Blob tx and instructions (#592)

Work towards #589. The VM PR FuelLabs/fuel-vm#780 contains a working implementation. To allow contracts larger than the contract size limit, and to allow upgradability, we'll be adding a new tx type that's just a bunch of bytes, and opcodes to load subsections of that. A *blob* will be like a contract, but without state or balances, allowing it to be quite a bit cheaper. The new instructions `BSIZ` and `BLDD` and the new blob mode for `LDC` closely follow how `CSIZ`, `LDC` and `CCP` work, but only operation on blob data. ### Before requesting review - [x] I have reviewed the code myself
FuelLabs · Jul 31, 2024 · f182c55 · f182c55
1 parent 6ebbe21
commit f182c55
Show file tree

Hide file tree

Showing 6 changed files with 102 additions and 10 deletions.
diff --git a/src/SUMMARY.md b/src/SUMMARY.md
@@ -13,6 +13,7 @@
  - [Transaction Pointer](./tx-format/tx-pointer.md)
 - [Computing Identifiers](./identifiers/index.md)
  - [Asset ID](./identifiers/asset.md)
+ - [Blob ID](./identifiers/blob-id.md)
  - [Contract ID](./identifiers/contract-id.md)
  - [Predicate ID](./identifiers/predicate-id.md)
  - [Transaction ID](./identifiers/transaction-id.md)

diff --git a/src/fuel-vm/instruction-set.md b/src/fuel-vm/instruction-set.md
@@ -91,7 +91,7 @@
  - [`CCP`: Code copy](#ccp-code-copy)
  - [`CROO`: Code Merkle root](#croo-code-merkle-root)
  - [`CSIZ`: Code size](#csiz-code-size)
- - [`LDC`: Load code from an external contract](#ldc-load-code-from-an-external-contract)
+ - [`LDC`: Load code from an external contract](#ldc-load-code-from-an-external-contract-or-blob)
  - [`LOG`: Log event](#log-log-event)
  - [`LOGD`: Log data event](#logd-log-data-event)
  - [`MINT`: Mint new coins](#mint-mint-new-coins)
@@ -106,6 +106,9 @@
  - [`TIME`: Timestamp at height](#time-timestamp-at-height)
  - [`TR`: Transfer coins to contract](#tr-transfer-coins-to-contract)
  - [`TRO`: Transfer coins to output](#tro-transfer-coins-to-output)
+- [Blob Instructions](#blob-instructions)
+ - [`BSIZ`: Blob size](#bsiz-blob-size)
+ - [`BLDD`: Load data from a blob](#bldd-load-data-from-a-blob)
 - [Cryptographic Instructions](#cryptographic-instructions)
  - [`ECK1`: Secp251k1 signature recovery](#eck1-secp256k1-signature-recovery)
  - [`ECR1`: Secp256r1 signature recovery](#ecr1-secp256r1-signature-recovery)
@@ -1749,7 +1752,7 @@ Panic if:
 | Notes | If `$rD` is greater than the code size, zero bytes are filled in. |
 
 This is used only for reading and inspecting code of other contracts.
-Use [`LDC`](#ldc-load-code-from-an-external-contract) to load code for executing.
+Use [`LDC`](#ldc-load-code-from-an-external-contract-or-blob) to load code for executing.
 
 Panic if:
 
@@ -1793,27 +1796,37 @@ Panic if:
 - `$rB + 32` overflows or `> VM_MAX_RAM`
 - Contract with ID `MEM[$rB, 32]` is not in `tx.inputs`
 
