Skip to content

Armv8.1-M: Add CFI directives for stack unwinding #1558

Open
mkannwischer wants to merge 2 commits intomainfrom
armv8-cfify
Open

Armv8.1-M: Add CFI directives for stack unwinding #1558
mkannwischer wants to merge 2 commits intomainfrom
armv8-cfify

Conversation

@mkannwischer
Copy link
Contributor

Extend scripts/cfify with Armv8.1-M architecture support, handling
push/pop, vpush/vpop, sub/add sp, and bx lr.
Enable cfify for armv81m in scripts/autogen.

Neither r3 nor r12 is callee-saved according to AAPCS.

Signed-off-by: Matthias J. Kannwischer <matthias@kannwischer.eu>
Extend scripts/cfify with Armv8.1-M architecture support, handling
push/pop, vpush/vpop, sub/add sp, and bx lr.
Enable cfify for armv81m in scripts/autogen.

- Resolves #1517

Signed-off-by: Matthias J. Kannwischer <matthias@kannwischer.eu>
@oqs-bot
Copy link
Contributor

oqs-bot commented Feb 9, 2026

CBMC Results (ML-KEM-512)

Full Results (151 proofs)
Proof Status Current Previous Change
**TOTAL** 1191s 1232s -3.3%
mlk_indcpa_keypair_derand 183s 190s -4%
mlk_indcpa_enc 161s 163s -1%
mlk_keccak_squeezeblocks_x4 151s 167s -10%
mlk_rej_uniform_c 75s 85s -12%
mlk_polyvec_basemul_acc_montgomery_cached_c 45s 41s +10%
mlk_poly_rej_uniform 42s 43s -2%
poly_ntt_native 27s 27s +0%
mlk_polyvec_add 26s 25s +4%
mlk_ntt_layer 22s 22s +0%
polyvec_basemul_acc_montgomery_cached_native 21s 21s +0%
keccakf1600x4_permute_native_x4 19s 18s +6%
mlk_poly_reduce_native 15s 17s -12%
mlk_indcpa_dec 10s 11s -9%
mlk_poly_sub 10s 9s +11%
mlk_keccak_absorb_once_x4 9s 8s +12%
mlk_ntt_butterfly_block 9s 9s +0%
mlk_poly_rej_uniform_x4 9s 8s +12%
mlk_keccak_squeeze_once 8s 7s +14%
mlk_keccak_squeezeblocks 7s 8s -12%
mlk_poly_frombytes_native 7s 10s -30%
poly_frombytes_native_x86_64 7s 4s +75%
mlk_poly_frommsg 6s 7s -14%
keccakf1600_permute_native 5s 5s +0%
kem_dec 5s 4s +25%
mlk_fqmul 5s 7s -29%
mlk_keccakf1600x4_xor_bytes 5s 2s +150%
mlk_poly_frombytes_c 5s 1s +400%
mlk_poly_getnoise_eta1_4x 5s 3s +67%
poly_getnoise_eta1122_4x_native 5s 3s +67%
kem_check_pk 4s 6s -33%
kem_enc 4s 2s +100%
mlk_check_pct 4s 4s +0%
mlk_invntt_layer 4s 6s -33%
