Use heap allocation + valgrind in backend unit test by hanno-becker · Pull Request #1633 · pq-code-package/mlkem-native

hanno-becker · 2026-03-19T15:33:04Z

Resolves Use heap allocation + valgrind in backend unit test #1364

oqs-bot · 2026-03-19T21:24:33Z

CBMC Results (ML-KEM-512)

Full Results (187 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1291s	1205s	+7.1%
`mlk_indcpa_keypair_derand`	✅	155s	145s	+7%
`mlk_indcpa_enc`	✅	145s	137s	+6%
`mlk_keccak_squeezeblocks_x4`	✅	123s	117s	+5%
`mlk_rej_uniform_c`	✅	76s	67s	+13%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	68s	64s	+6%
`mlk_poly_rej_uniform`	✅	39s	37s	+5%
`poly_ntt_native`	✅	36s	30s	+20%
`polyvec_basemul_acc_montgomery_cached_native`	✅	24s	23s	+4%
`keccakf1600x4_permute_native_x4`	✅	20s	21s	-5%
`mlk_poly_reduce_native`	✅	17s	14s	+21%
`mlk_polyvec_add`	✅	17s	18s	-6%
`mlk_poly_decompress_d10_native`	✅	16s	12s	+33%
`mlk_ntt_layer`	✅	13s	12s	+8%
`mlk_poly_decompress_d4_native`	✅	13s	12s	+8%
`mlk_poly_frommsg`	✅	11s	8s	+38%
`mlk_indcpa_dec`	✅	10s	11s	-9%
`mlk_keccak_absorb_once_x4`	✅	9s	7s	+29%
`mlk_ntt_butterfly_block`	✅	9s	7s	+29%
`mlk_poly_rej_uniform_x4`	✅	9s	8s	+12%
`mlk_polymat_permute_bitrev_to_custom`	✅	9s	10s	-10%
`mlk_fqmul`	✅	8s	6s	+33%
`mlk_invntt_layer`	✅	8s	6s	+33%
`mlk_keccak_squeeze_once`	✅	8s	9s	-11%
`mlk_poly_frombytes_native`	✅	8s	8s	+0%
`keccakf1600_permute_native`	✅	7s	5s	+40%
`mlk_keccak_squeezeblocks`	✅	7s	5s	+40%
`poly_frombytes_native_x86_64`	✅	7s	4s	+75%
`mlk_poly_add`	✅	6s	7s	-14%
`mlk_poly_getnoise_eta1_4x_native`	✅	6s	4s	+50%
`nttunpack_native_x86_64`	✅	6s	4s	+50%
`kem_check_pk`	✅	5s	2s	+150%
`mlk_poly_tomont_c`	✅	5s	3s	+67%
`mlk_polyvec_decompress_du`	✅	5s	2s	+150%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	5s	2s	+150%
`mlk_shake128x4_squeezeblocks`	✅	5s	1s	+400%
`mlk_value_barrier_u32`	✅	5s	3s	+67%
`poly_invntt_tomont_native`	✅	5s	2s	+150%
`poly_tobytes_native_x86_64`	✅	5s	5s	+0%
`keccak_f1600_x4_native_avx2`	✅	4s	2s	+100%
`keccakf1600x4_xor_bytes_native`	✅	4s	5s	-20%
`kem_dec`	✅	4s	4s	+0%
`kem_enc`	✅	4s	3s	+33%
`kem_keypair_derand`	✅	4s	2s	+100%
`mlk_check_pct`	✅	4s	2s	+100%
`mlk_keccak_absorb_once`	✅	4s	7s	-43%
`mlk_keccakf1600_permute`	✅	4s	5s	-20%
`mlk_matvec_mul`	✅	4s	2s	+100%
`mlk_montgomery_reduce`	✅	4s	3s	+33%
`mlk_poly_cbd_eta2`	✅	4s	5s	-20%
`mlk_poly_compress_d10_c`	✅	4s	4s	+0%
`mlk_poly_compress_d4_c`	✅	4s	3s	+33%
`mlk_poly_compress_du`	✅	4s	2s	+100%
`mlk_poly_decompress_d10_c`	✅	4s	5s	-20%
`mlk_poly_decompress_d4_c`	✅	4s	2s	+100%
`mlk_poly_decompress_du`	✅	4s	1s	+300%
`mlk_polyvec_compress_du`	✅	4s	3s	+33%
`mlk_polyvec_frombytes`	✅	4s	2s	+100%
`mlk_polyvec_ntt`	✅	4s	2s	+100%
`mlk_polyvec_permute_bitrev_to_custom`	✅	4s	3s	+33%
`mlk_shake256x4`	✅	4s	2s	+100%
`mlk_value_barrier_i32`	✅	4s	2s	+100%
`poly_decompress_d10_native_x86_64`	✅	4s	3s	+33%
`poly_decompress_d4_native_x86_64`	✅	4s	3s	+33%
`poly_decompress_d5_native_x86_64`	✅	4s	2s	+100%
`poly_getnoise_eta1122_4x_native`	✅	4s	2s	+100%
`rej_uniform_native_x86_64`	✅	4s	3s	+33%
`intt_native_aarch64`	✅	3s	1s	+200%
`keccakf1600x4_extract_bytes_native`	✅	3s	2s	+50%
`kem_check_sk`	✅	3s	2s	+50%
`kem_enc_derand`	✅	3s	2s	+50%
`mlk_barrett_reduce`	✅	3s	2s	+50%
`mlk_ct_cmask_nonzero_u16`	✅	3s	3s	+0%
`mlk_gen_matrix`	✅	3s	3s	+0%
`mlk_gen_matrix_serial`	✅	3s	4s	-25%
`mlk_keccakf1600_extract_bytes`	✅	3s	2s	+50%
`mlk_keccakf1600x4_permute`	✅	3s	1s	+200%
`mlk_keypair_getnoise_eta1`	✅	3s	3s	+0%
`mlk_poly_compress_d10_native`	✅	3s	1s	+200%
`mlk_poly_decompress_d4`	✅	3s	2s	+50%
`mlk_poly_decompress_d5`	✅	3s	2s	+50%
`mlk_poly_getnoise_eta1122_4x`	✅	3s	2s	+50%
`mlk_poly_getnoise_eta1_4x`	✅	3s	3s	+0%
`mlk_poly_invntt_tomont_c`	✅	3s	3s	+0%
`mlk_poly_mulcache_compute_c`	✅	3s	5s	-40%
`mlk_poly_ntt`	✅	3s	2s	+50%
`mlk_poly_reduce_c`	✅	3s	2s	+50%
`mlk_poly_tobytes_c`	✅	3s	1s	+200%
`mlk_poly_tomont_native`	✅	3s	3s	+0%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	3s	4s	-25%
