fstool 0.2.0

Build disk images and filesystems (ext2/3/4, MBR, GPT) from a directory tree and TOML spec, in the spirit of genext2fs.
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
//! Object Map (`omap_phys_t`) — translates virtual object IDs into physical
//! block addresses.
//!
//! Reference: *Apple File System Reference*, section "Object Map".
//!
//! ```text
//! omap_phys_t (header for the omap itself)
//!    0   om_o                obj_phys_t (32 B)
//!   32   om_flags            u32
//!   36   om_snap_count       u32
//!   40   om_tree_type         u32
//!   44   om_snapshot_tree_type u32
//!   48   om_tree_oid         u64  ← physical block of the omap B-tree root
//!   56   om_snapshot_tree_oid u64
//!   64   om_most_recent_snap  u64
//!   72   om_pending_revert_min u64
//!   80   om_pending_revert_max u64
//! ```
//!
//! ```text
//! omap_key_t (16 B):
//!   0   ok_oid  u64
//!   8   ok_xid  u64
//! omap_val_t (16 B):
//!   0   ov_flags  u32
//!   4   ov_size   u32
//!   8   ov_paddr  u64
//! ```
//!
//! The omap is a fixed-KV B-tree (16/16). Lookup of `(oid, xid)` walks the
//! tree as usual; we pick the entry with the largest `xid <= target_xid`.

use super::btree::{BTreeNode, NodeCache};
use super::obj::{OBJECT_TYPE_OMAP, ObjPhys};

/// Decoded `omap_phys_t`.
#[derive(Debug, Clone)]
pub struct OmapPhys {
    pub obj: ObjPhys,
    pub flags: u32,
    pub snap_count: u32,
    pub tree_type: u32,
    pub tree_oid: u64,
}

impl OmapPhys {
    pub const MIN_SIZE: usize = 88;
    pub fn decode(buf: &[u8]) -> crate::Result<Self> {
        if buf.len() < Self::MIN_SIZE {
            return Err(crate::Error::InvalidImage(
                "apfs: omap_phys buffer too short".into(),
            ));
        }
        let obj = ObjPhys::decode(buf)?;
        if obj.obj_type() != OBJECT_TYPE_OMAP {
            return Err(crate::Error::InvalidImage(format!(
                "apfs: o_type {:#x} is not OMAP",
                obj.obj_type()
            )));
        }
        let flags = u32::from_le_bytes(buf[32..36].try_into().unwrap());
        let snap_count = u32::from_le_bytes(buf[36..40].try_into().unwrap());
        let tree_type = u32::from_le_bytes(buf[40..44].try_into().unwrap());
        let tree_oid = u64::from_le_bytes(buf[48..56].try_into().unwrap());
        Ok(Self {
            obj,
            flags,
            snap_count,
            tree_type,
            tree_oid,
        })
    }
}

/// A decoded omap key.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub struct OmapKey {
    pub oid: u64,
    pub xid: u64,
}

impl OmapKey {
    pub const SIZE: usize = 16;
    pub fn decode(buf: &[u8]) -> crate::Result<Self> {
        if buf.len() < Self::SIZE {
            return Err(crate::Error::InvalidImage(
                "apfs: omap_key too short".into(),
            ));
        }
        Ok(Self {
            oid: u64::from_le_bytes(buf[0..8].try_into().unwrap()),
            xid: u64::from_le_bytes(buf[8..16].try_into().unwrap()),
        })
    }
}

/// A decoded omap value.
#[derive(Debug, Clone, Copy)]
pub struct OmapVal {
    pub flags: u32,
    pub size: u32,
    pub paddr: u64,
}

impl OmapVal {
    pub const SIZE: usize = 16;
    pub fn decode(buf: &[u8]) -> crate::Result<Self> {
        if buf.len() < Self::SIZE {
            return Err(crate::Error::InvalidImage(
                "apfs: omap_val too short".into(),
            ));
        }
        Ok(Self {
            flags: u32::from_le_bytes(buf[0..4].try_into().unwrap()),
            size: u32::from_le_bytes(buf[4..8].try_into().unwrap()),
            paddr: u64::from_le_bytes(buf[8..16].try_into().unwrap()),
        })
    }
}