mlk_keccak_absorb_once 4s 5s -20%
mlk_keccakf1600_permute 4s 4s +0%
mlk_poly_cbd_eta1 4s 4s +0%
mlk_poly_frombytes 4s 4s +0%
mlk_poly_mulcache_compute 4s 1s +300%
mlk_polymat_permute_bitrev_to_custom 4s 5s -20%
mlk_polyvec_compress_du 4s 4s +0%
mlk_scalar_signed_to_unsigned_q 4s 2s +100%
mlk_shake256x4 4s 4s +0%
ntt_native_aarch64 4s 1s +300%
ntt_native_x86_64 4s 2s +100%
poly_reduce_native_x86_64 4s 1s +300%
polyvec_basemul_acc_montgomery_cached_k3_native_x86_64 4s 1s +300%
keccak_f1600_x4_native_aarch64_v84a 3s 2s +50%
keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid 3s 3s +0%
keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid 3s 2s +50%
kem_check_sk 3s 3s +0%
kem_enc_derand 3s 4s -25%
mlk_barrett_reduce 3s 2s +50%
mlk_ct_cmask_nonzero_u8 3s 1s +200%
mlk_ct_get_optblocker_u32 3s 2s +50%
mlk_keccakf1600_extract_bytes (big endian) 3s 2s +50%
mlk_keccakf1600_xor_bytes 3s 3s +0%
mlk_keccakf1600x4_extract_bytes 3s 2s +50%
mlk_montgomery_reduce 3s 2s +50%
mlk_poly_decompress_du 3s 2s +50%
mlk_poly_getnoise_eta1122_4x 3s 5s -40%
mlk_poly_getnoise_eta1_4x_native 3s 2s +50%
mlk_poly_tobytes_native 3s 2s +50%
mlk_poly_tomsg 3s 2s +50%
mlk_polyvec_decompress_du 3s 4s -25%
mlk_polyvec_invntt_tomont 3s 2s +50%
mlk_polyvec_ntt 3s 2s +50%
mlk_polyvec_permute_bitrev_to_custom_native 3s 2s +50%
mlk_polyvec_tobytes 3s 2s +50%
mlk_scalar_compress_d10 3s 4s -25%
mlk_scalar_compress_d5 3s 2s +50%
mlk_scalar_decompress_d5 3s 2s +50%
mlk_shake256 3s 2s +50%
nttunpack_native_x86_64 3s 3s +0%
poly_mulcache_compute_native_aarch64 3s 4s -25%
poly_mulcache_compute_native_x86_64 3s 4s -25%
rej_uniform_native 3s 3s +0%
rej_uniform_native_aarch64 3s 2s +50%
rej_uniform_native_x86_64 3s 2s +50%
sys_check_capability 3s 1s +200%
intt_native_x86_64 2s 4s -50%
keccak_f1600_x1_native_aarch64 2s 1s +100%
kem_keypair 2s 1s +100%
mlk_ct_cmov_zero 2s 2s +0%
mlk_ct_get_optblocker_i32 2s 2s +0%
mlk_ct_sel_int16 2s 3s -33%
mlk_ct_sel_uint8 2s 2s +0%
mlk_gen_matrix 2s 2s +0%
mlk_gen_matrix_serial 2s 3s -33%
mlk_keccakf1600_extract_bytes 2s 1s +100%
mlk_keccakf1600_xor_bytes (big endian) 2s 2s +0%
mlk_keccakf1600x4_permute 2s 2s +0%
mlk_poly_add 2s 2s +0%
