Initial commit: Go 1.23 release state

src/runtime/libfuzzer_amd64.s

@@ -0,0 +1,158 @@
+// Copyright 2019 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.
+
+//go:build libfuzzer
+
+#include "go_asm.h"
+#include "go_tls.h"
+#include "textflag.h"
+
+// Based on race_amd64.s; see commentary there.
+
+#ifdef GOOS_windows
+#define RARG0 CX
+#define RARG1 DX
+#define RARG2 R8
+#define RARG3 R9
+#else
+#define RARG0 DI
+#define RARG1 SI
+#define RARG2 DX
+#define RARG3 CX
+#endif
+
+// void runtime·libfuzzerCall4(fn, hookId int, s1, s2 unsafe.Pointer, result uintptr)
+// Calls C function fn from libFuzzer and passes 4 arguments to it.
+TEXT	runtime·libfuzzerCall4(SB), NOSPLIT, $0-40
+	MOVQ	fn+0(FP), AX
+	MOVQ	hookId+8(FP), RARG0
+	MOVQ	s1+16(FP), RARG1
+	MOVQ	s2+24(FP), RARG2
+	MOVQ	result+32(FP), RARG3
+
+	get_tls(R12)
+	MOVQ	g(R12), R14
+	MOVQ	g_m(R14), R13
+
+	// Switch to g0 stack.
+	MOVQ	SP, R12		// callee-saved, preserved across the CALL
+	MOVQ	m_g0(R13), R10
+	CMPQ	R10, R14
+	JE	call	// already on g0
+	MOVQ	(g_sched+gobuf_sp)(R10), SP
+call:
+	ANDQ	$~15, SP	// alignment for gcc ABI
+	CALL	AX
+	MOVQ	R12, SP
+	RET
+
+// void runtime·libfuzzerCallTraceIntCmp(fn, arg0, arg1, fakePC uintptr)
+// Calls C function fn from libFuzzer and passes 2 arguments to it after
+// manipulating the return address so that libfuzzer's integer compare hooks
+// work
+// libFuzzer's compare hooks obtain the caller's address from the compiler
+// builtin __builtin_return_address. Since we invoke the hooks always
+// from the same native function, this builtin would always return the same
+// value. Internally, the libFuzzer hooks call through to the always inlined
+// HandleCmp and thus can't be mimicked without patching libFuzzer.
+//
+// We solve this problem via an inline assembly trampoline construction that
+// translates a runtime argument `fake_pc` in the range [0, 512) into a call to
+// a hook with a fake return address whose lower 9 bits are `fake_pc` up to a
+// constant shift. This is achieved by pushing a return address pointing into
+// 512 ret instructions at offset `fake_pc` onto the stack and then jumping
+// directly to the address of the hook.
+//
+// Note: We only set the lowest 9 bits of the return address since only these
+// bits are used by the libFuzzer value profiling mode for integer compares, see
+// https://github.com/llvm/llvm-project/blob/704d92607d26e696daba596b72cb70effe79a872/compiler-rt/lib/fuzzer/FuzzerTracePC.cpp#L390
+// as well as
+// https://github.com/llvm/llvm-project/blob/704d92607d26e696daba596b72cb70effe79a872/compiler-rt/lib/fuzzer/FuzzerValueBitMap.h#L34
+// ValueProfileMap.AddValue() truncates its argument to 16 bits and shifts the
+// PC to the left by log_2(128)=7, which means that only the lowest 16 - 7 bits
+// of the return address matter. String compare hooks use the lowest 12 bits,
+// but take the return address as an argument and thus don't require the
+// indirection through a trampoline.
+// TODO: Remove the inline assembly trampoline once a PC argument has been added to libfuzzer's int compare hooks.
+TEXT	runtime·libfuzzerCallTraceIntCmp(SB), NOSPLIT, $0-32
+	MOVQ	fn+0(FP), AX
+	MOVQ	arg0+8(FP), RARG0
+	MOVQ	arg1+16(FP), RARG1
+	MOVQ	fakePC+24(FP), R8
+
+	get_tls(R12)
+	MOVQ	g(R12), R14
+	MOVQ	g_m(R14), R13
+
+	// Switch to g0 stack.
+	MOVQ	SP, R12		// callee-saved, preserved across the CALL
+	MOVQ	m_g0(R13), R10
+	CMPQ	R10, R14
+	JE	call	// already on g0
+	MOVQ	(g_sched+gobuf_sp)(R10), SP
+call:
+	ANDQ	$~15, SP	// alignment for gcc ABI
+	SUBQ	$8, SP
+	// Load the address of the end of the function and push it into the stack.
+	// This address will be jumped to after executing the return instruction
+	// from the return sled. There we reset the stack pointer and return.
+	MOVQ    $end_of_function<>(SB), BX
+	PUSHQ   BX
+	// Load the starting address of the return sled into BX.
+	MOVQ    $ret_sled<>(SB), BX
+	// Load the address of the i'th return instruction from the return sled.
+	// The index is given in the fakePC argument.
+	ADDQ    R8, BX
+	PUSHQ   BX
+	// Call the original function with the fakePC return address on the stack.
+	// Function arguments arg0 and arg1 are passed in the registers specified
+	// by the x64 calling convention.
+	JMP     AX
+// This code will not be executed and is only there to satisfy assembler
+// check of a balanced stack.
+not_reachable:
+	POPQ    BX
+	POPQ    BX
+	RET
+
+TEXT end_of_function<>(SB), NOSPLIT, $0-0
+	MOVQ	R12, SP
+	RET
+
+#define REPEAT_8(a) a \
+  a \
+  a \
+  a \
+  a \
+  a \
+  a \
+  a
+
+#define REPEAT_512(a) REPEAT_8(REPEAT_8(REPEAT_8(a)))
+
+TEXT ret_sled<>(SB), NOSPLIT, $0-0
+	REPEAT_512(RET)
+
+// void runtime·libfuzzerCallWithTwoByteBuffers(fn, start, end *byte)
+// Calls C function fn from libFuzzer and passes 2 arguments of type *byte to it.
+TEXT	runtime·libfuzzerCallWithTwoByteBuffers(SB), NOSPLIT, $0-24
+	MOVQ	fn+0(FP), AX
+	MOVQ	start+8(FP), RARG0
+	MOVQ	end+16(FP), RARG1
+
+	get_tls(R12)
+	MOVQ	g(R12), R14
+	MOVQ	g_m(R14), R13
+
+	// Switch to g0 stack.
+	MOVQ	SP, R12		// callee-saved, preserved across the CALL
+	MOVQ	m_g0(R13), R10
+	CMPQ	R10, R14
+	JE	call	// already on g0
+	MOVQ	(g_sched+gobuf_sp)(R10), SP
+call:
+	ANDQ	$~15, SP	// alignment for gcc ABI
+	CALL	AX
+	MOVQ	R12, SP
+	RET