`mlk_polyvec_invntt_tomont`	✅	3s	2s	+50%
`mlk_polyvec_reduce`	✅	3s	3s	+0%
`mlk_scalar_compress_d11`	✅	3s	3s	+0%
`mlk_scalar_decompress_d11`	✅	3s	1s	+200%
`mlk_scalar_decompress_d4`	✅	3s	2s	+50%
`mlk_scalar_signed_to_unsigned_q`	✅	3s	2s	+50%
`ntt_native_x86_64`	✅	3s	4s	-25%
`poly_compress_d4_native_x86_64`	✅	3s	2s	+50%
`poly_mulcache_compute_native_aarch64`	✅	3s	2s	+50%
`poly_reduce_native_x86_64`	✅	3s	1s	+200%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	3s	1s	+200%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	3s	3s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	3s	1s	+200%
`rej_uniform_native`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	2s	3s	-33%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	2s	+0%
`kem_keypair`	✅	2s	2s	+0%
`mlk_ct_cmask_neg_i16`	✅	2s	2s	+0%
`mlk_ct_cmask_nonzero_u8`	✅	2s	3s	-33%
`mlk_ct_get_optblocker_u32`	✅	2s	2s	+0%
`mlk_ct_memcmp`	✅	2s	3s	-33%
`mlk_ct_sel_int16`	✅	2s	4s	-50%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	2s	4s	-50%
`mlk_poly_cbd_eta1`	✅	2s	2s	+0%
`mlk_poly_compress_d11_c`	✅	2s	2s	+0%
`mlk_poly_compress_d11_native`	✅	2s	1s	+100%
`mlk_poly_compress_d4`	✅	2s	2s	+0%
`mlk_poly_compress_d4_native`	✅	2s	4s	-50%
`mlk_poly_compress_d5`	✅	2s	2s	+0%
`mlk_poly_compress_dv`	✅	2s	2s	+0%
`mlk_poly_decompress_d11`	✅	2s	3s	-33%
`mlk_poly_decompress_d11_c`	✅	2s	3s	-33%
`mlk_poly_decompress_d5_c`	✅	2s	2s	+0%
`mlk_poly_decompress_d5_native`	✅	2s	2s	+0%
`mlk_poly_decompress_dv`	✅	2s	3s	-33%
`mlk_poly_frombytes_c`	✅	2s	3s	-33%
`mlk_poly_getnoise_eta2`	✅	2s	2s	+0%
`mlk_poly_invntt_tomont`	✅	2s	3s	-33%
`mlk_poly_mulcache_compute`	✅	2s	2s	+0%
`mlk_poly_ntt_c`	✅	2s	1s	+100%
`mlk_poly_reduce`	✅	2s	3s	-33%
`mlk_poly_tobytes`	✅	2s	2s	+0%
`mlk_poly_tobytes_native`	✅	2s	3s	-33%
`mlk_poly_tomont`	✅	2s	2s	+0%
`mlk_poly_tomsg`	✅	2s	2s	+0%
`mlk_polyvec_mulcache_compute`	✅	2s	2s	+0%
`mlk_polyvec_tomont`	✅	2s	2s	+0%
`mlk_scalar_compress_d1`	✅	2s	2s	+0%
`mlk_scalar_compress_d10`	✅	2s	2s	+0%
`mlk_scalar_compress_d4`	✅	2s	3s	-33%
`mlk_scalar_compress_d5`	✅	2s	1s	+100%
`mlk_scalar_decompress_d10`	✅	2s	2s	+0%
`mlk_sha3_256`	✅	2s	2s	+0%
`mlk_sha3_512`	✅	2s	4s	-50%
`mlk_shake128_absorb_once`	✅	2s	1s	+100%
`mlk_shake128_squeezeblocks`	✅	2s	3s	-33%
`mlk_value_barrier_u8`	✅	2s	1s	+100%
`ntt_native_aarch64`	✅	2s	5s	-60%
`poly_compress_d10_native_x86_64`	✅	2s	1s	+100%
`poly_compress_d11_native_x86_64`	✅	2s	2s	+0%
`poly_compress_d5_native_x86_64`	✅	2s	3s	-33%
`poly_decompress_d11_native_x86_64`	✅	2s	2s	+0%
`poly_mulcache_compute_native_x86_64`	✅	2s	4s	-50%
`poly_reduce_native_aarch64`	✅	2s	1s	+100%
`poly_tomont_native_x86_64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	2s	2s	+0%
`intt_native_x86_64`	✅	1s	3s	-67%
`keccak_f1600_x1_native_aarch64_v84a`	✅	1s	1s	+0%
`keccak_f1600_x4_native_aarch64_v84a`	✅	1s	3s	-67%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	3s	-67%
`mlk_ct_cmov_zero`	✅	1s	1s	+0%
`mlk_ct_get_optblocker_i32`	✅	1s	1s	+0%
`mlk_ct_get_optblocker_u8`	✅	1s	3s	-67%
`mlk_ct_sel_uint8`	✅	1s	3s	-67%
`mlk_keccakf1600_xor_bytes`	✅	1s	2s	-50%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	1s	2s	-50%
`mlk_keccakf1600x4_extract_bytes`	✅	1s	1s	+0%
`mlk_keccakf1600x4_xor_bytes`	✅	1s	3s	-67%
`mlk_poly_compress_d10`	✅	1s	2s	-50%
`mlk_poly_compress_d11`	✅	1s	3s	-67%
`mlk_poly_compress_d5_c`	✅	1s	1s	+0%
`mlk_poly_compress_d5_native`	✅	1s	3s	-67%
`mlk_poly_decompress_d10`	✅	1s	2s	-50%
`mlk_poly_decompress_d11_native`	✅	1s	4s	-75%
`mlk_poly_frombytes`	✅	1s	4s	-75%
`mlk_poly_mulcache_compute_native`	✅	1s	3s	-67%
`mlk_poly_sub`	✅	1s	3s	-67%
`mlk_polyvec_tobytes`	✅	1s	2s	-50%
`mlk_rej_uniform`	✅	1s	2s	-50%
`mlk_scalar_decompress_d5`	✅	1s	3s	-67%
`mlk_shake128x4_absorb_once`	✅	1s	1s	+0%
`mlk_shake256`	✅	1s	1s	+0%
`poly_tobytes_native_aarch64`	✅	1s	3s	-67%
`poly_tomont_native_aarch64`	✅	1s	2s	-50%
`rej_uniform_native_aarch64`	✅	1s	2s	-50%
`sys_check_capability`	✅	1s	2s	-50%