mlk_poly_cbd_eta2 2s 2s +0%
mlk_poly_compress_du 2s 3s -33%
mlk_poly_getnoise_eta2 2s 2s +0%
mlk_poly_invntt_tomont 2s 2s +0%
mlk_poly_invntt_tomont_c 2s 3s -33%
mlk_poly_mulcache_compute_c 2s 3s -33%
mlk_poly_mulcache_compute_native 2s 4s -50%
mlk_poly_ntt 2s 3s -33%
mlk_poly_reduce 2s 2s +0%
mlk_poly_reduce_c 2s 2s +0%
mlk_poly_tobytes 2s 3s -33%
mlk_poly_tobytes_c 2s 2s +0%
mlk_poly_tomont 2s 2s +0%
mlk_poly_tomont_c 2s 2s +0%
mlk_poly_tomont_native 2s 2s +0%
mlk_polyvec_basemul_acc_montgomery_cached 2s 3s -33%
mlk_polyvec_mulcache_compute 2s 2s +0%
mlk_polyvec_tomont 2s 4s -50%
mlk_rej_uniform 2s 4s -50%
mlk_scalar_compress_d1 2s 3s -33%
mlk_scalar_decompress_d11 2s 1s +100%
mlk_sha3_256 2s 1s +100%
mlk_shake128_absorb_once 2s 1s +100%
mlk_shake128_squeezeblocks 2s 5s -60%
mlk_shake128x4_absorb_once 2s 1s +100%
mlk_value_barrier_i32 2s 2s +0%
mlk_value_barrier_u32 2s 3s -33%
mlk_value_barrier_u8 2s 3s -33%
poly_invntt_tomont_native 2s 3s -33%
poly_tobytes_native_aarch64 2s 2s +0%
poly_tobytes_native_x86_64 2s 3s -33%
poly_tomont_native_aarch64 2s 4s -50%
poly_tomont_native_x86_64 2s 2s +0%
polyvec_basemul_acc_montgomery_cached_k2_native_x86_64 2s 4s -50%
polyvec_basemul_acc_montgomery_cached_k4_native_aarch64 2s 3s -33%
intt_native_aarch64 1s 2s -50%
keccak_f1600_x1_native_aarch64_v84a 1s 2s -50%
kem_keypair_derand 1s 3s -67%
mlk_ct_cmask_neg_i16 1s 2s -50%
mlk_ct_cmask_nonzero_u16 1s 2s -50%
mlk_ct_get_optblocker_u8 1s 2s -50%
mlk_ct_memcmp 1s 1s +0%
mlk_matvec_mul 1s 2s -50%
mlk_poly_compress_dv 1s 3s -67%
mlk_poly_decompress_dv 1s 1s +0%
mlk_poly_ntt_c 1s 1s +0%
mlk_polyvec_frombytes 1s 3s -67%
mlk_polyvec_permute_bitrev_to_custom 1s 3s -67%
mlk_polyvec_reduce 1s 2s -50%
mlk_scalar_compress_d11 1s 2s -50%
mlk_scalar_compress_d4 1s 2s -50%
mlk_scalar_decompress_d10 1s 3s -67%
mlk_scalar_decompress_d4 1s 2s -50%
mlk_sha3_512 1s 3s -67%
mlk_shake128x4_squeezeblocks 1s 2s -50%
poly_reduce_native_aarch64 1s 1s +0%
polyvec_basemul_acc_montgomery_cached_k2_native_aarch64 1s 2s -50%
polyvec_basemul_acc_montgomery_cached_k3_native_aarch64 1s 2s -50%
polyvec_basemul_acc_montgomery_cached_k4_native_x86_64 1s 3s -67%