/// Compare two omap keys per the spec's ordering rule: ascending `oid`
/// first, then ascending `xid`.
fn cmp_omap_key(a: &OmapKey, b: &OmapKey) -> std::cmp::Ordering {
    a.oid.cmp(&b.oid).then(a.xid.cmp(&b.xid))
}

/// Search an omap B-tree of arbitrary depth for the entry matching
/// `target_oid` with the largest `xid <= target_xid`. Returns the
/// resolved `OmapVal` or `None` when no such entry exists.
///
/// `read_block` is supplied by the caller; given a physical block
/// number it must fill the passed buffer (length = block size) with
/// that block's contents. The function makes no assumption about the
/// underlying device.
///
/// Internal nodes are descended by binary-searching for the smallest
/// key that compares **strictly greater** than the target `(oid,
/// target_xid)`, then stepping back one entry — the standard "largest
/// key ≤ target" rule, except that here every internal-node separator
/// is the *smallest* key in its child subtree (APFS uses this
/// convention).
///
/// Multi-level walks consult `cache` to avoid re-fetching blocks.
/// Callers that don't want caching can pass `NodeCache::new(0)`.
pub fn lookup<F>(
    root_block: &[u8],
    target_oid: u64,
    target_xid: u64,
    read_block: &mut F,
) -> crate::Result<Option<OmapVal>>
where
    F: FnMut(u64, &mut [u8]) -> crate::Result<()>,
{
    let mut cache = NodeCache::new(NodeCache::DEFAULT_CAP);
    lookup_with_cache(root_block, target_oid, target_xid, read_block, &mut cache)
}

/// Same as [`lookup`] but with caller-supplied cache so repeated lookups
/// in the same omap reuse loaded internal nodes.
pub fn lookup_with_cache<F>(
    root_block: &[u8],
    target_oid: u64,
    target_xid: u64,
    read_block: &mut F,
    cache: &mut NodeCache,
) -> crate::Result<Option<OmapVal>>
where
    F: FnMut(u64, &mut [u8]) -> crate::Result<()>,
{
    let target = OmapKey {
        oid: target_oid,
        xid: target_xid,
    };
    let block_size = root_block.len();
    let root = BTreeNode::decode(root_block)?;
    let (klen, vlen) = root
        .fixed_kv_size()
        .ok_or_else(|| crate::Error::InvalidImage("apfs: omap root missing btree_info".into()))?;
    if klen != OmapKey::SIZE || vlen != OmapVal::SIZE {
        return Err(crate::Error::InvalidImage(format!(
            "apfs: omap fixed kv size ({klen}, {vlen}) != (16, 16)"
        )));
    }

    if root.is_leaf() {
        return scan_leaf_for_best(&root, klen, target_oid, target_xid);
    }

    // Descend by re-binding to an owned child block each iteration.
    let child_idx = find_child_for_key(&root, klen, &target)?;
    let (_, mut child_paddr) = root.child_entry_at(child_idx, klen)?;
    let mut cur = fetch_block(child_paddr, block_size, read_block, cache)?;
    loop {
        let node = BTreeNode::decode(&cur)?;
        if node.is_leaf() {
            return scan_leaf_for_best(&node, klen, target_oid, target_xid);
        }
        let idx = find_child_for_key(&node, klen, &target)?;
        let (_, paddr) = node.child_entry_at(idx, klen)?;
        child_paddr = paddr;
        // Replace cur with the next child block (this drops the previous
        // `node` borrow since `cur` is rebound).
        cur = fetch_block(child_paddr, block_size, read_block, cache)?;
    }
}

