x86/alternative: Optimize single-byte NOPs at an arbitrary position

Up until now the assumption was that an alternative patching site would
have some instructions at the beginning and trailing single-byte NOPs
(0x90) padding. Therefore, the patching machinery would go and optimize
those single-byte NOPs into longer ones.

However, this assumption is broken on 32-bit when code like
hv_do_hypercall() in hyperv_init() would use the ratpoline speculation
killer CALL_NOSPEC. The 32-bit version of that macro would align certain
insns to 16 bytes, leading to the compiler issuing a one or more
single-byte NOPs, depending on the holes it needs to fill for alignment.

That would lead to the warning in optimize_nops() to fire:

  ------------[ cut here ]------------
  Not a NOP at 0xc27fb598
   WARNING: CPU: 0 PID: 0 at arch/x86/kernel/alternative.c:211 optimize_nops.isra.13

due to that function verifying whether all of the following bytes really
are single-byte NOPs.

Therefore, carve out the NOP padding into a separate function and call
it for each NOP range beginning with a single-byte NOP.

Fixes: 23c1ad5 ("x86/alternatives: Optimize optimize_nops()")
Reported-by: Richard Narron <richard@aaazen.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=213301
Link: https://lkml.kernel.org/r/20210601212125.17145-1-bp@alien8.de

arch/x86/kernel/alternative.c

@@ -182,42 +182,70 @@ recompute_jump(struct alt_instr *a, u8 *orig_insn, u8 *repl_insn, u8 *insn_buff)
 		n_dspl, (unsigned long)orig_insn + n_dspl + repl_len);
 }
 
+/*
+ * optimize_nops_range() - Optimize a sequence of single byte NOPs (0x90)
+ *
+ * @instr: instruction byte stream
+ * @instrlen: length of the above
+ * @off: offset within @instr where the first NOP has been detected
+ *
+ * Return: number of NOPs found (and replaced).
+ */
+static __always_inline int optimize_nops_range(u8 *instr, u8 instrlen, int off)
+{
+	unsigned long flags;
+	int i = off, nnops;
+
+	while (i < instrlen) {
+		if (instr[i] != 0x90)
+			break;
+
+		i++;
+	}
+
+	nnops = i - off;
+
+	if (nnops <= 1)
+		return nnops;
+
+	local_irq_save(flags);
+	add_nops(instr + off, nnops);
+	local_irq_restore(flags);
+
+	DUMP_BYTES(instr, instrlen, "%px: [%d:%d) optimized NOPs: ", instr, off, i);
+
+	return nnops;
+}
+
 /*
  * "noinline" to cause control flow change and thus invalidate I$ and
  * cause refetch after modification.
  */
 static void __init_or_module noinline optimize_nops(struct alt_instr *a, u8 *instr)
 {
-	unsigned long flags;
 	struct insn insn;
-	int nop, i = 0;
+	int i = 0;
 
 	/*
-	 * Jump over the non-NOP insns, the remaining bytes must be single-byte
-	 * NOPs, optimize them.
+	 * Jump over the non-NOP insns and optimize single-byte NOPs into bigger
+	 * ones.
 	 */
 	for (;;) {
 		if (insn_decode_kernel(&insn, &instr[i]))
 			return;
 
+		/*
+		 * See if this and any potentially following NOPs can be
+		 * optimized.
+		 */
 		if (insn.length == 1 && insn.opcode.bytes[0] == 0x90)
-			break;
-
-		if ((i += insn.length) >= a->instrlen)
-			return;
-	}
+			i += optimize_nops_range(instr, a->instrlen, i);
+		else
+			i += insn.length;
 
-	for (nop = i; i < a->instrlen; i++) {
-		if (WARN_ONCE(instr[i] != 0x90, "Not a NOP at 0x%px\n", &instr[i]))
+		if (i >= a->instrlen)
 			return;
 	}
-
-	local_irq_save(flags);
-	add_nops(instr + nop, i - nop);
-	local_irq_restore(flags);
-
-	DUMP_BYTES(instr, a->instrlen, "%px: [%d:%d) optimized NOPs: ",
-		   instr, nop, a->instrlen);
 }
 
 /*