about summary refs log tree commit diff
path: root/web/tvl/blog/2024-02-tvix-update.md

It's already been way too long since the last update here. It doesn't mean nothing has moved forward since, in fact a lot of things happened, but while there's been a talk about Tvix at NixCon 2023, as well as a Nix Developer Dialogue Interview, we never found the time to write something down - time to change that :-)

Evaluation regression testing

Most of the Evaluator work has been driven by evaluating nixpkgs, and ensuring we "evaluate it the same", which ultimately means we produce the same builds and end up with the same store paths.

We don't build things yet, but at least for nixpkgs (which doesn't do any IFD), it's possible to just peek at the outPath (and drvPath) of a package, and compare the calculated store path(s) with what Nix produces to determine if they use the same build recipe (for the package as well as all of its dependencies) 1.

We added some "integration tests" to our CI ensuring we keep getting the same output and drv paths as Nix, and already have quite complicated expressions in there and passing, like firefox, which exercises some of the more hairy bits in nixpkgs (like cross-compilation infrastructure for WASM). Yay!

Even though we're not getting into very fine-grained optimization until we're sure Tvix evaluates nixpkgs in a compatible fashion, we still want to have an idea about evaluation performance, and how it changes over time. To do this, we extended our benchmarks and wired them into Windtunnel, which now regularly runs benchmarks and provides graphs of how the benchmark results change over time, from commit to commit.

In the future, we plan to extend this to also run this as a part of code review, before changes are applied to our main branch. This will make it easier to see changes in performance right in the web interface, without having to do a manual benchmark locally before and after the change.

builtins.derivation[Strict], ATerms and output path calculation

These two are obviously needed to compare any derivation. As an interesting side note, in Nixcpp, the former is defined by a piece of bundled-in-Nix .nix code, that massages some parameters around and then calls the "native" derivationStrict - so implementing them required having the necessary tooling in Tvix to have builtins defined in .nix source code).

builtins.derivation[Strict] returns an attribute set with the previously mentioned outPath and drvPath paths. Implementing that correctly required implementing output path calculation the same way as Nix does (bit-by-bit).

Very little of how the output path calculation works is documented anywhere in Nixlang. It uses a subset of ATerm internally, produces "fingerprints" containing hashes of these ATerms, which are then hashed again. The intermediate hashes are not printed out anywhere (except if you [patch nix] nixcpp-patch-hashes to do so).

We already did parts of this correctly while starting this work on go-nix some while ago, but found some more edge cases and ultimately came up with a nicer interface than there.

All the Derivation internal data model, ATerm serialization and output path calculation have been sliced out into a more general-purpose nix-compat crate, alongside with more documentation unit tests and a Derivation ATerm parser, so hopefully this will now be more accessible for everyone now.

Note our builtin does not yet persist the Derivation anywhere "on disk" (though we have a debug CL that does write it to a temporary directory, in case we want to track down differences).

tvix-[ca]store

Tvix now has a store implementation!

The Nix model

Inside Nix, store path contents are normally hashed and communicated in NAR format, which is very coarse and often wasteful - a single bit of change in one file in a large store path causes a new NAR file to be uploaded to the binary cache, which then needs to be downloaded.

Additionally, identifying everything by the SHA256 digest of its NAR representation makes Nix store paths very incompatible with other content-addressed systems, as it's a very Nix-specific format.

The more granular Tvix model

After experimenting with some concepts and ideas in Golang, mostly around how to improve binary cache performance 2, both on-disk as well as over the network, we settled on a more granular, content-addressed and general-purpose format.

Internally, it behaves very similar to how git handles tree objects, except blobs are identified by their raw BLAKE3 digests rather than some custom encoding, and similarly, tree/directory objects use the BLAKE3 digest of its canonical protobuf serialization as identifiers.

This provides some immediate benefits:

  • We only need to keep the same data once, even if it's used across different store paths.
  • Transfers can be more granular and only need to fetch the data that's needed. Due to everything being content-addressed, it can be fetched from anything supporting BLAKE3 digests, immediately making it compatible with other P2P systems (IPFS blake3 blobs, …), or general-purpose content-addressed caches (bazel-remote).

There's a lot more details about the data model, certain decisions etc. in the docs.

Compatibility

We however still want to stay compatible with Nix, as in calculating "NAR-addressed" store paths the same, support substituting from regular Nix binary caches, as well as storing all the other additional metadata about store paths.

We accomplished this by splitting the two different concerns into two separate tvix-store and tvix-castore crates, with the former one holding all Nix-specific metadata and functionality, and the latter being a general-purpose content-addressed blob and filesystem tree storage system, which is usable in a lot of contexts outside of Tvix too. For example, if you want to use tvix-castore to write your own git alternative, or provide granular and authenticated access into large scientific datasets, you could!

Backends

In addition to a gRPC API and client bindings, there's support for local filesystem-based backends, as well as for sled, an embedded K/V database.

We're also currently working on a backend supporting most common object storages, as well as on more granular seeking and content-defined chunking for blobs.

FUSE/virtiofs

A tvix-store can be mounted via FUSE, or exposed through virtiofs 3. While doing the obvious thing - allowing mounting and browsing the contents of the store, this will allow lazy substitution of builds on remote builders, be in containerized or virtualized workloads.

We have an example in the repository seeding gnu hello into a throwaway store, then booting a MicroVM and executing it.

nar-bridge, bridging binary caches

nar-bridge and the NixHTTPPathInfoService bridge tvix-[ca]store with existing Nix binary caches and Nix.

The former exposes a tvix-[ca]store over the common Nix HTTP Binary Cache interface (both read and write).

The latter allows Tvix to substitute from regular Nix HTTP Binary caches, unpacking NARs and ingesting them on-the-fly into the castore model. The necessary parsers for NARInfo, signatures etc are also available in the nix-compat crate.