/// Locate the child of an internal omap node whose subtree may contain
/// `target`. Per the spec, internal keys are the *first* key in their
/// subtree, so we want the largest entry whose key ≤ target.
fn find_child_for_key(node: &BTreeNode<'_>, klen: usize, target: &OmapKey) -> crate::Result<u32> {
    if node.nkeys == 0 {
        return Err(crate::Error::InvalidImage(
            "apfs: empty omap internal node".into(),
        ));
    }
    // Binary search for the largest index whose key ≤ target.
    let mut lo: i64 = 0;
    let mut hi: i64 = node.nkeys as i64 - 1;
    let mut best: i64 = 0; // default: leftmost child
    while lo <= hi {
        let mid = (lo + hi) / 2;
        let (kb, _) = node.child_entry_at(mid as u32, klen)?;
        let k = OmapKey::decode(kb)?;
        match cmp_omap_key(&k, target) {
            std::cmp::Ordering::Less | std::cmp::Ordering::Equal => {
                best = mid;
                lo = mid + 1;
            }
            std::cmp::Ordering::Greater => {
                hi = mid - 1;
            }
        }
    }
    Ok(best as u32)
}

/// Scan a leaf node for the largest-xid entry matching `target_oid`
/// with `xid <= target_xid`.
fn scan_leaf_for_best(
    node: &BTreeNode<'_>,
    klen: usize,
    target_oid: u64,
    target_xid: u64,
) -> crate::Result<Option<OmapVal>> {
    let vlen = OmapVal::SIZE;
    let mut best: Option<OmapVal> = None;
    let mut best_xid: u64 = 0;
    for i in 0..node.nkeys {
        let (kb, vb) = node.entry_at(i, klen, vlen)?;
        let k = OmapKey::decode(kb)?;
        if k.oid != target_oid {
            continue;
        }
        if k.xid > target_xid {
            continue;
        }
        if best.is_none() || k.xid > best_xid {
            best_xid = k.xid;
            best = Some(OmapVal::decode(vb)?);
        }
    }
    Ok(best)
}

fn fetch_block<F>(
    paddr: u64,
    block_size: usize,
    read_block: &mut F,
    cache: &mut NodeCache,
) -> crate::Result<Vec<u8>>
where
    F: FnMut(u64, &mut [u8]) -> crate::Result<()>,
{
    if let Some(b) = cache.get(paddr) {
        return Ok(b.to_vec());
    }
    let mut buf = vec![0u8; block_size];
    read_block(paddr, &mut buf)?;
    cache.put(paddr, buf.clone());
    Ok(buf)
}

#[cfg(test)]
mod tests {
    use super::super::btree::{BTNODE_FIXED_KV_SIZE, BTNODE_LEAF, BTNODE_ROOT, BTREE_INFO_SIZE};
    use super::super::obj::OBJECT_TYPE_BTREE_NODE;
    use super::*;

    /// Build a leaf omap node with a list of `(oid, xid, paddr)` entries
    /// in the order given. Caller is responsible for keeping the list
    /// sorted by `(oid, xid)`.
    fn build_leaf_omap_node(
        block_size: usize,
        entries: &[(u64, u64, u64)],
        is_root: bool,
    ) -> Vec<u8> {
        let mut block = vec![0u8; block_size];
        block[24..28].copy_from_slice(&OBJECT_TYPE_BTREE_NODE.to_le_bytes());
        let mut flags = BTNODE_LEAF | BTNODE_FIXED_KV_SIZE;
        if is_root {
            flags |= BTNODE_ROOT;
        }
        block[32..34].copy_from_slice(&flags.to_le_bytes());
        block[34..36].copy_from_slice(&0u16.to_le_bytes()); // level
        block[36..40].copy_from_slice(&(entries.len() as u32).to_le_bytes());
        let toc_len = entries.len() * 4;
        block[40..42].copy_from_slice(&0u16.to_le_bytes()); // toc_off
        block[42..44].copy_from_slice(&(toc_len as u16).to_le_bytes());

        let toc_base = 56;
        let keys_start = toc_base + toc_len;
        let vals_end = if is_root {
            block_size - BTREE_INFO_SIZE
        } else {
            block_size
        };

        for (i, &(oid, xid, paddr)) in entries.iter().enumerate() {
            // ToC: kvoff with k.off = i*16, v.off = (i+1)*16
            let k_off = (i * 16) as u16;
            let v_off = ((i + 1) * 16) as u16;
            block[toc_base + i * 4..toc_base + i * 4 + 2].copy_from_slice(&k_off.to_le_bytes());
            block[toc_base + i * 4 + 2..toc_base + i * 4 + 4].copy_from_slice(&v_off.to_le_bytes());