oqs-bot · 2026-03-19T21:24:59Z

CBMC Results (ML-KEM-1024)

Full Results (187 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1402s	1426s	-1.7%
`mlk_indcpa_enc`	✅	259s	254s	+2%
`polyvec_basemul_acc_montgomery_cached_native`	✅	131s	132s	-1%
`mlk_keccak_squeezeblocks_x4`	✅	123s	121s	+2%
`mlk_indcpa_keypair_derand`	✅	81s	82s	-1%
`mlk_rej_uniform_c`	✅	75s	77s	-3%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	66s	70s	-6%
`mlk_poly_rej_uniform`	✅	34s	35s	-3%
`poly_ntt_native`	✅	29s	35s	-17%
`mlk_polyvec_add`	✅	25s	24s	+4%
`keccakf1600x4_permute_native_x4`	✅	21s	20s	+5%
`mlk_poly_decompress_d11_native`	✅	15s	12s	+25%
`mlk_poly_reduce_native`	✅	14s	16s	-12%
`mlk_poly_decompress_d5_native`	✅	13s	15s	-13%
`mlk_polyvec_ntt`	✅	13s	14s	-7%
`mlk_ntt_layer`	✅	12s	12s	+0%
`mlk_indcpa_dec`	✅	10s	13s	-23%
`mlk_polymat_permute_bitrev_to_custom`	✅	10s	9s	+11%
`mlk_invntt_layer`	✅	8s	6s	+33%
`mlk_keccak_squeezeblocks`	✅	8s	6s	+33%
`mlk_poly_compress_d11_c`	✅	8s	9s	-11%
`mlk_poly_frombytes_native`	✅	8s	8s	+0%
`mlk_poly_frommsg`	✅	8s	10s	-20%
`mlk_poly_rej_uniform_x4`	✅	8s	7s	+14%
`mlk_fqmul`	✅	7s	5s	+40%
`mlk_gen_matrix`	✅	7s	8s	-12%
`mlk_keccak_absorb_once_x4`	✅	7s	9s	-22%
`mlk_ntt_butterfly_block`	✅	7s	8s	-12%
`poly_frombytes_native_x86_64`	✅	7s	5s	+40%
`mlk_keccak_squeeze_once`	✅	6s	6s	+0%
`keccakf1600_permute_native`	✅	5s	7s	-29%
`kem_dec`	✅	5s	6s	-17%
`mlk_gen_matrix_serial`	✅	5s	7s	-29%
`mlk_matvec_mul`	✅	5s	2s	+150%
`poly_getnoise_eta1122_4x_native`	✅	5s	1s	+400%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	5s	3s	+67%
`keccakf1600x4_xor_bytes_native`	✅	4s	2s	+100%
`kem_keypair_derand`	✅	4s	3s	+33%
`mlk_ct_cmask_nonzero_u16`	✅	4s	2s	+100%
`mlk_keccak_absorb_once`	✅	4s	4s	+0%
`mlk_keccakf1600_permute`	✅	4s	3s	+33%
`mlk_keccakf1600x4_xor_bytes`	✅	4s	2s	+100%
`mlk_poly_add`	✅	4s	5s	-20%
`mlk_poly_compress_d5_c`	✅	4s	4s	+0%
`mlk_poly_decompress_d4`	✅	4s	4s	+0%
`mlk_poly_decompress_d5_c`	✅	4s	2s	+100%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	4s	4s	+0%
`mlk_scalar_compress_d10`	✅	4s	2s	+100%
`mlk_scalar_decompress_d4`	✅	4s	2s	+100%
`poly_decompress_d10_native_x86_64`	✅	4s	1s	+300%
`poly_decompress_d11_native_x86_64`	✅	4s	4s	+0%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	4s	2s	+100%
`intt_native_aarch64`	✅	3s	2s	+50%
`keccak_f1600_x4_native_avx2`	✅	3s	2s	+50%
`kem_check_pk`	✅	3s	4s	-25%
`kem_enc`	✅	3s	2s	+50%
`kem_enc_derand`	✅	3s	2s	+50%
`mlk_ct_cmov_zero`	✅	3s	3s	+0%
`mlk_ct_get_optblocker_u32`	✅	3s	3s	+0%
`mlk_ct_get_optblocker_u8`	✅	3s	1s	+200%
`mlk_ct_sel_int16`	✅	3s	2s	+50%
`mlk_ct_sel_uint8`	✅	3s	1s	+200%
`mlk_poly_cbd_eta1`	✅	3s	2s	+50%
`mlk_poly_decompress_d10`	✅	3s	2s	+50%
`mlk_poly_decompress_d10_c`	✅	3s	4s	-25%
`mlk_poly_decompress_d11_c`	✅	3s	1s	+200%
`mlk_poly_decompress_d4_native`	✅	3s	2s	+50%
`mlk_poly_decompress_d5`	✅	3s	3s	+0%
`mlk_poly_decompress_du`	✅	3s	5s	-40%
`mlk_poly_frombytes`	✅	3s	3s	+0%
`mlk_poly_getnoise_eta1122_4x`	✅	3s	2s	+50%
`mlk_poly_mulcache_compute_c`	✅	3s	3s	+0%
`mlk_poly_reduce_c`	✅	3s	2s	+50%
`mlk_polyvec_compress_du`	✅	3s	4s	-25%
`mlk_polyvec_permute_bitrev_to_custom`	✅	3s	3s	+0%
`mlk_polyvec_tobytes`	✅	3s	4s	-25%
`mlk_polyvec_tomont`	✅	3s	2s	+50%
`mlk_rej_uniform`	✅	3s	3s	+0%
`mlk_scalar_compress_d1`	✅	3s	1s	+200%
`mlk_scalar_decompress_d10`	✅	3s	1s	+200%
`mlk_shake128x4_absorb_once`	✅	3s	2s	+50%
`mlk_shake256x4`	✅	3s	4s	-25%
`ntt_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d11_native_x86_64`	✅	3s	2s	+50%
`poly_compress_d5_native_x86_64`	✅	3s	1s	+200%
`poly_decompress_d4_native_x86_64`	✅	3s	1s	+200%
`poly_decompress_d5_native_x86_64`	✅	3s	6s	-50%
`poly_mulcache_compute_native_x86_64`	✅	3s	3s	+0%
`poly_reduce_native_x86_64`	✅	3s	3s	+0%
`poly_tomont_native_aarch64`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	3s	2s	+50%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	3s	3s	+0%
