Solana supports writing onchain programs using theRust programming language.
Hello World: Get started with Solana development
To quickly get started with Solana development and build your first Rustprogram, take a look at these detailed quick start guides:
- Build and deploy your first Solana program using only your browser.No installation needed.
- Setup your local environment and use the localtest validator.
Project Layout #
Solana Rust programs follow the typicalRust project layout:
/inc//src//Cargo.toml
Solana Rust programs may depend directly on each other in order to gain accessto instruction helpers when makingcross-program invocations. When doing so it's important tonot pull in the dependent program's entrypoint symbols because they may conflictwith the program's own. To avoid this, programs should define an no-entrypoint
feature in Cargo.toml
and use to exclude the entrypoint.
Then when other programs include this program as a dependency, they should do sousing the no-entrypoint
feature.
Project Dependencies #
At a minimum, Solana Rust programs must pull in thesolana-program
crate.
Solana SBF programs have some restrictions that may prevent theinclusion of some crates as dependencies or require special handling.
For example:
- Crates that require the architecture be a subset of the ones supported by theofficial toolchain. There is no workaround for this unless that crate isforked and SBF added to that those architecture checks.
- Crates may depend on
rand
which is not supported in Solana's deterministicprogram environment. To include arand
dependent crate refer toDepending on Rand. - Crates may overflow the stack even if the stack overflowing code isn'tincluded in the program itself. For more information refer toStack.
How to Build #
First setup the environment:
- Install the latest Rust stable from https://rustup.rs/
- Install the latest Solana command-line tools
The normal cargo build is available for building programs against your hostmachine which can be used for unit testing:
cargo build
To build a specific program, such as SPL Token, for the Solana SBF target whichcan be deployed to the cluster:
cd <the program directory>cargo build-bpf
How to Test #
Solana programs can be unit tested via the traditional cargo test
mechanism byexercising program functions directly.
To help facilitate testing in an environment that more closely matches a livecluster, developers can use theprogram-test
crate. Theprogram-test
crate starts up a local instance of the runtime and allows teststo send multiple transactions while keeping state for the duration of the test.
For more information thetest in sysvar exampleshows how an instruction containing sysvar account is sent and processed by theprogram.
Program Entrypoint #
Programs export a known entrypoint symbol which the Solana runtime looks up andcalls when invoking a program. Solana supports multiple versions of the BPFloader and the entrypoints may vary between them. Programs must be written forand deployed to the same loader. For more details see theFAQ section on Loaders.
Currently there are two supported loadersBPF LoaderandBPF loader deprecated
They both have the same raw entrypoint definition, the following is the rawsymbol that the runtime looks up and calls:
#[no_mangle]pub unsafe extern "C" fn entrypoint(input: *mut u8) -> u64;
This entrypoint takes a generic byte array which contains the serialized programparameters (program id, accounts, instruction data, etc...). To deserialize theparameters each loader contains its own wrapper macro that exports the rawentrypoint, deserializes the parameters, calls a user defined instructionprocessing function, and returns the results.
You can find the entrypoint macros here:
The program defined instruction processing function that the entrypoint macroscall must be of this form:
pub type ProcessInstruction = fn(program_id: &Pubkey, accounts: &[AccountInfo], instruction_data: &[u8]) -> ProgramResult;
Parameter Deserialization #
Each loader provides a helper function that deserializes the program's inputparameters into Rust types. The entrypoint macros automatically calls thedeserialization helper:
Some programs may want to perform deserialization themselves and they can byproviding their own implementation of the raw entrypoint.Take note that the provided deserialization functions retain references back tothe serialized byte array for variables that the program is allowed to modify(lamports, account data). The reason for this is that upon return the loaderwill read those modifications so they may be committed. If a program implementstheir own deserialization function they need to ensure that any modificationsthe program wishes to commit be written back into the input byte array.
Details on how the loader serializes the program inputs can be found in theInput Parameter Serializationdocs.
Data Types #
The loader's entrypoint macros call the program defined instruction processorfunction with the following parameters:
program_id: &Pubkey,accounts: &[AccountInfo],instruction_data: &[u8]
The program id is the public key of the currently executing program.
The accounts is an ordered slice of the accounts referenced by the instructionand represented as anAccountInfostructures. An account's place in the array signifies its meaning, for example,when transferring lamports an instruction may define the first account as thesource and the second as the destination.
The members of the AccountInfo
structure are read-only except for lamports
and data
. Both may be modified by the program in accordance with the "runtimeenforcement policy". Both of these members are protected by the Rust RefCell
construct, so they must be borrowed to read or write to them. The reason forthis is they both point back to the original input byte array, but there may bemultiple entries in the accounts slice that point to the same account. UsingRefCell
ensures that the program does not accidentally perform overlappingread/writes to the same underlying data via multiple AccountInfo
structures.If a program implements their own deserialization function care should be takento handle duplicate accounts appropriately.
The instruction data is the general purpose byte array from theinstruction's instruction data beingprocessed.
Heap #
Rust programs implement the heap directly by defining a customglobal_allocator
Programs may implement their own global_allocator
based on its specific needs.Refer to the custom heap example for more information.