            // Key bytes at keys_start + k_off
            let ks = keys_start + k_off as usize;
            block[ks..ks + 8].copy_from_slice(&oid.to_le_bytes());
            block[ks + 8..ks + 16].copy_from_slice(&xid.to_le_bytes());

            // Value bytes at vals_end - v_off (16 B) — flags=0, size=0, paddr
            let vs = vals_end - v_off as usize;
            block[vs..vs + 4].copy_from_slice(&0u32.to_le_bytes());
            block[vs + 4..vs + 8].copy_from_slice(&0u32.to_le_bytes());
            block[vs + 8..vs + 16].copy_from_slice(&paddr.to_le_bytes());
        }

        if is_root {
            let info_off = block_size - BTREE_INFO_SIZE;
            block[info_off + 8..info_off + 12].copy_from_slice(&16u32.to_le_bytes());
            block[info_off + 12..info_off + 16].copy_from_slice(&16u32.to_le_bytes());
        }
        block
    }

    /// Build an internal omap node: keys are `(oid, xid)` separators
    /// and the values are 8-byte child paddrs (despite bt_val_size = 16
    /// for the leaves — internal nodes always store 8-byte child ptrs).
    fn build_internal_omap_node(
        block_size: usize,
        entries: &[(u64, u64, u64)],
        is_root: bool,
    ) -> Vec<u8> {
        let mut block = vec![0u8; block_size];
        block[24..28].copy_from_slice(&OBJECT_TYPE_BTREE_NODE.to_le_bytes());
        let mut flags = BTNODE_FIXED_KV_SIZE;
        if is_root {
            flags |= BTNODE_ROOT;
        }
        block[32..34].copy_from_slice(&flags.to_le_bytes());
        block[34..36].copy_from_slice(&1u16.to_le_bytes()); // level = 1
        block[36..40].copy_from_slice(&(entries.len() as u32).to_le_bytes());
        let toc_len = entries.len() * 4;
        block[40..42].copy_from_slice(&0u16.to_le_bytes());
        block[42..44].copy_from_slice(&(toc_len as u16).to_le_bytes());

        let toc_base = 56;
        let keys_start = toc_base + toc_len;
        let vals_end = if is_root {
            block_size - BTREE_INFO_SIZE
        } else {
            block_size
        };

        for (i, &(oid, xid, child_paddr)) in entries.iter().enumerate() {
            let k_off = (i * 16) as u16;
            // Internal node values are 8 bytes — offsets must reflect
            // that. Use 8B per slot.
            let v_off = ((i + 1) * 8) as u16;
            block[toc_base + i * 4..toc_base + i * 4 + 2].copy_from_slice(&k_off.to_le_bytes());
            block[toc_base + i * 4 + 2..toc_base + i * 4 + 4].copy_from_slice(&v_off.to_le_bytes());

            let ks = keys_start + k_off as usize;
            block[ks..ks + 8].copy_from_slice(&oid.to_le_bytes());
            block[ks + 8..ks + 16].copy_from_slice(&xid.to_le_bytes());

            let vs = vals_end - v_off as usize;
            block[vs..vs + 8].copy_from_slice(&child_paddr.to_le_bytes());
        }

        if is_root {
            // bt_key_size = 16, bt_val_size = 16 (leaf payload size).
            let info_off = block_size - BTREE_INFO_SIZE;
            block[info_off + 8..info_off + 12].copy_from_slice(&16u32.to_le_bytes());
            block[info_off + 12..info_off + 16].copy_from_slice(&16u32.to_le_bytes());
        }
        block
    }

