x86/ibt: Clean up is_endbr()

Pretty much every caller of is_endbr() actually wants to test something at an
address and ends up doing get_kernel_nofault(). Fold the lot into a more
convenient helper.

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Sami Tolvanen <samitolvanen@google.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: "Masami Hiramatsu (Google)" <mhiramat@kernel.org>
Link: https://lore.kernel.org/r/20250207122546.181367417@infradead.org

Author: Peter Zijlstra

arch/x86/events/core.c

@@ -2844,7 +2844,7 @@ static bool is_uprobe_at_func_entry(struct pt_regs *regs)
 		return true;
 
 	/* endbr64 (64-bit only) */
-	if (user_64bit_mode(regs) && is_endbr(*(u32 *)auprobe->insn))
+	if (user_64bit_mode(regs) && is_endbr((u32 *)auprobe->insn))
 		return true;
 
 	return false;

arch/x86/include/asm/ftrace.h

@@ -36,21 +36,9 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr)
 
 static inline unsigned long arch_ftrace_get_symaddr(unsigned long fentry_ip)
 {
-#ifdef CONFIG_X86_KERNEL_IBT
-	u32 instr;
-
-	/* We want to be extra safe in case entry ip is on the page edge,
-	 * but otherwise we need to avoid get_kernel_nofault()'s overhead.
-	 */
-	if ((fentry_ip & ~PAGE_MASK) < ENDBR_INSN_SIZE) {
-		if (get_kernel_nofault(instr, (u32 *)(fentry_ip - ENDBR_INSN_SIZE)))
-			return fentry_ip;
-	} else {
-		instr = *(u32 *)(fentry_ip - ENDBR_INSN_SIZE);
-	}
-	if (is_endbr(instr))
+	if (is_endbr((void*)(fentry_ip - ENDBR_INSN_SIZE)))
 		fentry_ip -= ENDBR_INSN_SIZE;
-#endif
+
 	return fentry_ip;
 }
 #define ftrace_get_symaddr(fentry_ip)	arch_ftrace_get_symaddr(fentry_ip)

arch/x86/include/asm/ibt.h

@@ -65,7 +65,7 @@ static inline __attribute_const__ u32 gen_endbr_poison(void)
 	return 0x001f0f66; /* osp nopl (%rax) */
 }
 
-static inline bool is_endbr(u32 val)
+static inline bool __is_endbr(u32 val)
 {
 	if (val == gen_endbr_poison())
 		return true;
@@ -74,6 +74,7 @@ static inline bool is_endbr(u32 val)
 	return val == gen_endbr();
 }
 
+extern __noendbr bool is_endbr(u32 *val);
 extern __noendbr u64 ibt_save(bool disable);
 extern __noendbr void ibt_restore(u64 save);
 
@@ -98,7 +99,7 @@ extern __noendbr void ibt_restore(u64 save);
 
 #define __noendbr
 
-static inline bool is_endbr(u32 val) { return false; }
+static inline bool is_endbr(u32 *val) { return false; }
 
 static inline u64 ibt_save(bool disable) { return 0; }
 static inline void ibt_restore(u64 save) { }

arch/x86/kernel/alternative.c

@@ -852,16 +852,24 @@ void __init_or_module noinline apply_returns(s32 *start, s32 *end) { }
 
 #ifdef CONFIG_X86_KERNEL_IBT
 
+__noendbr bool is_endbr(u32 *val)
+{
+	u32 endbr;
+
+	__get_kernel_nofault(&endbr, val, u32, Efault);
+	return __is_endbr(endbr);
+
+Efault:
+	return false;
+}
+
 static void poison_cfi(void *addr);
 
 static void __init_or_module poison_endbr(void *addr, bool warn)
 {
-	u32 endbr, poison = gen_endbr_poison();
-
-	if (WARN_ON_ONCE(get_kernel_nofault(endbr, addr)))
-		return;
+	u32 poison = gen_endbr_poison();
 
