Fixing source map locations when peephole optimizations are applied.

[vineethk]

Description

Fixes #14167.

Peephole optimization changes the original file format bytecode after it is generated from stackless bytecode. Source mapping is computed when the original file format bytecode is created, so the source mapping is incorrect after peephole optimization changes the file format bytecode.

This resulted in several downstream issues, including move test, prover, coverage, etc.

This PR contains the fix by having the peephole optimization maintain a mapping to the original offsets and propagating them, and updating the source map accordingly after the optimization.

How Has This Been Tested?

• Existing tests
• Added another test reported by a user
• Manually tested the user's report before and after the fix (details below).

When we ran aptos move test --move-2 for the following program:

module anto::Example {
    use std::option::{Option};

    const E_CONDITION_A: u64 = 1;
    const E_CONDITION_B: u64 = 2;
    const E_CONDITION_C: u64 = 3;
    const E_CONDITION_D: u64 = 4;

    fun validate_parameters(
        param_a: Option<u64>, // Optional first parameter to validate
        param_b: Option<u64>, // Optional second parameter to validate
        reference_value: u64, // Reference value for comparisons
        threshold_value: u64, // Threshold for additional validation
        flag: bool // Indicates a positive or negative condition
    ) {
        // Determine the context based on the flag
        let is_positive = flag;

        // Check and validate param_a if it exists
        if (param_a.is_some()) {
            // Extract the value from the Option
            let value_a = param_a.extract();

            // Ensure the value is non-zero
            assert!(value_a > 0, E_CONDITION_A);

            // Validate based on the condition (is_positive)
            let is_valid_a =
                if (is_positive) {
                    value_a > reference_value // For positive condition, value_a must be greater than reference_value
                } else {
                    value_a < reference_value // For negative condition, value_a must be less than reference_value
                };

            // Assert that the validation passed
            assert!(
                is_valid_a,
                if (is_positive) E_CONDITION_B else E_CONDITION_C
            );
        };

        // Check and validate param_b if it exists
        if (param_b.is_some()) {
            // Extract the value from the Option
            let value_b = param_b.extract();

            // Ensure the value is non-zero
            assert!(value_b > 0, E_CONDITION_A);

            // Validate based on the condition (is_positive)
            let is_valid_b =
                if (is_positive) {
                    value_b < reference_value // For positive condition, value_b must be less than reference_value
                } else {
                    value_b > reference_value // For negative condition, value_b must be greater than reference_value
                };

            // Assert that the validation passed
            assert!(
                is_valid_b,
                if (is_positive) E_CONDITION_C else E_CONDITION_D
            );

            // Additional validation against the threshold value if it exists
            if (threshold_value > 0) {
                let is_valid_threshold =
                    if (is_positive) {
                        value_b > threshold_value // For positive condition, value_b must be greater than threshold_value
                    } else {
                        value_b < threshold_value // For negative condition, value_b must be less than threshold_value
                    };

                // Assert that the threshold validation passed
                assert!(is_valid_threshold, E_CONDITION_A);
            }
        };
    }

    #[test_only]
    use std::option;

    #[test]
    fun test_validate_parameters() {
        // Passing Invalid param_a for positive condition
        // This should throw E_CONDITION_B error
        validate_parameters(
            option::some(40), // Less than reference_value (60)
            option::some(50),
            60,
            30,
            true
        );
    }
}

on the main branch, we get:

│  9 │     fun validate_parameters(
│    │         ------------------- In this function in 0xffeeede::Example
│    ·
│ 48 │             assert!(value_b > 0, E_CONDITION_A);
│    │             ^^^^^^ Test was not expected to error, but it aborted with code 2 originating in the module 000000000000000000000000000000000000000000000000000000000ffeeede::Example rooted here

with this PR, we get (correctly):

│  9 │     fun validate_parameters(
│    │         ------------------- In this function in 0xffeeede::Example
│    ·
│ 36 │             assert!(
│    │             ^^^^^^ Test was not expected to error, but it aborted with code 2 originating in the module 000000000000000000000000000000000000000000000000000000000ffeeede::Example rooted here

Key Areas to Review

• Correctness of the offset propagation
• It is expected that there are no changes to the generated file format bytecode itself, if you find anything that suggests otherwise, please comment.

Type of Change

• Bug fix

Which Components or Systems Does This Change Impact?

• Move Compiler

[trunk-io]

⏱️ 2h 55m total CI duration on this PR

Slowest 15 Jobs	Cumulative Duration	Recent Runs
execution-performance / single-node-performance	46m	🟩 🟩
rust-move-tests	13m	🟩
rust-move-tests	13m	🟩
rust-move-tests	12m	⬜
rust-move-tests	12m	🟩
rust-move-tests	12m	🟩
check-dynamic-deps	11m	🟩 🟩 🟩 🟩 🟩
test-target-determinator	10m	🟩 🟩
rust-cargo-deny	9m	🟩 🟩 🟩 🟩 🟩
execution-performance / test-target-determinator	9m	🟩 🟩
check	7m	🟩 🟩
rust-doc-tests	6m	🟩
rust-doc-tests	5m	🟩
fetch-last-released-docker-image-tag	3m	🟩 🟩
semgrep/ci	2m	🟩 🟩 🟩 🟩 🟩

🚨 1 job on the last run was significantly faster/slower than expected

Job	Duration	vs 7d avg	Delta
check-dynamic-deps	3m	1m

_{settings ⋅ feedback ⋅ docs ⋅ learn more about trunk.io}

[vineethk]

• Fixing source map locations when peephole optimizations are applied. #15384 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[rahxephon89]

LGTM, thanks

[brmataptos]

Looks ok, though I'd at least add a comment for code clarity.

