x86,static_call: Use alternative RET encoding

In addition to teaching static_call about the new way to spell 'RET',
there is an added complication in that static_call() is allowed to
rewrite text before it is known which particular spelling is required.

In order to deal with this; have a static_call specific fixup in the
apply_return() 'alternative' patching routine that will rewrite the
static_call trampoline to match the definite sequence.

This in turn creates the problem of uniquely identifying static call
trampolines. Currently trampolines are 8 bytes, the first 5 being the
jmp.d32/ret sequence and the final 3 a byte sequence that spells out
'SCT'.

This sequence is used in __static_call_validate() to ensure it is
patching a trampoline and not a random other jmp.d32. That is,
false-positives shouldn't be plenty, but aren't a big concern.

OTOH the new __static_call_fixup() must not have false-positives, and
'SCT' decodes to the somewhat weird but semi plausible sequence:

  push %rbx
  rex.XB push %r12

Additionally, there are SLS concerns with immediate jumps. Combined it
seems like a good moment to change the signature to a single 3 byte
trap instruction that is unique to this usage and will not ever get
generated by accident.

As such, change the signature to: '0x0f, 0xb9, 0xcc', which decodes
to:

  ud1 %esp, %ecx

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Borislav Petkov <bp@suse.de>

1 files changed, 38 insertions, 2 deletions

diff --git a/arch/x86/kernel/static_call.c b/arch/x86/kernel/static_call.c
index aa72cefdd5be..fe21fe778185 100644
--- a/arch/x86/kernel/static_call.c
+++ b/arch/x86/kernel/static_call.c
@@ -12,6 +12,13 @@ enum insn_type {
 };
 
 /*
+ * ud1 %esp, %ecx - a 3 byte #UD that is unique to trampolines, chosen such
+ * that there is no false-positive trampoline identification while also being a
+ * speculation stop.
+ */
+static const u8 tramp_ud[] = { 0x0f, 0xb9, 0xcc };
+
+/*
  * cs cs cs xorl %eax, %eax - a single 5 byte instruction that clears %[er]ax
  */
 static const u8 xor5rax[] = { 0x2e, 0x2e, 0x2e, 0x31, 0xc0 };
@@ -43,7 +50,10 @@ static void __ref __static_call_transform(void *insn, enum insn_type type, void
 		break;
 
 	case RET:
-		code = &retinsn;
+		if (cpu_feature_enabled(X86_FEATURE_RETHUNK))
+			code = text_gen_insn(JMP32_INSN_OPCODE, insn, &__x86_return_thunk);
+		else
+			code = &retinsn;
 		break;
 	}
 
@@ -60,7 +70,7 @@ static void __static_call_validate(void *insn, bool tail, bool tramp)
 {
 	u8 opcode = *(u8 *)insn;
 
-	if (tramp && memcmp(insn+5, "SCT", 3)) {
+	if (tramp && memcmp(insn+5, tramp_ud, 3)) {
 		pr_err("trampoline signature fail");
 		BUG();
 	}
@@ -115,3 +125,29 @@ void arch_static_call_transform(void *site, void *tramp, void *func, bool tail)
 	mutex_unlock(&text_mutex);
 }
 EXPORT_SYMBOL_GPL(arch_static_call_transform);
+
+#ifdef CONFIG_RETPOLINE
+/*
+ * This is called by apply_returns() to fix up static call trampolines,
+ * specifically ARCH_DEFINE_STATIC_CALL_NULL_TRAMP which is recorded as
+ * having a return trampoline.
+ *
+ * The problem is that static_call() is available before determining
+ * X86_FEATURE_RETHUNK and, by implication, running alternatives.
+ *
+ * This means that __static_call_transform() above can have overwritten the
+ * return trampoline and we now need to fix things up to be consistent.
+ */
+bool __static_call_fixup(void *tramp, u8 op, void *dest)
+{
+	if (memcmp(tramp+5, tramp_ud, 3)) {
+		/* Not a trampoline site, not our problem. */
+		return false;
+	}
+
+	if (op == RET_INSN_OPCODE || dest == &__x86_return_thunk)
+		__static_call_transform(tramp, RET, NULL);
+
+	return true;
+}
+#endif