perf: Enhance signing performance by caching scalar representation in PrivateKey

[jlucaso1]

Summary by CodeRabbit

• Performance

  • Further improved signing performance by caching the private scalar to avoid repeated recomputation.

• Refactor

  • Internal key state now stores ephemeral keys with indirection to reduce stack usage and change ownership semantics.

• Behavior

  • Some key constructors now require cached values for signing; attempting to sign with keys created without cache may fail.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds caching of the private-key scalar to Edwards cache structures and PrivateKey, adjusts constructors/accessors to populate and use the cached scalar for signing, and changes a PairCode state to heap-allocate ephemeral keypairs (Box).

Changes

Cohort / File(s)	Summary
EdwardsCacheData Scalar Addition wacore/libsignal/src/core/curve.rs	Adds scalar: Scalar to EdwardsCacheData, updates doc comments, and populates the scalar in get_edwards_cache and from_bytes_with_cache.
PrivateKey Cache Refactoring wacore/libsignal/src/core/curve/curve25519.rs	Adds scalar: Scalar to PrivateKey; renames compute_ed_public_key → compute_edwards_cache returning (Scalar, CompressedEdwardsY, u8); adds SIGN_BIT_NOT_INITIALIZED, new/adjusted constructors (new_without_cache, from_bytes_with_cache, from_bytes_without_cache), accessors (cached_scalar, cached_ed_public_key, cached_sign_bit), and makes calculate_signature use cached scalar and panic if cache uninitialized.
PairCode State Boxed KeyPair src/pair_code.rs, wacore/src/pair_code.rs	Changes PairCodeState::WaitingForPhoneConfirmation to store ephemeral_keypair as Box<EphemeralKeyPair> / Box<KeyPair> (heap-allocated) instead of by-value; public enum variant signature/type changed accordingly.

Sequence Diagram(s)

(Skipped — changes are internal caching and a local enum field type change; no new multi-component sequential flow to visualize.)

Estimated Code Review Effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• Feat pair code complete #202 — touches the same pair-code state/type and may overlap on PairCodeState/ephemeral keypair handling.
• perf: Cache Edwards public key for XEdDSA signatures (-27% signing, with lazy key gen) #240 — earlier changes to Edwards caching (lazy ed_public_key + sign_bit); this PR extends that caching to include the private scalar.

Poem

🐇 I nibble bytes and tuck them tight,
A scalar cached to speed the write,
No more repeats of costly grind,
A sentinel keeps errors confined,
Hooray — my hops make signing light!

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main performance optimization: caching the scalar representation in PrivateKey to avoid recomputation during signing operations.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

Bencher Report

Branch	feat-improvements-libsignal
Testbed	ubuntu-latest

🚨 1 Alert

Benchmark	Measure Units	View	Benchmark Result (Result Δ%)	Upper Boundary (Limit %)
binary_benchmark::unpack_group::bench_unpack_compressed	Instructions instructions x 1e3	📈 plot 🚷 threshold 🚨 alert (🔔)	556.53 x 1e3 (+13.70%) Baseline: 489.46 x 1e3	513.94 x 1e3 (108.29%)