@oqs-bot
Copy link
Contributor

oqs-bot commented Feb 9, 2026

CBMC Results (ML-KEM-768)

Full Results (151 proofs)
Proof Status Current Previous Change
**TOTAL** 1380s 1424s -3.1%
mlk_indcpa_keypair_derand 255s 254s +0%
mlk_indcpa_enc 199s 217s -8%
mlk_keccak_squeezeblocks_x4 163s 160s +2%
mlk_rej_uniform_c 70s 81s -14%
polyvec_basemul_acc_montgomery_cached_native 55s 61s -10%
mlk_polyvec_basemul_acc_montgomery_cached_c 49s 54s -9%
mlk_poly_rej_uniform 37s 37s +0%
poly_ntt_native 37s 34s +9%
mlk_polyvec_add 31s 30s +3%
keccakf1600x4_permute_native_x4 20s 19s +5%
mlk_ntt_layer 19s 24s -21%
mlk_indcpa_dec 18s 17s +6%
mlk_poly_reduce_native 15s 16s -6%
mlk_keccak_absorb_once_x4 10s 9s +11%
mlk_ntt_butterfly_block 10s 10s +0%
mlk_poly_sub 10s 11s -9%
mlk_keccak_squeezeblocks 9s 8s +12%
kem_dec 8s 6s +33%
mlk_fqmul 8s 7s +14%
mlk_poly_frombytes_native 8s 9s -11%
mlk_poly_rej_uniform_x4 8s 7s +14%
keccakf1600_permute_native 6s 6s +0%
mlk_keccak_squeeze_once 6s 8s -25%
mlk_poly_frommsg 6s 7s -14%
mlk_polymat_permute_bitrev_to_custom 6s 8s -25%
mlk_gen_matrix 5s 6s -17%
mlk_invntt_layer 5s 5s +0%
mlk_keccak_absorb_once 5s 3s +67%
mlk_polyvec_permute_bitrev_to_custom_native 5s 3s +67%
nttunpack_native_x86_64 5s 3s +67%
mlk_barrett_reduce 4s 2s +100%
mlk_ct_cmask_nonzero_u16 4s 1s +300%
mlk_gen_matrix_serial 4s 6s -33%
mlk_keccakf1600_permute 4s 3s +33%
mlk_poly_compress_du 4s 4s +0%
mlk_poly_frombytes_c 4s 2s +100%
mlk_poly_ntt 4s 3s +33%
mlk_poly_ntt_c 4s 4s +0%
mlk_polyvec_ntt 4s 2s +100%
mlk_scalar_decompress_d5 4s 3s +33%
mlk_sha3_512 4s 2s +100%
mlk_shake128x4_absorb_once 4s 2s +100%
mlk_shake256x4 4s 5s -20%
poly_frombytes_native_x86_64 4s 4s +0%
poly_getnoise_eta1122_4x_native 4s 2s +100%
poly_invntt_tomont_native 4s 2s +100%
poly_tomont_native_aarch64 4s 2s +100%
polyvec_basemul_acc_montgomery_cached_k2_native_aarch64 4s 2s +100%
polyvec_basemul_acc_montgomery_cached_k2_native_x86_64 4s 4s +0%
keccak_f1600_x4_native_aarch64_v84a 3s 2s +50%
kem_check_pk 3s 4s -25%
kem_check_sk 3s 4s -25%
kem_keypair 3s 3s +0%
kem_keypair_derand 3s 1s +200%
mlk_check_pct 3s 3s +0%
mlk_ct_cmask_nonzero_u8 3s 2s +50%
mlk_keccakf1600x4_permute 3s 2s +50%
mlk_poly_cbd_eta2 3s 3s +0%
mlk_poly_getnoise_eta1_4x 3s 5s -40%
mlk_poly_getnoise_eta2 3s 3s +0%
mlk_poly_invntt_tomont 3s 2s +50%
mlk_poly_mulcache_compute 3s 3s +0%
mlk_poly_mulcache_compute_c 3s 4s -25%
mlk_poly_tobytes_c 3s 3s +0%
mlk_polyvec_basemul_acc_montgomery_cached 3s 2s +50%
mlk_polyvec_compress_du 3s 3s +0%
mlk_polyvec_decompress_du 3s 4s -25%
mlk_polyvec_frombytes 3s 1s +200%
mlk_polyvec_permute_bitrev_to_custom 3s 2s +50%
mlk_scalar_decompress_d4 3s 2s +50%
mlk_scalar_signed_to_unsigned_q 3s 4s -25%
mlk_shake256 3s 2s +50%
mlk_value_barrier_i32 3s 3s +0%
mlk_value_barrier_u32 3s 1s +200%
ntt_native_x86_64 3s 2s +50%