`rej_uniform_native_aarch64`	✅	3s	2s	+50%
`rej_uniform_native_x86_64`	✅	3s	3s	+0%
`keccak_f1600_x1_native_aarch64`	✅	2s	2s	+0%
`keccak_f1600_x1_native_aarch64_v84a`	✅	2s	1s	+100%
`keccak_f1600_x4_native_aarch64_v84a`	✅	2s	3s	-33%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	2s	2s	+0%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	2s	2s	+0%
`kem_check_sk`	✅	2s	4s	-50%
`kem_keypair`	✅	2s	4s	-50%
`mlk_barrett_reduce`	✅	2s	4s	-50%
`mlk_ct_cmask_neg_i16`	✅	2s	1s	+100%
`mlk_ct_cmask_nonzero_u8`	✅	2s	2s	+0%
`mlk_ct_memcmp`	✅	2s	3s	-33%
`mlk_keccakf1600_extract_bytes`	✅	2s	5s	-60%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	2s	4s	-50%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	2s	2s	+0%
`mlk_keccakf1600x4_extract_bytes`	✅	2s	1s	+100%
`mlk_keypair_getnoise_eta1`	✅	2s	3s	-33%
`mlk_montgomery_reduce`	✅	2s	2s	+0%
`mlk_poly_cbd_eta2`	✅	2s	3s	-33%
`mlk_poly_compress_d10_c`	✅	2s	1s	+100%
`mlk_poly_compress_d10_native`	✅	2s	4s	-50%
`mlk_poly_compress_d11`	✅	2s	2s	+0%
`mlk_poly_compress_d4`	✅	2s	4s	-50%
`mlk_poly_compress_d4_native`	✅	2s	2s	+0%
`mlk_poly_compress_d5_native`	✅	2s	3s	-33%
`mlk_poly_compress_dv`	✅	2s	2s	+0%
`mlk_poly_decompress_d10_native`	✅	2s	2s	+0%
`mlk_poly_decompress_d11`	✅	2s	2s	+0%
`mlk_poly_frombytes_c`	✅	2s	3s	-33%
`mlk_poly_getnoise_eta1_4x`	✅	2s	2s	+0%
`mlk_poly_getnoise_eta1_4x_native`	✅	2s	4s	-50%
`mlk_poly_getnoise_eta2`	✅	2s	2s	+0%
`mlk_poly_invntt_tomont`	✅	2s	3s	-33%
`mlk_poly_invntt_tomont_c`	✅	2s	3s	-33%
`mlk_poly_mulcache_compute`	✅	2s	4s	-50%
`mlk_poly_mulcache_compute_native`	✅	2s	2s	+0%
`mlk_poly_ntt`	✅	2s	2s	+0%
`mlk_poly_reduce`	✅	2s	3s	-33%
`mlk_poly_sub`	✅	2s	2s	+0%
`mlk_poly_tobytes_c`	✅	2s	2s	+0%
`mlk_poly_tobytes_native`	✅	2s	3s	-33%
`mlk_poly_tomont_c`	✅	2s	2s	+0%
`mlk_poly_tomsg`	✅	2s	1s	+100%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	2s	2s	+0%
`mlk_polyvec_decompress_du`	✅	2s	1s	+100%
`mlk_polyvec_invntt_tomont`	✅	2s	3s	-33%
`mlk_polyvec_mulcache_compute`	✅	2s	3s	-33%
`mlk_polyvec_reduce`	✅	2s	1s	+100%
`mlk_scalar_compress_d11`	✅	2s	2s	+0%
`mlk_scalar_compress_d5`	✅	2s	4s	-50%
`mlk_scalar_decompress_d11`	✅	2s	2s	+0%
`mlk_scalar_signed_to_unsigned_q`	✅	2s	3s	-33%
`mlk_sha3_256`	✅	2s	1s	+100%
`mlk_shake128_absorb_once`	✅	2s	2s	+0%
`mlk_shake128_squeezeblocks`	✅	2s	4s	-50%
`mlk_shake128x4_squeezeblocks`	✅	2s	2s	+0%
`mlk_value_barrier_i32`	✅	2s	3s	-33%
`mlk_value_barrier_u32`	✅	2s	2s	+0%
`nttunpack_native_x86_64`	✅	2s	3s	-33%
`poly_compress_d4_native_x86_64`	✅	2s	1s	+100%
`poly_invntt_tomont_native`	✅	2s	3s	-33%
`poly_mulcache_compute_native_aarch64`	✅	2s	1s	+100%
`poly_reduce_native_aarch64`	✅	2s	2s	+0%
`poly_tobytes_native_aarch64`	✅	2s	4s	-50%
`poly_tobytes_native_x86_64`	✅	2s	1s	+100%
`rej_uniform_native`	✅	2s	2s	+0%
`sys_check_capability`	✅	2s	2s	+0%
`intt_native_x86_64`	✅	1s	4s	-75%
`keccakf1600x4_extract_bytes_native`	✅	1s	2s	-50%
`mlk_check_pct`	✅	1s	2s	-50%
`mlk_ct_get_optblocker_i32`	✅	1s	3s	-67%
`mlk_keccakf1600_xor_bytes`	✅	1s	2s	-50%
`mlk_keccakf1600x4_permute`	✅	1s	1s	+0%
`mlk_poly_compress_d10`	✅	1s	2s	-50%
`mlk_poly_compress_d11_native`	✅	1s	3s	-67%
`mlk_poly_compress_d4_c`	✅	1s	2s	-50%
`mlk_poly_compress_d5`	✅	1s	2s	-50%
`mlk_poly_compress_du`	✅	1s	2s	-50%
`mlk_poly_decompress_d4_c`	✅	1s	2s	-50%
`mlk_poly_decompress_dv`	✅	1s	2s	-50%
`mlk_poly_ntt_c`	✅	1s	2s	-50%
`mlk_poly_tobytes`	✅	1s	3s	-67%
`mlk_poly_tomont`	✅	1s	3s	-67%
`mlk_poly_tomont_native`	✅	1s	3s	-67%
`mlk_polyvec_frombytes`	✅	1s	1s	+0%
`mlk_scalar_compress_d4`	✅	1s	1s	+0%
`mlk_scalar_decompress_d5`	✅	1s	3s	-67%
`mlk_sha3_512`	✅	1s	1s	+0%
`mlk_shake256`	✅	1s	2s	-50%
`mlk_value_barrier_u8`	✅	1s	2s	-50%
`ntt_native_aarch64`	✅	1s	2s	-50%
`poly_compress_d10_native_x86_64`	✅	1s	2s	-50%
`poly_tomont_native_x86_64`	✅	1s	1s	+0%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	1s	2s	-50%