Click to view all benchmark results

Benchmark	Instructions	Benchmark Result instructions (Result Δ%)	Upper Boundary instructions (Limit %)
binary_benchmark::attr_parser_group::bench_attr_parser attr_lookup:setup_attr_marshaled()	📈 view plot 🚷 view threshold	7,031.00 (-4.69%) Baseline: 7,377.06	7,745.91 (90.77%)
binary_benchmark::marshal_group::bench_marshal_allocating	📈 view plot 🚷 view threshold	124,647.00 (-12.68%) Baseline: 142,740.04	149,877.04 (83.17%)
binary_benchmark::marshal_group::bench_marshal_reusing_buffer	📈 view plot 🚷 view threshold	124,747.00 (-12.67%) Baseline: 142,840.94	149,982.99 (83.17%)
binary_benchmark::roundtrip_group::bench_roundtrip large:setup_large_marshaled()	📈 view plot 🚷 view threshold	106,274.00 (0.00%) Baseline: 106,274.00	111,587.70 (95.24%)
binary_benchmark::roundtrip_group::bench_roundtrip small:setup_small_marshaled()	📈 view plot 🚷 view threshold	8,487.00 (0.00%) Baseline: 8,487.00	8,911.35 (95.24%)
binary_benchmark::unmarshal_group::bench_unmarshal large:setup_large_marshaled()	📈 view plot 🚷 view threshold	49,338.00 (-0.39%) Baseline: 49,533.20	52,009.86 (94.86%)
binary_benchmark::unmarshal_group::bench_unmarshal small:setup_small_marshaled()	📈 view plot 🚷 view threshold	3,025.00 (-5.84%) Baseline: 3,212.54	3,373.16 (89.68%)
binary_benchmark::unpack_group::bench_unpack_compressed	📈 view plot 🚷 view threshold 🚨 view alert (🔔)	556,531.00 (+13.70%) Baseline: 489,462.46	513,935.59 (108.29%)
binary_benchmark::unpack_group::bench_unpack_uncompressed	📈 view plot 🚷 view threshold	779.00 (+0.02%) Baseline: 778.87	817.81 (95.25%)
libsignal_benchmark::conversation_group::bench_full_dm_conversation full:setup_conversation_data()	📈 view plot 🚷 view threshold	27,922,065.00 (+0.55%) Baseline: 27,768,815.43	29,157,256.21 (95.76%)
libsignal_benchmark::dm_group::bench_dm_decrypt_first_message decrypt_prekey:setup_dm_with_first_message()	📈 view plot 🚷 view threshold	5,571,063.00 (+0.02%) Baseline: 5,569,837.35	5,848,329.22 (95.26%)
libsignal_benchmark::dm_group::bench_dm_encrypt_first_message first_msg:setup_dm_session()	📈 view plot 🚷 view threshold	178,206.00 (+0.32%) Baseline: 177,631.26	186,512.82 (95.55%)
libsignal_benchmark::dm_group::bench_dm_encrypt_subsequent_message subsequent:setup_established_dm_session()	📈 view plot 🚷 view threshold	179,022.00 (+0.33%) Baseline: 178,439.00	187,360.95 (95.55%)
libsignal_benchmark::dm_group::bench_dm_session_establishment setup:setup_dm_users()	📈 view plot 🚷 view threshold	17,224,679.00 (-0.23%) Baseline: 17,264,501.43	18,127,726.51 (95.02%)
libsignal_benchmark::group_messaging_group::bench_group_create_distribution_message create:setup_group_sender()	📈 view plot 🚷 view threshold	295,925.00 (+0.10%) Baseline: 295,641.61	310,423.69 (95.33%)
libsignal_benchmark::group_messaging_group::bench_group_decrypt_message decrypt:setup_group_with_encrypted_message()	📈 view plot 🚷 view threshold	12,702,357.00 (+0.72%) Baseline: 12,611,739.70	13,242,326.68 (95.92%)
libsignal_benchmark::group_messaging_group::bench_group_encrypt_message encrypt:setup_group_with_distribution()	📈 view plot 🚷 view threshold	716,092.00 (+0.02%) Baseline: 715,964.17	751,762.38 (95.26%)
libsignal_benchmark::session_optimization_group::bench_decrypt_with_previous_session previous_session:setup_with_archived_sessions()	📈 view plot 🚷 view threshold	41,798.00 (+0.54%) Baseline: 41,572.00	43,650.60 (95.76%)
libsignal_benchmark::session_optimization_group::bench_out_of_order_decryption out_of_order:setup_out_of_order_messages()	📈 view plot 🚷 view threshold	5,530,955.00 (-0.25%) Baseline: 5,544,757.25	5,821,995.11 (95.00%)
libsignal_benchmark::session_optimization_group::bench_promote_matching_session promote:setup_promote_matching_session()	📈 view plot 🚷 view threshold	965,435.00 (-0.75%) Baseline: 972,685.75	1,021,320.04 (94.53%)
libsignal_benchmark::signature_group::bench_key_generation keygen	📈 view plot 🚷 view threshold	2,823,499.00 (+0.02%) Baseline: 2,822,809.80	2,963,950.29 (95.26%)
libsignal_benchmark::signature_group::bench_signature_creation sign:setup_keypair_with_message()	📈 view plot 🚷 view threshold	3,445,806.00 (-22.86%) Baseline: 4,466,834.00	4,690,175.70 (73.47%)
libsignal_benchmark::signature_group::bench_signature_verification verify:setup_keypair_with_message()	📈 view plot 🚷 view threshold	124,956,624.00 (-0.15%) Baseline: 125,140,173.60	131,397,182.28 (95.10%)
reporting_token_benchmark::content_extraction_group::bench_content_extraction extended:setup_extended_message()	📈 view plot 🚷 view threshold	11,819.00 (+0.07%) Baseline: 11,811.28	12,401.84 (95.30%)
reporting_token_benchmark::content_extraction_group::bench_content_extraction simple:setup_simple_message()	📈 view plot 🚷 view threshold	3,837.00 (+0.17%) Baseline: 3,830.31	4,021.83 (95.40%)
reporting_token_benchmark::full_generation_group::bench_full_token_generation extended:setup_full_gen_extended()	📈 view plot 🚷 view threshold	88,033.00 (+0.03%) Baseline: 88,006.72	92,407.06 (95.27%)
reporting_token_benchmark::full_generation_group::bench_full_token_generation simple:setup_full_gen_simple()	📈 view plot 🚷 view threshold	80,074.00 (+0.03%) Baseline: 80,048.76	84,051.20 (95.27%)
reporting_token_benchmark::key_derivation_group::bench_key_derivation	📈 view plot 🚷 view threshold	51,035.00 (0.00%) Baseline: 51,035.00	53,586.75 (95.24%)
reporting_token_benchmark::message_encoding_group::bench_message_encoding extended:setup_extended_message()	📈 view plot 🚷 view threshold	5,738.00 (+0.13%) Baseline: 5,730.28	6,016.79 (95.37%)
reporting_token_benchmark::message_encoding_group::bench_message_encoding simple:setup_simple_message()	📈 view plot 🚷 view threshold	2,121.00 (+0.32%) Baseline: 2,114.31	2,220.03 (95.54%)
reporting_token_benchmark::token_calculation_group::bench_token_calculation	📈 view plot 🚷 view threshold	21,921.00 (+0.07%) Baseline: 21,904.89	23,000.13 (95.31%)

🐰 View full continuous benchmarking report in Bencher