poly_mulcache_compute_native_aarch64 3s 3s +0%
poly_reduce_native_x86_64 3s 2s +50%
polyvec_basemul_acc_montgomery_cached_k3_native_x86_64 3s 1s +200%
polyvec_basemul_acc_montgomery_cached_k4_native_aarch64 3s 3s +0%
polyvec_basemul_acc_montgomery_cached_k4_native_x86_64 3s 3s +0%
rej_uniform_native_x86_64 3s 3s +0%
sys_check_capability 3s 1s +200%
intt_native_aarch64 2s 2s +0%
keccak_f1600_x1_native_aarch64 2s 4s -50%
keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid 2s 2s +0%
keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid 2s 2s +0%
kem_enc 2s 2s +0%
kem_enc_derand 2s 2s +0%
mlk_ct_cmask_neg_i16 2s 2s +0%
mlk_ct_get_optblocker_i32 2s 1s +100%
mlk_ct_memcmp 2s 5s -60%
mlk_ct_sel_int16 2s 3s -33%
mlk_ct_sel_uint8 2s 4s -50%
mlk_keccakf1600_extract_bytes 2s 2s +0%
mlk_keccakf1600_extract_bytes (big endian) 2s 2s +0%
mlk_keccakf1600_xor_bytes (big endian) 2s 1s +100%
mlk_matvec_mul 2s 1s +100%
mlk_montgomery_reduce 2s 1s +100%
mlk_poly_add 2s 1s +100%
mlk_poly_cbd_eta1 2s 5s -60%
mlk_poly_compress_dv 2s 5s -60%
mlk_poly_decompress_du 2s 3s -33%
mlk_poly_frombytes 2s 1s +100%
mlk_poly_getnoise_eta1122_4x 2s 2s +0%
mlk_poly_getnoise_eta1_4x_native 2s 3s -33%
mlk_poly_invntt_tomont_c 2s 2s +0%
mlk_poly_tomont 2s 2s +0%
mlk_poly_tomont_c 2s 3s -33%
mlk_poly_tomont_native 2s 3s -33%
mlk_poly_tomsg 2s 3s -33%
mlk_polyvec_invntt_tomont 2s 3s -33%
mlk_polyvec_mulcache_compute 2s 3s -33%
mlk_polyvec_reduce 2s 3s -33%
mlk_polyvec_tomont 2s 1s +100%
mlk_rej_uniform 2s 2s +0%
mlk_scalar_compress_d1 2s 2s +0%
mlk_scalar_compress_d10 2s 2s +0%
mlk_scalar_compress_d5 2s 1s +100%
mlk_scalar_decompress_d10 2s 2s +0%
mlk_scalar_decompress_d11 2s 3s -33%
mlk_shake128_squeezeblocks 2s 1s +100%
mlk_shake128x4_squeezeblocks 2s 3s -33%
mlk_value_barrier_u8 2s 3s -33%
ntt_native_aarch64 2s 3s -33%
poly_mulcache_compute_native_x86_64 2s 1s +100%
poly_reduce_native_aarch64 2s 1s +100%
poly_tobytes_native_aarch64 2s 1s +100%
poly_tobytes_native_x86_64 2s 4s -50%
poly_tomont_native_x86_64 2s 2s +0%
rej_uniform_native_aarch64 2s 3s -33%
intt_native_x86_64 1s 3s -67%
keccak_f1600_x1_native_aarch64_v84a 1s 3s -67%
mlk_ct_cmov_zero 1s 2s -50%
mlk_ct_get_optblocker_u32 1s 2s -50%
mlk_ct_get_optblocker_u8 1s 4s -75%
mlk_keccakf1600_xor_bytes 1s 2s -50%
mlk_keccakf1600x4_extract_bytes 1s 3s -67%
mlk_keccakf1600x4_xor_bytes 1s 1s +0%
mlk_poly_decompress_dv 1s 1s +0%
mlk_poly_mulcache_compute_native 1s 3s -67%
mlk_poly_reduce 1s 3s -67%
mlk_poly_reduce_c 1s 1s +0%
mlk_poly_tobytes 1s 3s -67%
mlk_poly_tobytes_native 1s 2s -50%
mlk_polyvec_tobytes 1s 3s -67%
mlk_scalar_compress_d11 1s 1s +0%
mlk_scalar_compress_d4 1s 2s -50%
mlk_sha3_256 1s 3s -67%
mlk_shake128_absorb_once 1s 4s -75%
polyvec_basemul_acc_montgomery_cached_k3_native_aarch64 1s 2s -50%
rej_uniform_native 1s 2s -50%

