long depended on generating a C function that marshals the arguments into a struct, and then the actual ABI switch happens in functions with fixed signatures that simply take a pointer to this struct. In a way, this CL simply pushes this idea further. Currently, the cgo tool generates this argument struct in the exact layout of the Go stack frame and depends on reflectcall to unpack it into the appropriate Go call (even though it's actually reflectcall'ing a function generated by cgo). In this CL, we decouple this struct from the Go stack layout. Instead, cgo generates a Go function that takes the struct, unpacks it, and calls the exported function. Since this generated function has a generic signature (like the rest of the call path), we don't need reflectcall and can instead depend on the Go compiler itself to implement the call to the exported Go function. One complication is that syscall.NewCallback on Windows, which converts a Go function into a C function pointer, depends on cgocallback's current dynamic calling approach since the signatures of the callbacks aren't known statically. For this specific case, we continue to depend on reflectcall. Really, the current approach makes some overly simplistic assumptions about translating the C ABI to the Go ABI. Now we're at least in a much better position to do a proper ABI translation. For comparison, the current cgo call path looks like: GoF (generated C function) -> crosscall2 (in cgo/asm_*.s) -> _cgoexp_GoF (generated Go function) -> cgocallback (in asm_*.s) -> cgocallback_gofunc (in asm_*.s) -> cgocallbackg (in cgocall.go) -> cgocallbackg1 (in cgocall.go) -> reflectcall (in asm_*.s) -> _cgoexpwrap_GoF (generated Go function) -> p.GoF Now the call path looks like: GoF (generated C function) -> crosscall2 (in cgo/asm_*.s) -> cgocallback (in asm_*.s) -> cgocallbackg (in cgocall.go) -> cgocallbackg1 (in cgocall.go) -> _cgoexp_GoF (generated Go function) -> p.GoF Notably: 1. We combine _cgoexp_GoF and _cgoexpwrap_GoF and move the combined operation to the end of the sequence. This combined function also handles reflectcall's previous role. 2. We combined cgocallback and cgocallback_gofunc since the only purpose of having both was to convert a raw PC into a Go function value. We instead construct the Go function value in cgocallbackg1. 3. cgocallbackg1 no longer reaches backwards through the stack to get the arguments to cgocallback_gofunc. Instead, we just pass the arguments down. 4. Currently, we need an explicit msanwrite to mark the results struct as written because reflectcall doesn't do this. Now, the results are written by regular Go assignments, so the Go compiler generates the necessary MSAN annotations. This also means we no longer need to track the size of the arguments frame. Updates #40724, since now we don't need to teach cgo about the register ABI or change how it uses reflectcall. Change-Id: I7840489a2597962aeb670e0c1798a16a7359c94f Reviewed-on: https://go-review.googlesource.com/c/go/+/258938 Trust: Austin Clements Run-TryBot: Austin Clements TryBot-Result: Go Bot Reviewed-by: Cherry Zhang --- src/runtime/cgo/asm_arm.s | 61 ++++++++++++++++++++++++----------------------- 1 file changed, 31 insertions(+), 30 deletions(-) (limited to 'src/runtime/cgo/asm_arm.s') diff --git a/src/runtime/cgo/asm_arm.s b/src/runtime/cgo/asm_arm.s index 60132c14a8..ea55e173c1 100644 --- a/src/runtime/cgo/asm_arm.s +++ b/src/runtime/cgo/asm_arm.s @@ -5,51 +5,52 @@ #include "textflag.h" // Called by C code generated by cmd/cgo. -// func crosscall2(fn func(a unsafe.Pointer, n int32, ctxt uintptr), a unsafe.Pointer, n int32, ctxt uintptr) -// Saves C callee-saved registers and calls fn with three arguments. +// func crosscall2(fn, a unsafe.Pointer, n int32, ctxt uintptr) +// Saves C callee-saved registers and calls cgocallback with three arguments. +// fn is the PC of a func(a unsafe.Pointer) function. TEXT crosscall2(SB),NOSPLIT|NOFRAME,$0 - /* - * We still need to save all callee save register as before, and then - * push 3 args for fn (R1, R2, R3). - * Also note that at procedure entry in gc world, 4(R13) will be the - * first arg, so we must push another dummy reg (R0) for 0(R13). - * Additionally, runtime·load_g will clobber R0, so we need to save R0 - * nevertheless. - */ SUB $(8*9), R13 // Reserve space for the floating point registers. - MOVM.WP [R0, R1, R2, R3, R4, R5, R6, R7, R8, R9, g, R11, R12, R14], (R13) + // The C arguments arrive in R0, R1, R2, and R3. We want to + // pass R0, R1, and R3 to Go, so we push those on the stack. + // Also, save C callee-save registers R4-R12. + MOVM.WP [R0, R1, R3, R4, R5, R6, R7, R8, R9, g, R11, R12], (R13) + // Finally, save the link register R14. This also puts the + // arguments we pushed for cgocallback where they need to be, + // starting at 4(R13). + MOVW.W R14, -4(R13) // Skip floating point registers on GOARM < 6. MOVB runtime·goarm(SB), R11 CMP $6, R11 BLT skipfpsave - MOVD F8, (14*4+8*1)(R13) - MOVD F9, (14*4+8*2)(R13) - MOVD F10, (14*4+8*3)(R13) - MOVD F11, (14*4+8*4)(R13) - MOVD F12, (14*4+8*5)(R13) - MOVD F13, (14*4+8*6)(R13) - MOVD F14, (14*4+8*7)(R13) - MOVD F15, (14*4+8*8)(R13) + MOVD F8, (13*4+8*1)(R13) + MOVD F9, (13*4+8*2)(R13) + MOVD F10, (13*4+8*3)(R13) + MOVD F11, (13*4+8*4)(R13) + MOVD F12, (13*4+8*5)(R13) + MOVD F13, (13*4+8*6)(R13) + MOVD F14, (13*4+8*7)(R13) + MOVD F15, (13*4+8*8)(R13) skipfpsave: BL runtime·load_g(SB) - MOVW R15, R14 // R15 is PC. - MOVW 0(R13), R15 + // We set up the arguments to cgocallback when saving registers above. + BL runtime·cgocallback(SB) MOVB runtime·goarm(SB), R11 CMP $6, R11 BLT skipfprest - MOVD (14*4+8*1)(R13), F8 - MOVD (14*4+8*2)(R13), F9 - MOVD (14*4+8*3)(R13), F10 - MOVD (14*4+8*4)(R13), F11 - MOVD (14*4+8*5)(R13), F12 - MOVD (14*4+8*6)(R13), F13 - MOVD (14*4+8*7)(R13), F14 - MOVD (14*4+8*8)(R13), F15 + MOVD (13*4+8*1)(R13), F8 + MOVD (13*4+8*2)(R13), F9 + MOVD (13*4+8*3)(R13), F10 + MOVD (13*4+8*4)(R13), F11 + MOVD (13*4+8*5)(R13), F12 + MOVD (13*4+8*6)(R13), F13 + MOVD (13*4+8*7)(R13), F14 + MOVD (13*4+8*8)(R13), F15 skipfprest: - MOVM.IAW (R13), [R0, R1, R2, R3, R4, R5, R6, R7, R8, R9, g, R11, R12, R14] + MOVW.P 4(R13), R14 + MOVM.IAW (R13), [R0, R1, R3, R4, R5, R6, R7, R8, R9, g, R11, R12] ADD $(8*9), R13 MOVW R14, R15 -- cgit v1.3