[PATCH bpf-next v2 3/4] arm64: ftrace: Add ftrace direct call support

[Xu Kuohai]

From: Xu Kuohai <xukuohai at huawei.com>

Add ftrace direct support for arm64.

1. When there is custom trampoline only, patch fentry callsite to call
   the custom trampoline directly.

2. When ftrace caller and custom trampoline coexist, jump from fentry to
   ftrace caller first, then jump to custom trampoline when ftrace caller
   exits. As pt_regs->orig_x0 is currently unused by ftrace, its space
   is reused as an intermediary for jumping from ftrace caller to custom
   trampoline.

In short, this patch does the same thing as the x86 commit 562955fe6a55
("ftrace/x86: Add register_ftrace_direct() for custom trampolines").

Signed-off-by: Xu Kuohai <xukuohai at huawei.com>
Acked-by: Song Liu <songliubraving at fb.com>
---
 arch/arm64/Kconfig               |  2 ++
 arch/arm64/include/asm/ftrace.h  | 12 ++++++++++++
 arch/arm64/include/asm/ptrace.h  |  6 +++++-
 arch/arm64/kernel/asm-offsets.c  |  1 +
 arch/arm64/kernel/entry-ftrace.S | 18 +++++++++++++++---
 5 files changed, 35 insertions(+), 4 deletions(-)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 571cc234d0b3..e2f6ca75b881 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -180,6 +180,8 @@ config ARM64
 	select HAVE_DEBUG_KMEMLEAK
 	select HAVE_DMA_CONTIGUOUS
 	select HAVE_DYNAMIC_FTRACE
+	select HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS \
+		if DYNAMIC_FTRACE_WITH_REGS
 	select FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY \
 		if DYNAMIC_FTRACE_WITH_REGS
 	select HAVE_EFFICIENT_UNALIGNED_ACCESS
diff --git a/arch/arm64/include/asm/ftrace.h b/arch/arm64/include/asm/ftrace.h
index 40e63435965b..b07a3c24f918 100644
--- a/arch/arm64/include/asm/ftrace.h
+++ b/arch/arm64/include/asm/ftrace.h
@@ -67,6 +67,18 @@ extern void return_to_handler(void);
 
 unsigned long ftrace_call_adjust(unsigned long addr);
 
+#ifdef CONFIG_HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
+static inline void arch_ftrace_set_direct_caller(struct pt_regs *regs,
+						 unsigned long addr)
+{
+	/*
+	 * Place custom trampoline address in regs->custom_tramp to let ftrace
+	 * trampoline jump to it.
+	 */
+	regs->custom_tramp = addr;
+}
+#endif /* CONFIG_HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS */
+
 #ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS
 struct dyn_ftrace;
 struct ftrace_ops;
diff --git a/arch/arm64/include/asm/ptrace.h b/arch/arm64/include/asm/ptrace.h
index 41b332c054ab..9701c38fcc5f 100644
--- a/arch/arm64/include/asm/ptrace.h
+++ b/arch/arm64/include/asm/ptrace.h
@@ -185,7 +185,11 @@ struct pt_regs {
 			u64 pstate;
 		};
 	};
-	u64 orig_x0;
+	union {
+		u64 orig_x0;
+		/* Only used by ftrace to save custom trampoline address */
+		u64 custom_tramp;
+	};
 #ifdef __AARCH64EB__
 	u32 unused2;
 	s32 syscallno;
diff --git a/arch/arm64/kernel/asm-offsets.c b/arch/arm64/kernel/asm-offsets.c
index 1197e7679882..56d4acc52a86 100644
--- a/arch/arm64/kernel/asm-offsets.c
+++ b/arch/arm64/kernel/asm-offsets.c
@@ -80,6 +80,7 @@ int main(void)
   DEFINE(S_SDEI_TTBR1,		offsetof(struct pt_regs, sdei_ttbr1));
   DEFINE(S_PMR_SAVE,		offsetof(struct pt_regs, pmr_save));
   DEFINE(S_STACKFRAME,		offsetof(struct pt_regs, stackframe));
+  DEFINE(S_CUSTOM_TRAMP,	offsetof(struct pt_regs, custom_tramp));
   DEFINE(PT_REGS_SIZE,		sizeof(struct pt_regs));
   BLANK();
 #ifdef CONFIG_COMPAT
diff --git a/arch/arm64/kernel/entry-ftrace.S b/arch/arm64/kernel/entry-ftrace.S
index 0bebe3ffdb58..ae03df89d031 100644
--- a/arch/arm64/kernel/entry-ftrace.S
+++ b/arch/arm64/kernel/entry-ftrace.S
@@ -62,6 +62,9 @@
 	str	x29, [sp, #S_FP]
 	.endif
 
+	/* Set custom_tramp to zero  */
+	str     xzr, [sp, #S_CUSTOM_TRAMP]
+
 	/* Save the callsite's SP and LR */
 	add	x10, sp, #(PT_REGS_SIZE + 16)
 	stp	x9, x10, [sp, #S_LR]
@@ -114,12 +117,21 @@ SYM_INNER_LABEL(ftrace_call, SYM_L_GLOBAL)
 	/* Restore the callsite's FP, LR, PC */
 	ldr	x29, [sp, #S_FP]
 	ldr	x30, [sp, #S_LR]
-	ldr	x9, [sp, #S_PC]
-
+	ldr	x10, [sp, #S_PC]
+
+	ldr	x11, [sp, #S_CUSTOM_TRAMP]
+	cbz	x11, 1f
+	/* Set x9 to parent ip before jump to custom trampoline */
+	mov	x9,  x30
+	/* Set lr to self ip */
+	ldr	x30, [sp, #S_PC]
+	/* Set x10 (used for return address) to custom trampoline */
+	mov	x10, x11
+1:
 	/* Restore the callsite's SP */
 	add	sp, sp, #PT_REGS_SIZE + 16
 
-	ret	x9
+	ret	x10
 SYM_CODE_END(ftrace_common)
 
 #else /* CONFIG_DYNAMIC_FTRACE_WITH_REGS */
-- 
2.30.2