@oqs-bot
Copy link
Contributor

oqs-bot commented Feb 9, 2026

CBMC Results (ML-KEM-1024)

Full Results (151 proofs)
Proof Status Current Previous Change
**TOTAL** 2671s 2514s +6.2%
mlk_indcpa_enc 1493s 1270s +18%
mlk_indcpa_keypair_derand 207s 222s -7%
mlk_keccak_squeezeblocks_x4 148s 154s -4%
polyvec_basemul_acc_montgomery_cached_native 114s 127s -10%
mlk_rej_uniform_c 65s 72s -10%
mlk_polyvec_basemul_acc_montgomery_cached_c 56s 58s -3%
mlk_poly_rej_uniform 31s 33s -6%
mlk_poly_decompress_dv 22s 22s +0%
poly_ntt_native 22s 26s -15%
keccakf1600x4_permute_native_x4 19s 21s -10%
mlk_ntt_layer 17s 26s -35%
mlk_indcpa_dec 16s 20s -20%
mlk_poly_reduce_native 15s 17s -12%
mlk_polyvec_ntt 13s 16s -19%
mlk_polyvec_add 12s 11s +9%
mlk_keccak_absorb_once_x4 11s 8s +38%
mlk_poly_compress_du 10s 7s +43%
mlk_poly_sub 10s 11s -9%
mlk_gen_matrix 8s 7s +14%
mlk_keccak_squeeze_once 8s 6s +33%
mlk_keccak_squeezeblocks 8s 9s -11%
mlk_ntt_butterfly_block 8s 9s -11%
mlk_poly_frombytes_native 8s 10s -20%
mlk_poly_rej_uniform_x4 8s 8s +0%
keccakf1600_permute_native 7s 6s +17%
mlk_gen_matrix_serial 6s 6s +0%
poly_frombytes_native_x86_64 6s 4s +50%
kem_check_pk 5s 4s +25%
mlk_fqmul 5s 5s +0%
mlk_poly_mulcache_compute_c 5s 1s +400%
mlk_polyvec_permute_bitrev_to_custom 5s 2s +150%
mlk_shake256x4 5s 4s +25%
mlk_value_barrier_u32 5s 1s +400%
mlk_value_barrier_u8 5s 5s +0%
poly_tobytes_native_x86_64 5s 4s +25%
kem_dec 4s 5s -20%
mlk_ct_get_optblocker_u32 4s 3s +33%
mlk_invntt_layer 4s 6s -33%
mlk_keccak_absorb_once 4s 3s +33%
mlk_keccakf1600_permute 4s 5s -20%
mlk_poly_frommsg 4s 7s -43%
mlk_poly_ntt 4s 4s +0%
mlk_poly_tomont_native 4s 3s +33%
mlk_polyvec_basemul_acc_montgomery_cached 4s 3s +33%
mlk_polyvec_compress_du 4s 2s +100%
mlk_polyvec_mulcache_compute 4s 2s +100%
mlk_scalar_compress_d5 4s 4s +0%
poly_mulcache_compute_native_aarch64 4s 2s +100%
polyvec_basemul_acc_montgomery_cached_k2_native_aarch64 4s 1s +300%
polyvec_basemul_acc_montgomery_cached_k2_native_x86_64 4s 2s +100%
polyvec_basemul_acc_montgomery_cached_k4_native_aarch64 4s 1s +300%
keccak_f1600_x1_native_aarch64 3s 5s -40%
keccak_f1600_x1_native_aarch64_v84a 3s 2s +50%
keccak_f1600_x4_native_aarch64_v8a_scalar_hybrid 3s 3s +0%
kem_check_sk 3s 2s +50%
kem_enc_derand 3s 4s -25%
mlk_barrett_reduce 3s 2s +50%
mlk_check_pct 3s 3s +0%
mlk_ct_cmov_zero 3s 1s +200%
mlk_ct_sel_int16 3s 2s +50%
mlk_keccakf1600_extract_bytes 3s 1s +200%
mlk_keccakf1600_extract_bytes (big endian) 3s 2s +50%
mlk_keccakf1600x4_extract_bytes 3s 1s +200%
mlk_keccakf1600x4_permute 3s 1s +200%
mlk_matvec_mul 3s 2s +50%
mlk_poly_cbd_eta1 3s 1s +200%
mlk_poly_decompress_du 3s 4s -25%
mlk_poly_frombytes_c 3s 2s +50%
mlk_poly_getnoise_eta1_4x 3s 3s +0%
mlk_poly_mulcache_compute 3s 2s +50%
mlk_poly_reduce_c 3s 4s -25%
mlk_poly_tobytes_c 3s 2s +50%
mlk_poly_tomsg 3s 3s +0%
mlk_polymat_permute_bitrev_to_custom 3s 6s -50%
mlk_polyvec_decompress_du 3s 2s +50%
mlk_polyvec_permute_bitrev_to_custom_native 3s 5s -40%
mlk_polyvec_tomont 3s 4s -25%
mlk_scalar_compress_d1 3s 2s +50%
mlk_scalar_decompress_d10 3s 3s +0%
mlk_scalar_decompress_d4 3s 2s +50%
mlk_sha3_256 3s 3s +0%
mlk_shake128_squeezeblocks 3s 2s +50%
ntt_native_aarch64 3s 3s +0%
nttunpack_native_x86_64 3s 2s +50%
