ARM: rethook: Add rethook arm implementation

Add rethook arm implementation. Most of the code has been copied from
kretprobes on arm.
Since the arm's ftrace implementation is a bit special, this needs a
special care using from fprobe.

Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Tested-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/164735289643.1084943.15184590256680485720.stgit@devnote2

arch/arm/Kconfig

@@ -107,6 +107,7 @@ config ARM
 	select HAVE_MOD_ARCH_SPECIFIC
 	select HAVE_NMI
 	select HAVE_OPTPROBES if !THUMB2_KERNEL
+	select HAVE_RETHOOK
 	select HAVE_PERF_EVENTS
 	select HAVE_PERF_REGS
 	select HAVE_PERF_USER_STACK_DUMP

arch/arm/include/asm/stacktrace.h

@@ -14,7 +14,7 @@ struct stackframe {
 	unsigned long sp;
 	unsigned long lr;
 	unsigned long pc;
-#ifdef CONFIG_KRETPROBES
+#if defined(CONFIG_KRETPROBES) || defined(CONFIG_RETHOOK)
 	struct llist_node *kr_cur;
 	struct task_struct *tsk;
 #endif
@@ -27,7 +27,7 @@ void arm_get_current_stackframe(struct pt_regs *regs, struct stackframe *frame)
 		frame->sp = regs->ARM_sp;
 		frame->lr = regs->ARM_lr;
 		frame->pc = regs->ARM_pc;
-#ifdef CONFIG_KRETPROBES
+#if defined(CONFIG_KRETPROBES) || defined(CONFIG_RETHOOK)
 		frame->kr_cur = NULL;
 		frame->tsk = current;
 #endif

arch/arm/kernel/stacktrace.c

@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0-only
 #include <linux/export.h>
 #include <linux/kprobes.h>
+#include <linux/rethook.h>
 #include <linux/sched.h>
 #include <linux/sched/debug.h>
 #include <linux/stacktrace.h>
@@ -66,6 +67,11 @@ int notrace unwind_frame(struct stackframe *frame)
 	frame->sp = *(unsigned long *)(fp - 8);
 	frame->pc = *(unsigned long *)(fp - 4);
 #endif
+#ifdef CONFIG_RETHOOK
+	if (is_rethook_trampoline(frame->pc))
+		frame->pc = rethook_find_ret_addr(frame->tsk, frame->fp,
+						  &frame->kr_cur);
+#endif
 #ifdef CONFIG_KRETPROBES
 	if (is_kretprobe_trampoline(frame->pc))
 		frame->pc = kretprobe_find_ret_addr(frame->tsk,

arch/arm/probes/Makefile

@@ -6,3 +6,4 @@ obj-$(CONFIG_KPROBES)		+= decode-thumb.o
 else
 obj-$(CONFIG_KPROBES)		+= decode-arm.o
 endif
+obj-$(CONFIG_RETHOOK)		+= rethook.o

arch/arm/probes/rethook.c

@@ -0,0 +1,103 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * arm implementation of rethook. Mostly copied from arch/arm/probes/kprobes/core.c
+ */
+
+#include <linux/kprobes.h>
+#include <linux/rethook.h>
+
+/* Called from arch_rethook_trampoline */
+static __used unsigned long arch_rethook_trampoline_callback(struct pt_regs *regs)
+{
+	return rethook_trampoline_handler(regs, regs->ARM_fp);
+}
+NOKPROBE_SYMBOL(arch_rethook_trampoline_callback);
+
+/*
+ * When a rethook'ed function returns, it returns to arch_rethook_trampoline
+ * which calls rethook callback. We construct a struct pt_regs to
+ * give a view of registers r0-r11, sp, lr, and pc to the user
+ * return-handler. This is not a complete pt_regs structure, but that
+ * should be enough for stacktrace from the return handler with or
+ * without pt_regs.
+ */
+asm(
+	".text\n"
+	".global arch_rethook_trampoline\n"
+	".type arch_rethook_trampoline, %function\n"
+	"arch_rethook_trampoline:\n"
+#ifdef CONFIG_FRAME_POINTER
+	"ldr	lr, =arch_rethook_trampoline	\n\t"
+	/* this makes a framepointer on pt_regs. */
+#ifdef CONFIG_CC_IS_CLANG
+	"stmdb	sp, {sp, lr, pc}	\n\t"
+	"sub	sp, sp, #12		\n\t"
+	/* In clang case, pt_regs->ip = lr. */
+	"stmdb	sp!, {r0 - r11, lr}	\n\t"
+	/* fp points regs->r11 (fp) */
+	"add	fp, sp,	#44		\n\t"
+#else /* !CONFIG_CC_IS_CLANG */
+	/* In gcc case, pt_regs->ip = fp. */
+	"stmdb	sp, {fp, sp, lr, pc}	\n\t"
+	"sub	sp, sp, #16		\n\t"
+	"stmdb	sp!, {r0 - r11}		\n\t"
+	/* fp points regs->r15 (pc) */
+	"add	fp, sp, #60		\n\t"
+#endif /* CONFIG_CC_IS_CLANG */
+#else /* !CONFIG_FRAME_POINTER */
+	"sub	sp, sp, #16		\n\t"
+	"stmdb	sp!, {r0 - r11}		\n\t"
+#endif /* CONFIG_FRAME_POINTER */
+	"mov	r0, sp			\n\t"
+	"bl	arch_rethook_trampoline_callback	\n\t"
+	"mov	lr, r0			\n\t"
+	"ldmia	sp!, {r0 - r11}		\n\t"
+	"add	sp, sp, #16		\n\t"
+#ifdef CONFIG_THUMB2_KERNEL
+	"bx	lr			\n\t"
+#else
+	"mov	pc, lr			\n\t"
+#endif
+	".size arch_rethook_trampoline, .-arch_rethook_trampoline\n"
+);
+NOKPROBE_SYMBOL(arch_rethook_trampoline);
+
+/*
+ * At the entry of function with mcount. The stack and registers are prepared
+ * for the mcount function as below.
+ *
+ * mov     ip, sp
+ * push    {fp, ip, lr, pc}
+ * sub     fp, ip, #4	; FP[0] = PC, FP[-4] = LR, and FP[-12] = call-site FP.
+ * push    {lr}
+ * bl      <__gnu_mcount_nc> ; call ftrace
+ *
+ * And when returning from the function, call-site FP, SP and PC are restored
+ * from stack as below;
+ *
+ * ldm     sp, {fp, sp, pc}
+ *
+ * Thus, if the arch_rethook_prepare() is called from real function entry,
+ * it must change the LR and save FP in pt_regs. But if it is called via
+ * mcount context (ftrace), it must change the LR on stack, which is next
+ * to the PC (= FP[-4]), and save the FP value at FP[-12].
+ */
+void arch_rethook_prepare(struct rethook_node *rh, struct pt_regs *regs, bool mcount)
+{
+	unsigned long *ret_addr, *frame;
+
+	if (mcount) {
+		ret_addr = (unsigned long *)(regs->ARM_fp - 4);
+		frame = (unsigned long *)(regs->ARM_fp - 12);
+	} else {
+		ret_addr = &regs->ARM_lr;
+		frame = &regs->ARM_fp;
+	}
+
+	rh->ret_addr = *ret_addr;
+	rh->frame = *frame;
+
+	/* Replace the return addr with trampoline addr. */
+	*ret_addr = (unsigned long)arch_rethook_trampoline;
+}
+NOKPROBE_SYMBOL(arch_rethook_prepare);