buffer: improve Buffer.from performance

Using == null in code paths that are expected to mostly receive
objects, arrays or other more complex data types is not
ideal because typecasting these types is very slow. Change
to instead check === null || === undefined. Also move one
variable assignment in fromString after an if condition
that doesn't need it (and returns if truthy).

PR-URL: https://github.com/nodejs/node/pull/15178
Refs: https://jsperf.com/triple-equals-vs-double-equals/3
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Timothy Gu <timothygu99@gmail.com>

lib/buffer.js

@@ -193,7 +193,7 @@ Buffer.from = function from(value, encodingOrOffset, length) {
   if (isAnyArrayBuffer(value))
     return fromArrayBuffer(value, encodingOrOffset, length);
 
-  if (value == null) {
+  if (value === null || value === undefined) {
     throw new errors.TypeError(
       'ERR_INVALID_ARG_TYPE',
       'first argument',
@@ -208,7 +208,7 @@ Buffer.from = function from(value, encodingOrOffset, length) {
     );
 
   const valueOf = value.valueOf && value.valueOf();
-  if (valueOf != null && valueOf !== value)
+  if (valueOf !== null && valueOf !== undefined && valueOf !== value)
     return Buffer.from(valueOf, encodingOrOffset, length);
 
   var b = fromObject(value);
@@ -322,9 +322,9 @@ function allocate(size) {
 function fromString(string, encoding) {
   var length;
   if (typeof encoding !== 'string' || encoding.length === 0) {
-    encoding = 'utf8';
     if (string.length === 0)
       return new FastBuffer();
+    encoding = 'utf8';
     length = byteLengthUtf8(string);
   } else {
     length = byteLength(string, encoding, true);