Generate the same keys with Rfc2898DeriveBytes in c# and pbkdf2 in go

Question

Why do Rfc2898DeriveBytes in C# and pbkdf2 in go lang generate different keys?

my C# code

using System; using System.Security.Cryptography; using System.Text; public class Test { private static byte[] passBytes = new byte[] {164,176,124,62,244,154,226,211,177,90,202,180,12,142,25,225}; private static byte[] saltBytes = new byte[] {173,205,190,172,239,190,242,63,219,205,173,196,218,171,142,214}; public static byte[] GetKey() { var key = new Rfc2898DeriveBytes(Encoding.UTF8.GetString(passBytes, 0, 16), saltBytes).GetBytes(16); return key; } public static void Main() { System.Console.WriteLine(Convert.ToBase64String(GetKey())); } }

output: 77U85CphtSEwPP9a2T/jaQ==

golang code

package main import ( b64 "encoding/base64" "golang.org/x/crypto/pbkdf2" "crypto/sha1" ) var ( pass[]byte = []byte{164,176,124,62,244,154,226,211,177,90,202,180,12,142,25,225} salt[]byte = []byte{173,205,190,172,239,190,242,63,219,205,173,196,218,171,142,214} ) func getKey() (key[]byte){ key = pbkdf2.Key(pass,salt,1000,16,sha1.New) return } func main() { print(b64.StdEncoding.EncodeToString(getKey())) }

output: hnuuu+he4aF7vAzA8rfQtw==

Is there something different i must do?

Do they both use the same iteration count and SHA1? Also using UTF8 on passBytes seem suspect because the passBytes may not be a valid UTF-8 string. — zaph
– zaph, Commented Jun 30, 2017 at 14:03
The Rfc2898DeriveBytes ctors that takes a string just immediately turn it into bytes via the UTF-8 encoding. You can skip the creation of the string by just passing the bytes directly to a byte[]-based ctor. — bartonjs
– bartonjs, Commented Jun 30, 2017 at 15:01
@bartonjs whats wrong in the c# code? Is Encoding.UTF8.GetString(passBytes, 0, 16) not the right encoding supposed to be used on passBytes ? — avut
– avut, Commented Jul 8, 2017 at 0:27
@avut My point is that you have bytes which you turn into a string (via UTF8) that you pass to a constructor that turns the string into bytes (via UTF8). You'd save work by just using one of the constructors that took bytes directly. — bartonjs
– bartonjs, Commented Jul 8, 2017 at 1:50
@bartonjs Alright! I get it what you mean. But why is the output different shouldn't it be same? What is making output different. Isn't default encoding in C# is UTF-16? So Turning string into UTF-8 should give the same output? I'm confused. — avut
– avut, Commented Jul 8, 2017 at 13:43

putu · Accepted Answer · 2017-06-30 14:48:02Z

You're using different variant (the constructor that takes UTF-8 string) when initializing C# instance. In addition, as already pointed by zaph, you need to use same iteration count for both C# and golang codes. The golang version takes []byte arguments both for password and salt, and the C# counterpart is Rfc2898DeriveBytes Constructor (Byte[] password, Byte[] salt, Int32 iterations).

byte[] passBytes = new byte[] {164,176,124,62,244,154,226,211,177,90,202,180,12,142,25,225}; byte[] saltBytes = new byte[] {173,205,190,172,239,190,242,63,219,205,173,196,218,171,142,214}; var pbkdf2 = new Rfc2898DeriveBytes(passBytes, saltBytes, 1000); var key = Convert.ToBase64String(pbkdf2.GetBytes(16));

The output of above code is the same with golang version.

Thanks for reply. And in case I need to replicate the C# code in golang??
@avut It depends on which C# functionality that you want to replicate. If iterations is not specified in the C# constructor, I think the default value will be 1000. However, there are C# constructors that use random salt (Rfc2898DeriveBytes(String, Int32, Int32) and Rfc2898DeriveBytes(String, Int32, Int32)), and in this case you can't replicate.

Collectives™ on Stack Overflow

Generate the same keys with Rfc2898DeriveBytes in c# and pbkdf2 in go

1 Answer 1

2 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Related