EvalIO / builtins interacting with the store more closely

tvix-eval itself is designed to be quite pure when it comes to IO - it doesn't do any IO directly on its own, but for the very little IO functionality it does as part of "basic interaction with paths"4 (like importing other .nix files), it goes through an EvalIO interface, which is provided to the Evaluator struct on instantiation.

This allows us to be a bit more flexible with how IO looks like in practice, which becomes interesting for specific store implementations that might not expose a POSIX filesystem directly, or targets where we don't have a filesystem at all (like WASM).

Using the EvalIO trait also allows avoiding the tvix-eval crate to get too strongly coupled to a specific store implementation, hashing scheme etc 4. As we can extend the set of builtins available to the evaluator with "foreign builtins", these can live in other crates.

Following this pattern, we started implementing some of the "basic" builtins that deal with path access in tvix-eval, like:

  • builtins.pathExists
  • builtins.readFile

We also recently started working on more complicated builtins like builtins.filterSource and builtins.path, which are also used in nixpkgs.

Both import a path into the store, and allow passing a Nix expression that's used as a filter function for each path. builtins.path can also ensuring the imported contents match a certain hash.

This required the builtin to interact with the store and evaluator in a very tight fashion, as the filter function (written in Nix) needs to be repeatedly executed for each path, and its return value is able to cause the store to skip over certain paths (which it previously couldn't).

Getting the abstractions right there required some back-and-forth, but the remaining changes should land quite soon.

Catchables / tryEval

As you may know, Nix has a limited exception system for dealing with user-generated errors: builtins.tryEval can be used to detect if an expression fails (if builtins.throw or assert are used to generate it). This feature requires extra support in any Nix implementation, as errors may not necessarily cause the Nix program to abort.

The C++ Nix implementation just reuses the C++ language-provided Exception system for builtins.tryEval which Tvix can't (even if Rust had an equivalent system): In C++ Nix the runtime representation of the program in execution corresponds to the Nix expression tree of the relevant source files. This means that an exception raised in C++ code will automatically bubble up correctly since the C++ and Nix call stacks are equivalent to each other. Tvix compiles the Nix expressions to a byte code program which may be mutated by extra optimization rules (for example, we hope to eliminate as many thunks as possible in the future). This means that such a correspondence between Nix and the (in this case) VM runtime is not guaranteed.

Previously, builtins.tryEval (which is implemented in Rust and can access VM internals) just allowed the VM to recover from certain kinds of errors. This proved to be insufficient as it blew up as soon as a builtins.tryEval-ed thunk is forced again—extra bookkeeping was needed. As a solution, we now store thunk evaluation errors that can be recovered from as Value::Catchable which mitigates this problem.

As you can imagine, storing evaluation failures as "normal" values quickly leads to all sorts of bugs because most VM/builtins code is written with only ordinary values like attribute sets, strings etc. in mind. While ironing those out, we made sure to supplement those fixes with as many test cases for builtins.tryEval as possible. This will hopefully prevent any regressions if or rather when we touch this system again. We already have some ideas for replacing the Catchable value type with a cleaner representation.

String contexts

For a long time, we had the working theory of being able to get away with not implementing string contexts, but instead do reference scanning on a set of "known paths" (and not implement builtins.unsafeDiscardStringContext).

Unfortunately, we ultimately discovered we won't be able to do that, mostly due to a bug in Nix that's worked around in the nixpkgs' stdenv.mkDerivation implementation, but impossible to fix in Nix without breaking all hashes.

So we recently added support for string contexts into our NixString implementation, implemented the (unsafeDiscardStringContext, getContext) builtins, as well as some more unit tests that introspect string context behaviour of various builtins.

Strings as bstr

C++ nix uses C-style zero-terminated char pointers to represent strings internally - however, until recently, Tvix has used Rust String and str for string values. Since those are required to be valid utf-8, we haven't been able to properly represent all the string values that Nix supports.

We recently converted our internal representation to BString, which allows treating a Vec<u8> as a "morally-string-like" value.

JSON/TOML/XML

We added support for the toJSON, toXML, fromJSON and fromTOML builtins.

toXML is particularly exciting, as it's the only format that allows expressing (partially applied) functions. It's also used in some of Nix' own test suite, so we can now include these in our unit test suite (and pass, yay!).

Builder protocol, drv->builder

Some work went into the Builder protocol, and how Tvix represents builds internally.

Nix uses Derivations (in A-Term) as nodes in its build graph, but it refers to other store paths used in that build simply by these store paths only, and doesn't encode their expected content hashes anywhere.

In Nix, this poses a big problem as soon as these builds are scheduled on remote builders: Builds scheduled to a builder are not truly hermetic, but rely on that referred store path to actually have the same contents as the machine orchestrating the build (or at least very similar).

If a package is not binary reproducible, this can lead to so-called frankenbuilds. Even ignoring this problem, we still rely on state (the exact contents of that output in the remote builders' Nix Store), and making them appear there (if they can't be substituted) requires the one scheduling the build to copy them over, and be a trusted user.

We wanted to eliminate this hermiticity problem in the internal data model that Tvix uses to manage builds, by not relying on external state, and instead explicitly tracking both build inputs as well as produced outputs by their contents, too.


That's it for now!


  1. Why this means it's using the same build recipes is due to how the output path calculation for input-addressed paths works, more of that in the next section.

  2. See nix-casync for one example - investing content-defined chunking (while still keeping the NAR format)

  3. Strictly speaking, not limited to tvix-store - literally anything providing a listing into tvix-castore nodes.

  4. That's the same reason why builtins.derivation[Strict] also lives in tvix-glue, not in tvix-eval.