-### `LDC`: Load code from an external contract
+### `LDC`: Load code from an external contract or blob
 
 | | |
 |-------------|---------------------------------------------------------------------------------------------------------------------------------------------------|
-| Description | Copy `$rC` bytes of code starting at `$rB` for contract with ID equal to the 32 bytes in memory starting at `$rA` into memory starting at `$ssp`. |
-| Operation | ```MEM[$ssp, $rC] = code($rA, $rB, $rC);```  |
-| Syntax | `ldc $rA, $rB, $rC`  |
-| Encoding | `0x00 rA rB rC -`  |
+| Description | Copy `$rC` bytes of code at offset `$rB` from object with 32 byte id starting at `$rA` into memory starting at `$ssp`. Object type is in `imm`.  |
+| Operation | `id = mem[$rA,32]; code = match imm { 0 => contract_code($id), 1 => blob_payload($id) }; MEM[$ssp, $rC] = code[$rB, $rC];` |
+| Syntax | `ldc $rA, $rB, $rC, imm` |
+| Encoding | `0x00 rA rB rC imm` |
 | Notes | If `$rC` is greater than the code size, zero bytes are filled in. |
 
+Object type from `imm` determined the source for loading as follows:
+
+| `imm` | Object type |
+|-------|---------------|
+| `0` | Contract code |
+| `1` | Blob payload |
+| other | _reserved_ |
+
 Panic if:
 
 - `$ssp + $rC` overflows or `> VM_MAX_RAM`
 - `$rA + 32` overflows or `> VM_MAX_RAM`
 - `$ssp + $rC >= $hp`
-- `$rC > CONTRACT_MAX_SIZE`
-- Contract with ID `MEM[$rA, 32]` is not in `tx.inputs`
+- `imm == 0` and `$rC > CONTRACT_MAX_SIZE`
+- `imm == 0` and contract with ID `MEM[$rA, 32]` is not in `tx.inputs`
+- `imm == 1` and contract with ID `MEM[$rA, 32]` is not found in the chain state
+- `imm >= 2` (reserved value)
 
 Increment `$fp->codesize`, `$ssp` by `$rC` padded to word alignment. Then set `$sp` to `$ssp`.
 
-This instruction can be used to concatenate the code of multiple contracts together. It can only be used when the stack area of the call frame is zero-sized.
+This instruction can be used to concatenate the code of multiple contracts or blobs together. It can only be used when the stack area of the call frame is zero-sized.
 
 ### `LOG`: Log event
 
@@ -2217,6 +2230,41 @@ In an external context, decrease `MEM[balanceOfStart(MEM[$rD, 32]), 8]` by `$rC`
 
 This modifies the `balanceRoot` field of the appropriate output(s).
 
+## Blob Instructions
+
+All these instructions advance the program counter `$pc` by `4` after performing their operation.
+
+### `BSIZ`: Blob size
+
+| | |
+|-------------|-----------------------------------------------------------------------------------------------------------|
+| Description | Set `$rA` to the size of the blob with ID equal to the 32 bytes in memory starting at `$rB`. |
+| Operation | `$rA = len(blob(MEM[$rB, 32]));` |
+| Syntax | `bsiz $rA, $rB` |
+| Encoding | `0x00 rA rB - -` |
+| Notes | |
+
+Panic if:
+
+- `$rA` is a [reserved register](./index.md#semantics)
+- `$rB + 32` overflows or `> VM_MAX_RAM`
+- Blob ID `MEM[$rB, 32]` is not found
+
+### `BLDD`: Load data from a blob
+
+|-------------|-------------------------------------------------------------------------------------------------------------|
+| Description | Load 32-byte blob id at `$rB`, and copy `$rD` bytes starting from `$rC` into `$sA`. |
+| Operation | `MEM[$rA, $rD] = blob(MEM[$rB, 32])[$rC, $rD];` |
+| Syntax | `bldd $rA, $rB, rC, $rD` |
+| Encoding | `0x00 rA rB rC rD` |
+| Notes | If `$rC >` blob size, zero bytes are filled in. |
+
+Panic if:
+
+- `$rA + $rD` overflows or `> VM_MAX_RAM` or `> $hp`
+- `$rB + 32` overflows or `> VM_MAX_RAM`
+- Blob ID `MEM[$rB, 32]` is not found
+
 ## Cryptographic Instructions
 
 All these instructions advance the program counter `$pc` by `4` after performing their operation.

diff --git a/src/identifiers/blob-id.md b/src/identifiers/blob-id.md
@@ -0,0 +1,10 @@
+# Blob ID
+
+The _blob ID_ (also called _blob hash_) of a transaction is computed as
+the [hash](../protocol/cryptographic-primitives.md#hashing) of the blob data.
+
+Blob ID calculation doesn't vary between chains.
+
+```python
+sha256(blob_data)
+```
diff --git a/src/identifiers/index.md b/src/identifiers/index.md
@@ -3,6 +3,7 @@
 This chapter defines how to compute unique identifiers.
 