oqs-bot · 2026-03-19T21:25:50Z

CBMC Results (ML-KEM-768)

Full Results (187 proofs)

Proof	Status	Current	Previous	Change
`TOTAL`	✅	1446s	1499s	-3.5%
`mlk_indcpa_enc`	✅	268s	277s	-3%
`mlk_indcpa_keypair_derand`	✅	212s	224s	-5%
`mlk_keccak_squeezeblocks_x4`	✅	123s	127s	-3%
`mlk_rej_uniform_c`	✅	69s	86s	-20%
`polyvec_basemul_acc_montgomery_cached_native`	✅	59s	60s	-2%
`mlk_polyvec_basemul_acc_montgomery_cached_c`	✅	43s	50s	-14%
`mlk_poly_rej_uniform`	✅	33s	37s	-11%
`mlk_polyvec_add`	✅	28s	26s	+8%
`poly_ntt_native`	✅	21s	26s	-19%
`keccakf1600x4_permute_native_x4`	✅	18s	23s	-22%
`mlk_indcpa_dec`	✅	15s	14s	+7%
`mlk_poly_decompress_d4_native`	✅	15s	16s	-6%
`mlk_ntt_layer`	✅	14s	13s	+8%
`mlk_poly_decompress_d10_native`	✅	14s	15s	-7%
`mlk_poly_reduce_native`	✅	14s	15s	-7%
`mlk_poly_frommsg`	✅	11s	11s	+0%
`mlk_keccak_absorb_once_x4`	✅	9s	8s	+12%
`keccakf1600_permute_native`	✅	8s	5s	+60%
`mlk_poly_rej_uniform_x4`	✅	8s	7s	+14%
`mlk_fqmul`	✅	7s	7s	+0%
`mlk_invntt_layer`	✅	7s	5s	+40%
`mlk_keccak_squeeze_once`	✅	7s	8s	-12%
`mlk_keccak_squeezeblocks`	✅	7s	6s	+17%
`mlk_ntt_butterfly_block`	✅	7s	10s	-30%
`mlk_poly_frombytes_native`	✅	7s	11s	-36%
`poly_decompress_d10_native_x86_64`	✅	7s	3s	+133%
`kem_keypair_derand`	✅	6s	3s	+100%
`kem_check_sk`	✅	5s	2s	+150%
`kem_dec`	✅	5s	5s	+0%
`mlk_poly_compress_d10_c`	✅	5s	4s	+25%
`mlk_poly_decompress_d10_c`	✅	5s	6s	-17%
`mlk_poly_decompress_d5_c`	✅	5s	3s	+67%
`mlk_poly_getnoise_eta1122_4x`	✅	5s	3s	+67%
`mlk_polymat_permute_bitrev_to_custom`	✅	5s	7s	-29%
`mlk_shake256x4`	✅	5s	2s	+150%
`ntt_native_aarch64`	✅	5s	3s	+67%
`sys_check_capability`	✅	5s	1s	+400%
`intt_native_x86_64`	✅	4s	4s	+0%
`mlk_ct_cmov_zero`	✅	4s	3s	+33%
`mlk_ct_sel_int16`	✅	4s	1s	+300%
`mlk_gen_matrix_serial`	✅	4s	6s	-33%
`mlk_poly_add`	✅	4s	5s	-20%
`mlk_poly_compress_d11_native`	✅	4s	3s	+33%
`mlk_poly_compress_d4`	✅	4s	4s	+0%
`mlk_poly_compress_d4_c`	✅	4s	3s	+33%
`mlk_poly_getnoise_eta1_4x`	✅	4s	1s	+300%
`mlk_poly_mulcache_compute`	✅	4s	3s	+33%
`mlk_poly_mulcache_compute_native`	✅	4s	1s	+300%
`mlk_poly_tobytes_native`	✅	4s	3s	+33%
`mlk_polyvec_decompress_du`	✅	4s	3s	+33%
`mlk_polyvec_ntt`	✅	4s	1s	+300%
`mlk_polyvec_permute_bitrev_to_custom`	✅	4s	2s	+100%
`mlk_scalar_compress_d1`	✅	4s	2s	+100%
`mlk_scalar_decompress_d10`	✅	4s	3s	+33%
`poly_compress_d10_native_x86_64`	✅	4s	2s	+100%
`poly_compress_d5_native_x86_64`	✅	4s	2s	+100%
`poly_tobytes_native_aarch64`	✅	4s	1s	+300%
`polyvec_basemul_acc_montgomery_cached_k4_native_x86_64`	✅	4s	3s	+33%
`intt_native_aarch64`	✅	3s	2s	+50%
`keccak_f1600_x1_native_aarch64_v84a`	✅	3s	1s	+200%
`keccakf1600x4_xor_bytes_native`	✅	3s	3s	+0%