Restrictions #
On-chain Rust programs support most of Rust's libstd, libcore, and liballoc, aswell as many 3rd party crates.
There are some limitations since these programs run in a resource-constrained,single-threaded environment, as well as being deterministic:
- No access to
rand
std::fs
std::net
std::future
std::process
std::sync
std::task
std::thread
std::time
- Limited access to:
std::hash
std::os
- Bincode is extremely computationally expensive in both cycles and call depthand should be avoided
- String formatting should be avoided since it is also computationallyexpensive.
- No support for
println!
,print!
, the Solana logging helpersshould be used instead. - The runtime enforces a limit on the number of instructions a program canexecute during the processing of one instruction. Seecomputation budget for more information.
Depending on Rand #
Programs are constrained to run deterministically, so random numbers are notavailable. Sometimes a program may depend on a crate that depends itself onrand
even if the program does not use any of the random number functionality.If a program depends on rand
, the compilation will fail because there is noget-random
support for Solana. The error will typically look like this:
error: target is not supported, for more information see: https://docs.rs/getrandom/#unsupported-targets --> /Users/jack/.cargo/registry/src/github.com-1ecc6299db9ec823/getrandom-0.1.14/src/lib.rs:257:9 |257 | / compile_error!("\258 | | target is not supported, for more information see: \259 | | https://docs.rs/getrandom/#unsupported-targets\260 | | "); | |___________^
To work around this dependency issue, add the following dependency to theprogram's Cargo.toml
:
getrandom = { version = "0.1.14", features = ["dummy"] }
or if the dependency is on getrandom v0.2 add:
getrandom = { version = "0.2.2", features = ["custom"] }
Logging #
Rust's println!
macro is computationally expensive and not supported. Insteadthe helper macromsg!
is provided.
msg!
has two forms:
msg!("A string");
or
msg!(0_64, 1_64, 2_64, 3_64, 4_64);
Both forms output the results to the program logs. If a program so wishes theycan emulate println!
by using format!
:
msg!("Some variable: {:?}", variable);
The debugging section has moreinformation about working with program logs the Rust examplescontains a logging example.
Panicking #
Rust's panic!
, assert!
, and internal panic results are printed to theprogram logs by default.
INFO solana_runtime::message_processor] Finalized account CGLhHSuWsp1gT4B7MY2KACqp9RUwQRhcUFfVSuxpSajZINFO solana_runtime::message_processor] Call SBF program CGLhHSuWsp1gT4B7MY2KACqp9RUwQRhcUFfVSuxpSajZINFO solana_runtime::message_processor] Program log: Panicked at: 'assertion failed: `(left == right)` left: `1`, right: `2`', rust/panic/src/lib.rs:22:5INFO solana_runtime::message_processor] SBF program consumed 5453 of 200000 unitsINFO solana_runtime::message_processor] SBF program CGLhHSuWsp1gT4B7MY2KACqp9RUwQRhcUFfVSuxpSajZ failed: BPF program panicked
Custom Panic Handler #
Programs can override the default panic handler by providing their ownimplementation.
First define the custom-panic
feature in the program's Cargo.toml
[features]default = ["custom-panic"]custom-panic = []
Then provide a custom implementation of the panic handler:
#[cfg(all(feature = "custom-panic", target_os = "solana"))]#[no_mangle]fn custom_panic(info: &core::panic::PanicInfo<'_>) { solana_program::msg!("program custom panic enabled"); solana_program::msg!("{}", info);}
In the above snippit, the default implementation is shown, but developers mayreplace that with something that better suits their needs.
One of the side effects of supporting full panic messages by default is thatprograms incur the cost of pulling in more of Rust's libstd
implementationinto program's shared object. Typical programs will already be pulling in a fairamount of libstd
and may not notice much of an increase in the shared objectsize. But programs that explicitly attempt to be very small by avoiding libstd
may take a significant impact (~25kb). To eliminate that impact, programs canprovide their own custom panic handler with an empty implementation.
#[cfg(all(feature = "custom-panic", target_os = "solana"))]#[no_mangle]fn custom_panic(info: &core::panic::PanicInfo<'_>) { // Do nothing to save space}
Compute Budget #
Use the system call sol_remaining_compute_units()
to return a u64
indicatingthe number of compute units remaining for this transaction.
Use the system callsol_log_compute_units()
to log a message containing the remaining number of compute units the programmay consume before execution is halted
See the Compute Budget documentation formore information.
ELF Dump #
The SBF shared object internals can be dumped to a text file to gain moreinsight into a program's composition and what it may be doing at runtime. Thedump will contain both the ELF information as well as a list of all the symbolsand the instructions that implement them. Some of the BPF loader's error logmessages will reference specific instruction numbers where the error occurred.These references can be looked up in the ELF dump to identify the offendinginstruction and its context.
To create a dump file:
cd <program directory>cargo build-bpf --dump
Examples #
TheSolana Program Library GitHubrepo contains a collection of Rust examples.
TheSolana Developers Program Examples GitHubrepo also contains a collection of beginner to intermediate Rust programexamples.