apfs.ksy
apfs.ksy copied to clipboard
cugu
Significant difference between space manager blocks documented here vs. Apple docs
This is a table of the differences between what's documented here and what Apple eventually released in the docs:
| Current KSY | Apple Docs |
|---|---|
| sm_o (obj_phys_t) | |
| block_size (u4) | sm_block_size (uint32_t) |
| blocks_per_chunk (u4) | sm_blocks_per_chunk (uint32_t) |
| chunks_per_cib (u4) | sm_chunks_per_cib (uint32_t) |
| cibs_per_cab (u4) | sm_cibs_per_cab (uint32_t) |
| block_count (u4) | sm_dev[0].sm_block_count (uint64_t) |
| chunk_count (u4) | |
| cib_count (u4) | sm_dev[0].sm_chunk_count (uint64_t) |
| cab_count (u4) | |
| entry_count (u4) | sm_dev[0].sm_cib_count (uint32_t) |
| unknown_68 (u4) | sm_dev[0].sm_cab_count (uint32_t) |
| free_block_count (u8) | sm_dev[0].sm_free_count (uint64_t) |
| entries_offset (u4) | sm_dev[0].sm_addr_offset (uint32_t) |
| unknown_84 (92) | sm_dev[0].sm_reserved (uint32_t) |
| sm_dev[0].sm_reserved2 (uint64_t) | |
| sm_dev[1].sm_block_count (uint64_t) | |
| sm_dev[1].sm_chunk_count (uint64_t) | |
| sm_dev[1].sm_cib_count (uint32_t) | |
| sm_dev[1].sm_cab_count (uint32_t) | |
| sm_dev[1].sm_free_count (uint64_t) | |
| sm_dev[1].sm_addr_offset (uint32_t) | |
| sm_dev[1].sm_reserved (uint32_t) | |
| sm_dev[1].sm_reserved2 (uint64_t) | |
| sm_flags (uint32_t) | |
| sm_ip_bm_tx_multiplier (uint32_t) | |
| sm_ip_block_count (uint64_t) | |
| sm_ip_bm_size_in_blocks (uint32_t) | |
| sm_ip_bm_block_count (uint32_t) | |
| sm_ip_bm_base (paddr_t) | |
| prev_spaceman_internal_pool_block (u8) | sm_ip_base (paddr_t) |
| sm_fs_reserve_block_count (uint64_t) | |
| sm_fs_reserve_alloc_count (uint64_t) | |
| sm_fq[SFQ_COUNT] (spaceman_free_queue_t) | |
| sm_ip_bm_free_head (uint16_t) | |
| sm_ip_bm_free_tail (uint16_t) | |
| sm_ip_bm_xid_offset (uint32_t) | |
| sm_ip_bitmap_offset (uint32_t) | |
| sm_ip_bm_free_next_offset (uint32_t) | |
| sm_version (uint32_t) | |
| sm_struct_size (uint32_t) | |
| sm_datazone (spaceman_datazone_info_phys_t) |
I tried updating my own parsing code to match theirs, but some of my existing data didn't seem to parse using their documented structure. However, I noticed that there's a version field in the structure, and even in my own data set, I see many different values for that field. I haven't quite figured out whether we've screwed up writing the parsing code, or whether there really are multiple versions of the structure for different versions of APFS, or it could possibly even be a little of each.
Thanks for the listing. The space manager was never really relevant to me and I did not spend time to incorporate the information Apple released so far... I'm happy to merge any pull requests, though ;-)