`kem_check_pk`	✅	3s	3s	+0%
`kem_enc`	✅	3s	2s	+50%
`kem_enc_derand`	✅	3s	2s	+50%
`kem_keypair`	✅	3s	2s	+50%
`mlk_ct_get_optblocker_u32`	✅	3s	1s	+200%
`mlk_gen_matrix`	✅	3s	6s	-50%
`mlk_keccak_absorb_once`	✅	3s	3s	+0%
`mlk_keccakf1600_permute`	✅	3s	4s	-25%
`mlk_keccakf1600x4_xor_bytes`	✅	3s	2s	+50%
`mlk_montgomery_reduce`	✅	3s	3s	+0%
`mlk_poly_compress_d10`	✅	3s	2s	+50%
`mlk_poly_compress_du`	✅	3s	2s	+50%
`mlk_poly_decompress_d10`	✅	3s	2s	+50%
`mlk_poly_decompress_d4_c`	✅	3s	2s	+50%
`mlk_poly_decompress_d5_native`	✅	3s	2s	+50%
`mlk_poly_decompress_du`	✅	3s	1s	+200%
`mlk_poly_invntt_tomont`	✅	3s	1s	+200%
`mlk_poly_invntt_tomont_c`	✅	3s	1s	+200%
`mlk_poly_mulcache_compute_c`	✅	3s	2s	+50%
`mlk_polyvec_compress_du`	✅	3s	3s	+0%
`mlk_polyvec_mulcache_compute`	✅	3s	3s	+0%
`mlk_scalar_compress_d5`	✅	3s	4s	-25%
`mlk_scalar_decompress_d11`	✅	3s	3s	+0%
`mlk_scalar_decompress_d4`	✅	3s	1s	+200%
`mlk_scalar_decompress_d5`	✅	3s	1s	+200%
`poly_decompress_d4_native_x86_64`	✅	3s	5s	-40%
`poly_decompress_d5_native_x86_64`	✅	3s	2s	+50%
`poly_frombytes_native_x86_64`	✅	3s	6s	-50%
`poly_mulcache_compute_native_aarch64`	✅	3s	2s	+50%
`poly_reduce_native_x86_64`	✅	3s	3s	+0%
`poly_tobytes_native_x86_64`	✅	3s	2s	+50%
`rej_uniform_native`	✅	3s	4s	-25%
`rej_uniform_native_aarch64`	✅	3s	4s	-25%
`keccak_f1600_x4_native_avx2`	✅	2s	3s	-33%
`mlk_barrett_reduce`	✅	2s	4s	-50%
`mlk_check_pct`	✅	2s	2s	+0%
`mlk_ct_cmask_neg_i16`	✅	2s	4s	-50%
`mlk_ct_cmask_nonzero_u8`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_i32`	✅	2s	2s	+0%
`mlk_ct_get_optblocker_u8`	✅	2s	2s	+0%
`mlk_ct_memcmp`	✅	2s	3s	-33%
`mlk_ct_sel_uint8`	✅	2s	3s	-33%
`mlk_keccakf1600_extract_bytes`	✅	2s	2s	+0%
`mlk_keccakf1600_extract_bytes (big endian)`	✅	2s	1s	+100%
`mlk_keccakf1600_xor_bytes`	✅	2s	3s	-33%
`mlk_keccakf1600_xor_bytes (big endian)`	✅	2s	2s	+0%
`mlk_keccakf1600x4_extract_bytes`	✅	2s	3s	-33%
`mlk_matvec_mul`	✅	2s	5s	-60%
`mlk_poly_cbd_eta1`	✅	2s	3s	-33%
`mlk_poly_cbd_eta2`	✅	2s	1s	+100%
`mlk_poly_compress_d10_native`	✅	2s	2s	+0%
`mlk_poly_compress_d11_c`	✅	2s	2s	+0%
`mlk_poly_compress_d5`	✅	2s	4s	-50%
`mlk_poly_compress_d5_c`	✅	2s	1s	+100%
`mlk_poly_compress_d5_native`	✅	2s	2s	+0%
`mlk_poly_decompress_d11`	✅	2s	3s	-33%
`mlk_poly_decompress_d11_c`	✅	2s	2s	+0%
`mlk_poly_decompress_d11_native`	✅	2s	2s	+0%
`mlk_poly_decompress_d4`	✅	2s	1s	+100%
`mlk_poly_decompress_d5`	✅	2s	2s	+0%
`mlk_poly_decompress_dv`	✅	2s	4s	-50%
`mlk_poly_frombytes`	✅	2s	2s	+0%
`mlk_poly_frombytes_c`	✅	2s	3s	-33%
`mlk_poly_getnoise_eta1_4x_native`	✅	2s	3s	-33%
`mlk_poly_ntt`	✅	2s	4s	-50%
`mlk_poly_reduce_c`	✅	2s	1s	+100%