    /// Single-leaf omap: lookup picks the largest xid ≤ target.
    #[test]
    fn lookup_single_leaf_picks_latest_xid_below_target() {
        let block_size = 512usize;
        // Entries sorted (oid, xid): (5,1)→0x100, (5,3)→0x200, (5,7)→0x300
        let root = build_leaf_omap_node(
            block_size,
            &[(5, 1, 0x100), (5, 3, 0x200), (5, 7, 0x300)],
            true,
        );
        let mut no_read =
            |_p: u64, _b: &mut [u8]| panic!("single-leaf lookup must not read from device");
        let mut cache = NodeCache::new(NodeCache::DEFAULT_CAP);
        let v = lookup_with_cache(&root, 5, 5, &mut no_read, &mut cache).unwrap();
        assert_eq!(v.unwrap().paddr, 0x200);
        let v = lookup_with_cache(&root, 5, 10, &mut no_read, &mut cache).unwrap();
        assert_eq!(v.unwrap().paddr, 0x300);
        let v = lookup_with_cache(&root, 5, 0, &mut no_read, &mut cache).unwrap();
        assert!(v.is_none());
        let v = lookup_with_cache(&root, 99, 5, &mut no_read, &mut cache).unwrap();
        assert!(v.is_none());
    }

    /// Two-level omap: descend through an internal root to a leaf
    /// child. We register the leaf at paddr 10 in a fake device.
    #[test]
    fn lookup_two_level_descends_to_correct_leaf() {
        let block_size = 512usize;
        // Build two leaves:
        //  leaf_a (paddr 10) holds oid=5 entries
        //  leaf_b (paddr 11) holds oid=9 entries
        let leaf_a = build_leaf_omap_node(block_size, &[(5, 1, 0x100), (5, 3, 0x200)], false);
        let leaf_b = build_leaf_omap_node(block_size, &[(9, 2, 0x900)], false);

        // Build an internal root: first entry (5, 0) → child paddr 10,
        // second entry (9, 0) → child paddr 11.
        let root = build_internal_omap_node(block_size, &[(5, 0, 10), (9, 0, 11)], true);

        let mut device: std::collections::HashMap<u64, Vec<u8>> = Default::default();
        device.insert(10, leaf_a);
        device.insert(11, leaf_b);

        let mut read = |paddr: u64, buf: &mut [u8]| -> crate::Result<()> {
            let b = device
                .get(&paddr)
                .ok_or_else(|| crate::Error::InvalidImage(format!("no block at {paddr}")))?;
            buf.copy_from_slice(b);
            Ok(())
        };
        let mut cache = NodeCache::new(NodeCache::DEFAULT_CAP);

        let v = lookup_with_cache(&root, 5, 3, &mut read, &mut cache).unwrap();
        assert_eq!(v.unwrap().paddr, 0x200);

        let v = lookup_with_cache(&root, 9, 10, &mut read, &mut cache).unwrap();
        assert_eq!(v.unwrap().paddr, 0x900);

        // Missing oid: descends to the appropriate leaf, finds nothing.
        let v = lookup_with_cache(&root, 7, 10, &mut read, &mut cache).unwrap();
        assert!(v.is_none());
    }

    #[test]
    fn decode_omap_phys() {
        let mut buf = vec![0u8; 4096];
        buf[24..28].copy_from_slice(&OBJECT_TYPE_OMAP.to_le_bytes());
        buf[32..36].copy_from_slice(&1u32.to_le_bytes()); // flags
        buf[40..44].copy_from_slice(&0xeu32.to_le_bytes()); // tree_type
        buf[48..56].copy_from_slice(&0x1234u64.to_le_bytes()); // tree_oid
        let o = OmapPhys::decode(&buf).unwrap();
        assert_eq!(o.tree_oid, 0x1234);
        assert_eq!(o.flags, 1);
    }

    #[test]
    fn omap_key_val_decode() {
        let mut kb = [0u8; 16];
        kb[0..8].copy_from_slice(&7u64.to_le_bytes());
        kb[8..16].copy_from_slice(&9u64.to_le_bytes());
        let k = OmapKey::decode(&kb).unwrap();
        assert_eq!(k.oid, 7);
        assert_eq!(k.xid, 9);

        let mut vb = [0u8; 16];
        vb[0..4].copy_from_slice(&2u32.to_le_bytes());
        vb[4..8].copy_from_slice(&1u32.to_le_bytes());
        vb[8..16].copy_from_slice(&0x4000u64.to_le_bytes());
        let v = OmapVal::decode(&vb).unwrap();
        assert_eq!(v.flags, 2);
        assert_eq!(v.size, 1);
        assert_eq!(v.paddr, 0x4000);
    }
}