commit-graph: use chunk-format read API
Instead of parsing the table of contents directly, use the chunk-format API methods read_table_of_contents() and pair_chunk(). While the current implementation loses the duplicate-chunk detection, that will be added in a future change. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Derrick Stolee committed
Feb 18, 2021 at 14:07 UTC
2692c2f6fd1cc9f74e433cb05d5756575682b9ab
2 files changed
+55
-106
commit-graph.c
+54
-105
@@ -59,8 +59,7 @@ void git_test_write_commit_graph_or_die(void)
59
60
#define GRAPH_HEADER_SIZE 8
61
#define GRAPH_FANOUT_SIZE (4 * 256)
62
-#define GRAPH_CHUNKLOOKUP_WIDTH 12
63
-#define GRAPH_MIN_SIZE (GRAPH_HEADER_SIZE + 4 * GRAPH_CHUNKLOOKUP_WIDTH \
62
+#define GRAPH_MIN_SIZE (GRAPH_HEADER_SIZE + 4 * CHUNK_TOC_ENTRY_SIZE \
63
+ GRAPH_FANOUT_SIZE + the_hash_algo->rawsz)
64
65
#define CORRECTED_COMMIT_DATE_OFFSET_OVERFLOW (1ULL << 31)
@@ -298,15 +297,43 @@ static int verify_commit_graph_lite(struct commit_graph *g)
297
return 0;
298
}
299
300
+static int graph_read_oid_lookup(const unsigned char *chunk_start,
301
+ size_t chunk_size, void *data)
302
+{
303
+ struct commit_graph *g = data;
304
+ g->chunk_oid_lookup = chunk_start;
305
+ g->num_commits = chunk_size / g->hash_len;
306
+ return 0;
307
+}
308
+
309
+static int graph_read_bloom_data(const unsigned char *chunk_start,
310
+ size_t chunk_size, void *data)
311
+{
312
+ struct commit_graph *g = data;
313
+ uint32_t hash_version;
314
+ g->chunk_bloom_data = chunk_start;
315
+ hash_version = get_be32(chunk_start);
316
+
317
+ if (hash_version != 1)
318
+ return 0;
319
+
320
+ g->bloom_filter_settings = xmalloc(sizeof(struct bloom_filter_settings));
321
+ g->bloom_filter_settings->hash_version = hash_version;
322
+ g->bloom_filter_settings->num_hashes = get_be32(chunk_start + 4);
323
+ g->bloom_filter_settings->bits_per_entry = get_be32(chunk_start + 8);
324
+ g->bloom_filter_settings->max_changed_paths = DEFAULT_BLOOM_MAX_CHANGES;
325
+
326
+ return 0;
327
+}
328
+
329
struct commit_graph *parse_commit_graph(struct repository *r,
330
void *graph_map, size_t graph_size)
331
{
304
- const unsigned char *data, *chunk_lookup;
305
- uint32_t i;
332
+ const unsigned char *data;
333
struct commit_graph *graph;
307
- uint64_t next_chunk_offset;
334
uint32_t graph_signature;
335
unsigned char graph_version, hash_version;
336
+ struct chunkfile *cf = NULL;
337
338
if (!graph_map)
339
return NULL;
@@ -347,7 +374,7 @@ struct commit_graph *parse_commit_graph(struct repository *r,
374
graph->data_len = graph_size;
375
376
if (graph_size < GRAPH_HEADER_SIZE +
350
- (graph->num_chunks + 1) * GRAPH_CHUNKLOOKUP_WIDTH +
377
+ (graph->num_chunks + 1) * CHUNK_TOC_ENTRY_SIZE +
378
GRAPH_FANOUT_SIZE + the_hash_algo->rawsz) {
379
error(_("commit-graph file is too small to hold %u chunks"),
380
graph->num_chunks);
@@ -355,108 +382,28 @@ struct commit_graph *parse_commit_graph(struct repository *r,
382
return NULL;
383
}
384
358
- chunk_lookup = data + 8;
359
- next_chunk_offset = get_be64(chunk_lookup + 4);
360
- for (i = 0; i < graph->num_chunks; i++) {
361
- uint32_t chunk_id;
362
- uint64_t chunk_offset = next_chunk_offset;
363
- int chunk_repeated = 0;
364
-
365
- chunk_id = get_be32(chunk_lookup + 0);
366
-
367
- chunk_lookup += GRAPH_CHUNKLOOKUP_WIDTH;
368
- next_chunk_offset = get_be64(chunk_lookup + 4);
369
-
370
- if (chunk_offset > graph_size - the_hash_algo->rawsz) {
371
- error(_("commit-graph improper chunk offset %08x%08x"), (uint32_t)(chunk_offset >> 32),
372
- (uint32_t)chunk_offset);
373
- goto free_and_return;
374
- }
375
-
376
- switch (chunk_id) {
377
- case GRAPH_CHUNKID_OIDFANOUT:
378
- if (graph->chunk_oid_fanout)
379
- chunk_repeated = 1;
380
- else
381
- graph->chunk_oid_fanout = (uint32_t*)(data + chunk_offset);
382
- break;
383
-
384
- case GRAPH_CHUNKID_OIDLOOKUP:
385
- if (graph->chunk_oid_lookup)
386
- chunk_repeated = 1;
387
- else {
388
- graph->chunk_oid_lookup = data + chunk_offset;