`mlk_poly_sub`	✅	2s	3s	-33%
`mlk_poly_tobytes`	✅	2s	2s	+0%
`mlk_poly_tobytes_c`	✅	2s	3s	-33%
`mlk_poly_tomont`	✅	2s	3s	-33%
`mlk_poly_tomont_c`	✅	2s	2s	+0%
`mlk_poly_tomont_native`	✅	2s	2s	+0%
`mlk_poly_tomsg`	✅	2s	2s	+0%
`mlk_polyvec_basemul_acc_montgomery_cached`	✅	2s	2s	+0%
`mlk_polyvec_frombytes`	✅	2s	1s	+100%
`mlk_polyvec_invntt_tomont`	✅	2s	3s	-33%
`mlk_polyvec_permute_bitrev_to_custom_native`	✅	2s	2s	+0%
`mlk_polyvec_reduce`	✅	2s	2s	+0%
`mlk_polyvec_tobytes`	✅	2s	1s	+100%
`mlk_polyvec_tomont`	✅	2s	2s	+0%
`mlk_rej_uniform`	✅	2s	1s	+100%
`mlk_scalar_compress_d10`	✅	2s	4s	-50%
`mlk_scalar_compress_d11`	✅	2s	2s	+0%
`mlk_scalar_compress_d4`	✅	2s	2s	+0%
`mlk_scalar_signed_to_unsigned_q`	✅	2s	2s	+0%
`mlk_sha3_256`	✅	2s	4s	-50%
`mlk_sha3_512`	✅	2s	2s	+0%
`mlk_shake128x4_absorb_once`	✅	2s	1s	+100%
`mlk_shake128x4_squeezeblocks`	✅	2s	2s	+0%
`mlk_shake256`	✅	2s	2s	+0%
`mlk_value_barrier_i32`	✅	2s	3s	-33%
`mlk_value_barrier_u32`	✅	2s	3s	-33%
`mlk_value_barrier_u8`	✅	2s	1s	+100%
`ntt_native_x86_64`	✅	2s	1s	+100%
`nttunpack_native_x86_64`	✅	2s	3s	-33%
`poly_compress_d11_native_x86_64`	✅	2s	3s	-33%
`poly_compress_d4_native_x86_64`	✅	2s	2s	+0%
`poly_getnoise_eta1122_4x_native`	✅	2s	3s	-33%
`poly_invntt_tomont_native`	✅	2s	2s	+0%
`poly_mulcache_compute_native_x86_64`	✅	2s	2s	+0%
`poly_reduce_native_aarch64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k2_native_aarch64`	✅	2s	2s	+0%
`polyvec_basemul_acc_montgomery_cached_k2_native_x86_64`	✅	2s	3s	-33%
`polyvec_basemul_acc_montgomery_cached_k3_native_aarch64`	✅	2s	4s	-50%
`polyvec_basemul_acc_montgomery_cached_k3_native_x86_64`	✅	2s	4s	-50%
`polyvec_basemul_acc_montgomery_cached_k4_native_aarch64`	✅	2s	2s	+0%
`rej_uniform_native_x86_64`	✅	2s	5s	-60%
`keccak_f1600_x1_native_aarch64`	✅	1s	2s	-50%
`keccak_f1600_x4_native_aarch64_v84a`	✅	1s	2s	-50%
`keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid`	✅	1s	2s	-50%
`keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid`	✅	1s	4s	-75%
`keccakf1600x4_extract_bytes_native`	✅	1s	2s	-50%
`mlk_ct_cmask_nonzero_u16`	✅	1s	3s	-67%
`mlk_keccakf1600x4_permute`	✅	1s	1s	+0%
`mlk_keypair_getnoise_eta1`	✅	1s	4s	-75%
`mlk_poly_compress_d11`	✅	1s	1s	+0%
`mlk_poly_compress_d4_native`	✅	1s	2s	-50%
`mlk_poly_compress_dv`	✅	1s	3s	-67%
`mlk_poly_getnoise_eta2`	✅	1s	3s	-67%
`mlk_poly_ntt_c`	✅	1s	2s	-50%
`mlk_poly_reduce`	✅	1s	1s	+0%
`mlk_shake128_absorb_once`	✅	1s	2s	-50%
`mlk_shake128_squeezeblocks`	✅	1s	2s	-50%
`poly_decompress_d11_native_x86_64`	✅	1s	2s	-50%
`poly_tomont_native_aarch64`	✅	1s	1s	+0%
`poly_tomont_native_x86_64`	✅	1s	2s	-50%

