The Sudo Story in Substrate

In this post, I will explain how the Sudo module is used to access permissioned functions in Substrate.

If you have ever run a local Substrate node for testing or development, you have probably interacted with the Sudo module. More specifically, you might have noticed that the “Alice” account is special, and can do powerful things to your blockchain!

In this blog post, I will show you end to end how the Sudo module works, why “Alice” is able to use this module, and how it enables access to permissioned functions like the one which enables Substrate runtime upgrades.

What is Sudo?

sudo is a program for Unix-like computer operating systems that allows users to run programs with the security privileges of another user, by default the superuser. It originally stood for “superuser do”…

Wikipedia

In short, sudo is a term used to represent the execution of some highly privileged operation by some highly privileged user. If you are trying to relate this to smart contracts on Ethereum, this is very similar to the “contract owner”, an account who is allowed to call onlyOwner functions.

The Sudo Module

The Substrate runtime module library (SRML) provides a Sudo module which provides this same functionality, but at the runtime level of your blockchain. At the time of writing this post, the Sudo module is only 60 lines of code, so you can easily look through the source code yourself to understand exactly what it does. But I will break it down for you just in case you are unfamiliar with the structure of Runtime modules.

The Sudo Key

The Sudo module has a single storage item: the “Sudo key”.

decl_storage! {
	trait Store for Module<T: Trait> as Sudo {
		/// The `AccountId` of the sudo key.
		Key get(key) config(): T::AccountId;
	}
}

This holds the AccountId of the person who is the “superuser” of your blockchain. Notice that it has the config() parameter, which means that this value can be set using the “genesis configuration” of your blockchain. We will talk about that more below.

The Sudo Module Functions

The Sudo module has two dispatchable functions which allow users to interact with the module.

The first function available in the Sudo module is set_key(origin, new), which allows only the Sudo key to change who the Sudo key is. This is not that interesting, so we won’t go into details.

The second function is sudo(origin, proposal), which allows only the Sudo key to dispatch a privileged call. This authorization check is done in the first two lines of the function:

let sender = ensure_signed(origin)?;
ensure!(sender == Self::key(), "only the current sudo key can sudo");

Then the function dispatches actually happens:

let ok = proposal.dispatch(system::RawOrigin::Root.into()).is_ok();

There are a few different things to note about this innocuous line:

At this point, you might be asking “What does this whole RawOrigin::Root mean?”

Privileged Functions

The Sudo module wouldn’t do much unless there were also “sudo-able” functions, and that is precisely what we will talk about next.

Substrate has the concept of “Privileged Functions” which are functions which specifically require Root origin. The origin of a call describes where the call has come from, and every dispatchable function should check at the beginning of the function that the call origin matches what is expected. The origin could be Signed as it was in the Sudo module, which represents a basic signed transaction, but Substrate also provides a Root origin which describes a call that comes from within the runtime itself. There is no way to produce a Root origin other than through internal runtime logic, and as such, we can treat functions that require this origin as privileged functions.

This is what a privileged function look like:

decl_module! {
    pub struct Module<T: Trait> for enum Call where origin: T::Origin {
        pub fn privileged_function(origin) -> Result {
            ensure_root(origin)?;
            // do something...
            Ok(())
        }
    }
}

However, macro magic makes this a bit more confusing. Generally, there is a rule for dispatchable functions where the first parameter must always be origin. However, when using the decl_module! macro, if you omit the origin parameter, then it will be added automatically, and ensure_root(origin)? will also be added. So an equivalent way to write the privileged_function above would be:

pub fn privileged_function() -> Result {
    // do something...
    Ok(())
}

So these kinds of functions will only be callable by internal runtime logic like what is implemented in the Sudo module.

A Substrate Runtime Upgrade

So let’s take a look at a real privileged function which is available within most Substrate runtimes, the runtime upgrade. From the Consensus module:

/// Set the new code.
pub fn set_code(new: Vec<u8>) {
    storage::unhashed::put_raw(well_known_keys::CODE, &new);
}

This single line of logic is enough to power the entire “upgrade” feature of Substrate forkless runtime upgrades. This privileged function checks that the caller must have Root origin, thanks to the decl_module! macro, then it puts the Wasm bytecode into a well_known_key called CODE.

Thus, when you use something like the Polkadot UI to do a runtime upgrade, it will look like this:

Walking through the UI you will see that:

When this transaction is dispatched, the following logic is executed:

  1. The sudo function checks that “Alice” is the Sudo key.
  2. She is, so the rest of the function runs.
  3. A Root origin call is dispatched to the setCode function.
  4. The setCode function checks that the origin is Root.
  5. It is, so the rest of the function runs.
  6. The storage value is updated for the CODE well known key.

And that is the magic of the Sudo module! The last thing we should probably talk about is how “Alice” became the Sudo module to begin with, and for that we need to look at the genesis configuration of our blockchain.

Initializing the Sudo key

Have you wondered why your substrate test network gives “Alice” a bunch of initial “balance units” and makes her the Sudo key for your runtime? Well, this is all controlled in your blockchain genesis configuration which is defined in a file called chain_spec.rs.

We can see that this code ultimately creates a GenesisConfig object with the following initial setting:

sudo: Some(SudoConfig {
    key: root_key,
}),

If we follow the logic back, this root_key is defined as account_key("Alice") which generates an AccountId using the seed string “Alice”.

fn account_key(s: &str) -> AccountId {
    sr25519::Pair::from_string(&format!("//{}", s), None)
        .expect("static values are valid; qed")
        .public()
}

This works great for test networks since “Alice” can be treated as a well known account that is automatically configured in your UI. However, for a real network, this is probably not what you want to do. Instead, you should pass an AccountId directly to this genesis configuration, and keep the seed for this account VERY secret.

The End

That is the entire sudo story in Substrate. I hope you learned something new and that this shed light to some of the things happening behind the scenes of Substrate runtimes.

If you enjoyed this content and want to support me producing more, feel free to check out my donation page.