389
- graph->num_commits = (next_chunk_offset - chunk_offset)
390
- / graph->hash_len;
391
- }
392
- break;
385
+ cf = init_chunkfile(NULL);
386
394
- case GRAPH_CHUNKID_DATA:
395
- if (graph->chunk_commit_data)
396
- chunk_repeated = 1;
397
- else
398
- graph->chunk_commit_data = data + chunk_offset;
399
- break;
400
-
401
- case GRAPH_CHUNKID_GENERATION_DATA:
402
- if (graph->chunk_generation_data)
403
- chunk_repeated = 1;
404
- else
405
- graph->chunk_generation_data = data + chunk_offset;
406
- break;
407
-
408
- case GRAPH_CHUNKID_GENERATION_DATA_OVERFLOW:
409
- if (graph->chunk_generation_data_overflow)
410
- chunk_repeated = 1;
411
- else
412
- graph->chunk_generation_data_overflow = data + chunk_offset;
413
- break;
414
-
415
- case GRAPH_CHUNKID_EXTRAEDGES:
416
- if (graph->chunk_extra_edges)
417
- chunk_repeated = 1;
418
- else
419
- graph->chunk_extra_edges = data + chunk_offset;
420
- break;
421
-
422
- case GRAPH_CHUNKID_BASE:
423
- if (graph->chunk_base_graphs)
424
- chunk_repeated = 1;
425
- else
426
- graph->chunk_base_graphs = data + chunk_offset;
427
- break;
428
-
429
- case GRAPH_CHUNKID_BLOOMINDEXES:
430
- if (graph->chunk_bloom_indexes)
431
- chunk_repeated = 1;
432
- else if (r->settings.commit_graph_read_changed_paths)
433
- graph->chunk_bloom_indexes = data + chunk_offset;
434
- break;
435
-
436
- case GRAPH_CHUNKID_BLOOMDATA:
437
- if (graph->chunk_bloom_data)
438
- chunk_repeated = 1;
439
- else if (r->settings.commit_graph_read_changed_paths) {
440
- uint32_t hash_version;
441
- graph->chunk_bloom_data = data + chunk_offset;
442
- hash_version = get_be32(data + chunk_offset);
443
-
444
- if (hash_version != 1)
445
- break;
387
+ if (read_table_of_contents(cf, graph->data, graph_size,
388
+ GRAPH_HEADER_SIZE, graph->num_chunks))
389
+ goto free_and_return;
390
447
- graph->bloom_filter_settings = xmalloc(sizeof(struct bloom_filter_settings));
448
- graph->bloom_filter_settings->hash_version = hash_version;
449
- graph->bloom_filter_settings->num_hashes = get_be32(data + chunk_offset + 4);
450
- graph->bloom_filter_settings->bits_per_entry = get_be32(data + chunk_offset + 8);
451
- graph->bloom_filter_settings->max_changed_paths = DEFAULT_BLOOM_MAX_CHANGES;
452
- }
453
- break;
454
- }
391
+ pair_chunk(cf, GRAPH_CHUNKID_OIDFANOUT,
392
+ (const unsigned char **)&graph->chunk_oid_fanout);
393
+ read_chunk(cf, GRAPH_CHUNKID_OIDLOOKUP, graph_read_oid_lookup, graph);
394
+ pair_chunk(cf, GRAPH_CHUNKID_DATA, &graph->chunk_commit_data);
395
+ pair_chunk(cf, GRAPH_CHUNKID_EXTRAEDGES, &graph->chunk_extra_edges);
396
+ pair_chunk(cf, GRAPH_CHUNKID_BASE, &graph->chunk_base_graphs);
397
+ pair_chunk(cf, GRAPH_CHUNKID_GENERATION_DATA,
398
+ &graph->chunk_generation_data);
399
+ pair_chunk(cf, GRAPH_CHUNKID_GENERATION_DATA_OVERFLOW,
400
+ &graph->chunk_generation_data_overflow);
401
456
- if (chunk_repeated) {
457
- error(_("commit-graph chunk id %08x appears multiple times"), chunk_id);
458
- goto free_and_return;
459
- }
402
+ if (r->settings.commit_graph_read_changed_paths) {
403
+ pair_chunk(cf, GRAPH_CHUNKID_BLOOMINDEXES,
404
+ &graph->chunk_bloom_indexes);
405
+ read_chunk(cf, GRAPH_CHUNKID_BLOOMDATA,
406
+ graph_read_bloom_data, graph);
407
}
408
409
if (graph->chunk_bloom_indexes && graph->chunk_bloom_data) {
@@ -473,9 +420,11 @@ struct commit_graph *parse_commit_graph(struct repository *r,
420
if (verify_commit_graph_lite(graph))
421
goto free_and_return;
422
423
+ free_chunkfile(cf);
424
return graph;
425
426
free_and_return:
427
+ free_chunkfile(cf);
428
free(graph->bloom_filter_settings);
429
free(graph);
430
return NULL;
t/t5318-commit-graph.sh
+1
-1
@@ -564,7 +564,7 @@ test_expect_success 'detect bad hash version' '
564
565
test_expect_success 'detect low chunk count' '
566
corrupt_graph_and_verify $GRAPH_BYTE_CHUNK_COUNT "\01" \
567
- "missing the .* chunk"
567
+ "final chunk has non-zero id"
568
'
569
570
test_expect_success 'detect missing OID fanout chunk' '