Replace aligned_alloc + MLK_ALIGN_UP with posix_memalign in custom_heap_alloc_config.h. Unlike aligned_alloc, posix_memalign does not require the size to be a multiple of the alignment, removing the need for MLK_ALIGN_UP rounding. This ensures that allocations are exact-sized, allowing memory-safety tests like valgrind and ASan to detect overflows at precise buffer boundaries. On Windows, where posix_memalign is not available, we use _aligned_malloc instead. This, too, does not require the size to be a multiple of the alignment. Signed-off-by: Hanno Becker <beckphan@amazon.co.uk>

Replace all stack-allocated buffers in test_unit.c with heap allocations via MLK_ALLOC/MLK_FREE, using the custom_heap_alloc_config. This enables valgrind to detect buffer overflows in assembly backends, which operate on these buffers. Build the unit test objects with custom_heap_alloc_config.h by adding the appropriate -DMLK_CONFIG_FILE, -std=c11, and -D_GNU_SOURCE flags in components.mk. Signed-off-by: Hanno Becker <beckphan@amazon.co.uk>

Add native-vs-C consistency tests for previously untested backends: - mlk_rej_uniform_native: compare against mlk_rej_uniform_c - mlk_poly_compress_d{4,5,10,11}_native: compare against C reference - mlk_poly_decompress_d{4,5,10,11}_native: compare against C reference These tests call the assembly backends directly with heap-allocated buffers, enabling valgrind to detect buffer overflows. In particular, the rej_uniform test would have caught the 4-byte overread in the AVX2 rejection sampling fixed in commit f10b801. Previous ad-hoc tests for detecting overflow in (de)compression routines are now subsumed by the unit tests, and removed. Signed-off-by: Hanno Becker <beckphan@amazon.co.uk>

Add unit_valgrind job to ci.yml that runs the unit tests under valgrind on x86_64 and aarch64 runners. This catches buffer overflows in hand-written assembly that ASan cannot detect, since ASan only instruments compiler-generated code. Signed-off-by: Hanno Becker <beckphan@amazon.co.uk>

hanno-becker force-pushed the unit_valgrind branch 4 times, most recently from 88001da to d97ce09 Compare March 19, 2026 21:03

hanno-becker added 2 commits March 20, 2026 04:28

hanno-becker force-pushed the unit_valgrind branch from d97ce09 to 79a002b Compare March 20, 2026 05:36

hanno-becker added 2 commits March 20, 2026 05:42

hanno-becker force-pushed the unit_valgrind branch from 79a002b to 72e929a Compare March 20, 2026 05:42

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Use heap allocation + valgrind in backend unit test#1633

Use heap allocation + valgrind in backend unit test#1633
hanno-becker wants to merge 4 commits intomainfrom
unit_valgrind

hanno-becker commented Mar 19, 2026

Uh oh!

oqs-bot commented Mar 19, 2026 •

edited

Loading

Uh oh!

oqs-bot commented Mar 19, 2026 •

edited

Loading

Uh oh!

oqs-bot commented Mar 19, 2026 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

hanno-becker commented Mar 19, 2026

Uh oh!

oqs-bot commented Mar 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-KEM-512)

Uh oh!

oqs-bot commented Mar 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-KEM-1024)

Uh oh!

oqs-bot commented Mar 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CBMC Results (ML-KEM-768)

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

oqs-bot commented Mar 19, 2026 •

edited

Loading

oqs-bot commented Mar 19, 2026 •

edited

Loading

oqs-bot commented Mar 19, 2026 •

edited

Loading