 - [Asset ID](./asset.md)
+- [Blob ID](./blob-id.md)
 - [Contract ID](./contract-id.md)
 - [Predicate ID](./predicate-id.md)
 - [Transaction ID](./transaction-id.md)

diff --git a/src/tx-format/index.md b/src/tx-format/index.md
@@ -7,6 +7,9 @@ The Fuel Transaction Format.
  - [`TransactionScript`](./transaction.md#transactionscript)
  - [`TransactionCreate`](./transaction.md#transactioncreate)
  - [`TransactionMint`](./transaction.md#transactionmint)
+ - [`TransactionUpgrade`](./transaction.md#transactionupgrade)
+ - [`TransactionUpload`](./transaction.md#transactionupload)
+ - [`TransactionBlob`](./transaction.md#transactionblob)
 - [Input](./input.md)
  - [`InputCoin`](./input.md#inputcoin)
  - [`InputContract`](./input.md#inputcontract)

diff --git a/src/tx-format/transaction.md b/src/tx-format/transaction.md
@@ -7,6 +7,7 @@ enum TransactionType : uint8 {
  Mint = 2,
  Upgrade = 3,
  Upload = 4,
+ Blob = 5,
 }
 ```
 
@@ -232,3 +233,31 @@ Transaction is invalid if:
 - `subsectionIndex` >= `subsectionsNumber`
 - `subsectionsNumber > MAX_BYTECODE_SUBSECTIONS`
 - The [Binary Merkle tree](../protocol/cryptographic-primitives.md#binary-merkle-tree) root calculated from `(witnesses[witnessIndex], subsectionIndex, subsectionsNumber, proofSet)` is not equal to the `root`. Root calculation is affected by all fields, so modification of one of them invalidates the proof.
+
+## `TransactionBlob`
+
+The `Blob` inserts a simple binary blob in the chain. It's raw immutable data that can be cheaply loaded by the VM and used as instructions or just data. Unlike `Create`, it doesn't hold any state or balances.
+
+`Blob`s are content-addressed, i.e. the they are uniquely identified by hash of the data field. Programs running on the VM can load an already-posted blob just by the hash, without having to specify it in contract inputs.
+
+| name | type | description |
+|---------------------|-----------------------------|----------------------------------|
+| `id` | `byte[32]` | Blob id, i.e. hash of the data. |
+| `witnessIndex` | `uint16` | The witness index of the data. |
+| `policyTypes` | `uint32` | Bitfield of used policy types. |
+| `inputsCount` | `uint16` | Number of inputs. |
+| `outputsCount` | `uint16` | Number of outputs. |
+| `witnessesCount` | `uint16` | Number of witnesses. |
+| `policies` | [Policy](./policy.md)`[]` | List of policies. |
+| `inputs` | [Input](./input.md)`[]` | List of inputs. |
+| `outputs` | [Output](./output.md)`[]` | List of outputs. |
+| `witnesses` | [Witness](./witness.md)`[]` | List of witnesses. |
+
+Transaction is invalid if:
+
+- Any input is of type `InputType.Contract` or `InputType.Message` where `input.dataLength > 0`
+- Any input uses non-base asset.
+- Any output is of type `OutputType.Contract` or `OutputType.Variable` or `OutputType.Message` or `OutputType.ContractCreated`
+- Any output is of type `OutputType.Change` with non-base `asset_id`
+- `witnessIndex >= tx.witnessesCount`
+- `sha256(witnesses[witnessIndex]) != id`