[github-actions]

✅ Forge suite realistic_env_max_load success on 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7

two traffics test: inner traffic : committed: 14204.33 txn/s, latency: 2799.72 ms, (p50: 2700 ms, p70: 2700, p90: 3000 ms, p99: 6600 ms), latency samples: 5400820
two traffics test : committed: 99.89 txn/s, latency: 1815.46 ms, (p50: 1400 ms, p70: 1900, p90: 2200 ms, p99: 12400 ms), latency samples: 1660
Latency breakdown for phase 0: ["MempoolToBlockCreation: max: 2.061, avg: 1.365", "ConsensusProposalToOrdered: max: 0.326, avg: 0.293", "ConsensusOrderedToCommit: max: 0.379, avg: 0.367", "ConsensusProposalToCommit: max: 0.669, avg: 0.660"]
Max non-epoch-change gap was: 0 rounds at version 0 (avg 0.00) [limit 4], 0.81s no progress at version 2374168 (avg 0.20s) [limit 15].
Max epoch-change gap was: 0 rounds at version 0 (avg 0.00) [limit 4], 15.66s no progress at version 2374166 (avg 15.66s) [limit 16].
Test Ok

• Grafana dashboard
• Humio Logs
• Axiom Logs
• Validator CPU Profile
• Fullnode CPU Profile
• Test runner output
• Test run is land-blocking

[github-actions]

✅ Forge suite framework_upgrade success on f436adbe4384d6c5fd296addbb7f52d4be77231b ==> 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7

Compatibility test results for f436adbe4384d6c5fd296addbb7f52d4be77231b ==> 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7 (PR)
Upgrade the nodes to version: 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7
framework_upgrade::framework-upgrade::full-framework-upgrade : committed: 1383.85 txn/s, submitted: 1386.39 txn/s, failed submission: 2.54 txn/s, expired: 2.54 txn/s, latency: 2204.49 ms, (p50: 1800 ms, p70: 2300, p90: 3300 ms, p99: 5100 ms), latency samples: 119960
framework_upgrade::framework-upgrade::full-framework-upgrade : committed: 1434.38 txn/s, submitted: 1437.54 txn/s, failed submission: 3.16 txn/s, expired: 3.16 txn/s, latency: 2196.27 ms, (p50: 2100 ms, p70: 2100, p90: 3600 ms, p99: 4500 ms), latency samples: 127060
5. check swarm health
Compatibility test for f436adbe4384d6c5fd296addbb7f52d4be77231b ==> 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7 passed
Upgrade the remaining nodes to version: 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7
framework_upgrade::framework-upgrade::full-framework-upgrade : committed: 1514.82 txn/s, submitted: 1517.94 txn/s, failed submission: 3.12 txn/s, expired: 3.12 txn/s, latency: 2190.45 ms, (p50: 2100 ms, p70: 2300, p90: 3400 ms, p99: 4200 ms), latency samples: 126140
Test Ok

• Grafana dashboard
• Humio Logs
• Axiom Logs
• Validator CPU Profile
• Fullnode CPU Profile
• Test runner output
• Test run is land-blocking

[github-actions]

✅ Forge suite compat success on f436adbe4384d6c5fd296addbb7f52d4be77231b ==> 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7

Compatibility test results for f436adbe4384d6c5fd296addbb7f52d4be77231b ==> 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7 (PR)
1. Check liveness of validators at old version: f436adbe4384d6c5fd296addbb7f52d4be77231b
compatibility::simple-validator-upgrade::liveness-check : committed: 17228.60 txn/s, latency: 1970.77 ms, (p50: 2100 ms, p70: 2100, p90: 2200 ms, p99: 2500 ms), latency samples: 554520
2. Upgrading first Validator to new version: 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7
compatibility::simple-validator-upgrade::single-validator-upgrading : committed: 7665.05 txn/s, latency: 3759.01 ms, (p50: 4200 ms, p70: 4400, p90: 4500 ms, p99: 4600 ms), latency samples: 144100
compatibility::simple-validator-upgrade::single-validator-upgrade : committed: 7034.80 txn/s, latency: 4534.80 ms, (p50: 4500 ms, p70: 4600, p90: 6800 ms, p99: 7000 ms), latency samples: 233640
3. Upgrading rest of first batch to new version: 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7
compatibility::simple-validator-upgrade::half-validator-upgrading : committed: 7717.34 txn/s, latency: 3771.19 ms, (p50: 4200 ms, p70: 4400, p90: 4400 ms, p99: 4700 ms), latency samples: 145260
compatibility::simple-validator-upgrade::half-validator-upgrade : committed: 7074.72 txn/s, latency: 4525.80 ms, (p50: 4400 ms, p70: 4600, p90: 6800 ms, p99: 7100 ms), latency samples: 268180
4. upgrading second batch to new version: 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7
compatibility::simple-validator-upgrade::rest-validator-upgrading : committed: 11031.78 txn/s, latency: 2528.09 ms, (p50: 2400 ms, p70: 2800, p90: 4100 ms, p99: 4500 ms), latency samples: 201400
compatibility::simple-validator-upgrade::rest-validator-upgrade : committed: 10933.26 txn/s, latency: 2830.92 ms, (p50: 2500 ms, p70: 2900, p90: 5100 ms, p99: 6900 ms), latency samples: 357600
5. check swarm health
Compatibility test for f436adbe4384d6c5fd296addbb7f52d4be77231b ==> 2e87c37dbc23907eac35c6cdcba79f2b4652e4c7 passed
Test Ok

• Grafana dashboard
• Humio Logs
• Axiom Logs
• Validator CPU Profile
• Fullnode CPU Profile
• Test runner output
• Test run is land-blocking