poly_getnoise_eta1122_4x_native 3s 2s +50%
poly_invntt_tomont_native 3s 3s +0%
poly_mulcache_compute_native_x86_64 3s 2s +50%
polyvec_basemul_acc_montgomery_cached_k3_native_x86_64 3s 3s +0%
rej_uniform_native_aarch64 3s 3s +0%
intt_native_aarch64 2s 2s +0%
intt_native_x86_64 2s 4s -50%
keccak_f1600_x4_native_aarch64_v84a 2s 1s +100%
keccak_f1600_x4_native_aarch64_v8a_v84a_scalar_hybrid 2s 1s +100%
kem_enc 2s 3s -33%
kem_keypair 2s 2s +0%
mlk_ct_cmask_neg_i16 2s 2s +0%
mlk_ct_cmask_nonzero_u16 2s 3s -33%
mlk_ct_cmask_nonzero_u8 2s 2s +0%
mlk_ct_get_optblocker_i32 2s 3s -33%
mlk_ct_get_optblocker_u8 2s 3s -33%
mlk_ct_memcmp 2s 2s +0%
mlk_ct_sel_uint8 2s 1s +100%
mlk_keccakf1600_xor_bytes 2s 2s +0%
mlk_keccakf1600_xor_bytes (big endian) 2s 2s +0%
mlk_keccakf1600x4_xor_bytes 2s 2s +0%
mlk_montgomery_reduce 2s 2s +0%
mlk_poly_compress_dv 2s 1s +100%
mlk_poly_frombytes 2s 2s +0%
mlk_poly_getnoise_eta1_4x_native 2s 5s -60%
mlk_poly_getnoise_eta2 2s 3s -33%
mlk_poly_invntt_tomont 2s 3s -33%
mlk_poly_mulcache_compute_native 2s 2s +0%
mlk_poly_tomont 2s 1s +100%
mlk_poly_tomont_c 2s 2s +0%
mlk_polyvec_frombytes 2s 2s +0%
mlk_polyvec_invntt_tomont 2s 4s -50%
mlk_polyvec_reduce 2s 3s -33%
mlk_polyvec_tobytes 2s 2s +0%
mlk_rej_uniform 2s 2s +0%
mlk_scalar_compress_d10 2s 3s -33%
mlk_scalar_decompress_d5 2s 3s -33%
mlk_sha3_512 2s 3s -33%
mlk_shake128_absorb_once 2s 3s -33%
mlk_shake128x4_squeezeblocks 2s 2s +0%
mlk_shake256 2s 2s +0%
mlk_value_barrier_i32 2s 2s +0%
poly_reduce_native_x86_64 2s 4s -50%
poly_tomont_native_aarch64 2s 1s +100%
polyvec_basemul_acc_montgomery_cached_k3_native_aarch64 2s 3s -33%
polyvec_basemul_acc_montgomery_cached_k4_native_x86_64 2s 2s +0%
rej_uniform_native 2s 4s -50%
rej_uniform_native_x86_64 2s 2s +0%
kem_keypair_derand 1s 2s -50%
mlk_poly_add 1s 3s -67%
mlk_poly_cbd_eta2 1s 3s -67%
mlk_poly_getnoise_eta1122_4x 1s 3s -67%
mlk_poly_invntt_tomont_c 1s 4s -75%
mlk_poly_ntt_c 1s 2s -50%
mlk_poly_reduce 1s 2s -50%
mlk_poly_tobytes 1s 3s -67%
mlk_poly_tobytes_native 1s 2s -50%
mlk_scalar_compress_d11 1s 2s -50%
mlk_scalar_compress_d4 1s 2s -50%
mlk_scalar_decompress_d11 1s 2s -50%
mlk_scalar_signed_to_unsigned_q 1s 3s -67%
mlk_shake128x4_absorb_once 1s 3s -67%
ntt_native_x86_64 1s 2s -50%
poly_reduce_native_aarch64 1s 2s -50%
poly_tobytes_native_aarch64 1s 4s -75%
poly_tomont_native_x86_64 1s 2s -50%
sys_check_capability 1s 1s +0%

@mkannwischer mkannwischer marked this pull request as ready for review February 10, 2026 01:57
@mkannwischer mkannwischer requested a review from a team as a code owner February 10, 2026 01:57
mkannwischer added a commit to pq-code-package/mldsa-native that referenced this pull request Feb 11, 2026
Extend scripts/cfify with Armv8.1-M architecture support, handling
push/pop, vpush/vpop, sub/add sp, and bx lr.
Enable cfify for armv81m in scripts/autogen.

Also removes pushing/poping r3 and r12 in keccak_f1600_x4_mve_asm - they
do not have to be preserved according ot the AAPCS.

- Ports pq-code-package/mlkem-native#1558

Signed-off-by: Matthias J. Kannwischer <matthias@kannwischer.eu>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Armv8.1-M: Extend cfify for the Armv8.1-M backend

2 participants