code wiki / (root) / nx_pe_perf_probe.nx

nx_pe_perf_probe.nx

buildroot/runtime/nx_pe_perf_probe.nx

2137 B61 linesdepth 6pulls 8 transitivereach 0 importersview sourcekind probetopic pe
docsdependenciesstructsconstsfunctions

about

nx_pe_perf_probe.nx -- measure, IN THE WINDOWS PE, where the transfer time goes. Loops 100x the live per-page workload (23 records x 13KB) so the ~15ms-granular PE clock has signal. Splits AEAD-decrypt compute vs sys_mmap(13KB) cost. Build as PE (overwrites nishi_fetch.exe), run, read stdout. expect_exit: 0

dependencies 2 imports · 0 importers

nx_syscalls.nx nx_chacha20_poly1305.nx nx_pe_perf_probe.nx

imports: nx_syscalls.nxnx_chacha20_poly1305.nx

imported by: nobody (leaf or entry point)

call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown

main sys_mmap nx_chacha20_poly1305_encry sys_mmap ↻ aead_derive_otk sys_mmap ↻ chacha20_block sys_mmap ↻ _chacha20_set_consts_key load_u32_le load_u32_le ↻ chacha20_perm sys_mmap ↻ chacha20_perm_core store_u32_le chacha20_encrypt sys_mmap ↻ _chacha20_set_consts_key ↻ load_u32_le ↻ chacha20_perm_core ↻ aead_build_mac_data aead_put_u64_le poly1305_mac sys_mmap ↻ poly1305_clamp p_load_u32_le p_store_u32_le sys_now_ms sys_mmap ↻ sys_clock_gettime_mono nx_chacha20_poly1305_decry sys_mmap ↻ sys_write aead_derive_otk ↻ sys_now_ms ↻ aead_build_mac_data ↻ poly1305_mac ↻ poly1305_tag_equal chacha20_encrypt ↻ p_ps

structs

none

consts

10const K_MAGIC_13000: i64 = 13000
11const K_MAGIC_2300: i64 = 2300

functions

13func p_ps(s: *u8, n: i64) -> i64 { sys_write(1, s, n); return 0 }
called by 1: main calls 1: sys_write
14func p_pn(v: i64) -> i64
called by 1: main calls 2: sys_mmapsys_write
25func main() -> i64
calls 7: sys_mmapnx_chacha20_poly1305_encryptsys_now_msnx_chacha20_poly1305_decryptp_psp_pn+1