-	if (!is_endbr(endbr)) {
+	if (!is_endbr(addr)) {
 		WARN_ON_ONCE(warn);
 		return;
 	}
@@ -988,7 +996,7 @@ static u32  cfi_seed __ro_after_init;
 static u32 cfi_rehash(u32 hash)
 {
 	hash ^= cfi_seed;
-	while (unlikely(is_endbr(hash) || is_endbr(-hash))) {
+	while (unlikely(__is_endbr(hash) || __is_endbr(-hash))) {
 		bool lsb = hash & 1;
 		hash >>= 1;
 		if (lsb)

arch/x86/kernel/kprobes/core.c

@@ -373,16 +373,7 @@ static bool can_probe(unsigned long paddr)
 kprobe_opcode_t *arch_adjust_kprobe_addr(unsigned long addr, unsigned long offset,
 					 bool *on_func_entry)
 {
-	u32 insn;
-
-	/*
-	 * Since 'addr' is not guaranteed to be safe to access, use
-	 * copy_from_kernel_nofault() to read the instruction:
-	 */
-	if (copy_from_kernel_nofault(&insn, (void *)addr, sizeof(u32)))
-		return NULL;
-
-	if (is_endbr(insn)) {
+	if (is_endbr((u32 *)addr)) {
 		*on_func_entry = !offset || offset == 4;
 		if (*on_func_entry)
 			offset = 4;

arch/x86/net/bpf_jit_comp.c

@@ -641,7 +641,7 @@ int bpf_arch_text_poke(void *ip, enum bpf_text_poke_type t,
 	 * See emit_prologue(), for IBT builds the trampoline hook is preceded
 	 * with an ENDBR instruction.
 	 */
-	if (is_endbr(*(u32 *)ip))
+	if (is_endbr(ip))
 		ip += ENDBR_INSN_SIZE;
 
 	return __bpf_arch_text_poke(ip, t, old_addr, new_addr);
@@ -3036,7 +3036,7 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *rw_im
 		/* skip patched call instruction and point orig_call to actual
 		 * body of the kernel function.
 		 */
-		if (is_endbr(*(u32 *)orig_call))
+		if (is_endbr(orig_call))
 			orig_call += ENDBR_INSN_SIZE;
 		orig_call += X86_PATCH_SIZE;
 	}

kernel/trace/bpf_trace.c

@@ -1038,27 +1038,14 @@ static const struct bpf_func_proto bpf_get_func_ip_proto_tracing = {
 	.arg1_type	= ARG_PTR_TO_CTX,
 };
 
-#ifdef CONFIG_X86_KERNEL_IBT
-static unsigned long get_entry_ip(unsigned long fentry_ip)
+static inline unsigned long get_entry_ip(unsigned long fentry_ip)
 {
-	u32 instr;
-
-	/* We want to be extra safe in case entry ip is on the page edge,
-	 * but otherwise we need to avoid get_kernel_nofault()'s overhead.
-	 */
-	if ((fentry_ip & ~PAGE_MASK) < ENDBR_INSN_SIZE) {
-		if (get_kernel_nofault(instr, (u32 *)(fentry_ip - ENDBR_INSN_SIZE)))
-			return fentry_ip;
-	} else {
-		instr = *(u32 *)(fentry_ip - ENDBR_INSN_SIZE);
-	}
-	if (is_endbr(instr))
+#ifdef CONFIG_X86_KERNEL_IBT
+	if (is_endbr((void *)(fentry_ip - ENDBR_INSN_SIZE)))
 		fentry_ip -= ENDBR_INSN_SIZE;
+#endif
 	return fentry_ip;
 }
-#else
-#define get_entry_ip(fentry_ip) fentry_ip
-#endif
 
 BPF_CALL_1(bpf_get_func_ip_kprobe, struct pt_regs *, regs)
 {