From 8918e379aac11b1e2ed7bb5be8433516c4ff5e8f Mon Sep 17 00:00:00 2001 From: Derrick Stolee Date: Thu, 16 Apr 2020 20:14:02 +0000 Subject: revision: complicated pathspecs disable filters The changed-path Bloom filters work only when we can compute an explicit Bloom filter key in advance. When a pathspec is given that allows case-insensitive checks or wildcard matching, we must disable the Bloom filter performance checks. By checking the pathspec in prepare_to_use_bloom_filters(), we avoid setting up the Bloom filter data and thus revert to the usual logic. Before this change, the following tests would fail*: t6004-rev-list-path-optim.sh (Tests 6-7) t6130-pathspec-noglob.sh (Tests 3-6) t6131-pathspec-icase.sh (Tests 3-5) *These tests would fail when using GIT_TEST_COMMIT_GRAPH and GIT_TEST_COMMIT_GRAPH_BLOOM_FILTERS except that the latter environment variable was not set up correctly to write the changed- path Bloom filters in the test suite. That will be fixed in the next change. Helped-by: Taylor Blau Signed-off-by: Derrick Stolee Signed-off-by: Junio C Hamano diff --git a/revision.c b/revision.c index 2b06ee7..f78c636 100644 --- a/revision.c +++ b/revision.c @@ -650,6 +650,20 @@ static void trace2_bloom_filter_statistics_atexit(void) jw_release(&jw); } +static int forbid_bloom_filters(struct pathspec *spec) +{ + if (spec->has_wildcard) + return 1; + if (spec->nr > 1) + return 1; + if (spec->magic & ~PATHSPEC_LITERAL) + return 1; + if (spec->nr && (spec->items[0].magic & ~PATHSPEC_LITERAL)) + return 1; + + return 0; +} + static void prepare_to_use_bloom_filter(struct rev_info *revs) { struct pathspec_item *pi; @@ -659,7 +673,10 @@ static void prepare_to_use_bloom_filter(struct rev_info *revs) int len; if (!revs->commits) - return; + return; + + if (forbid_bloom_filters(&revs->prune_data)) + return; repo_parse_commit(revs->repo, revs->commits->item); -- cgit v0.10.2-6-g49f6 From b23ea9790d30e80c7c79a77aab2e9d150a189463 Mon Sep 17 00:00:00 2001 From: Derrick Stolee Date: Thu, 16 Apr 2020 20:14:03 +0000 Subject: tests: write commit-graph with Bloom filters The GIT_TEST_COMMIT_GRAPH environment variable updates the commit- graph file whenever "git commit" is run, ensuring that we always have an updated commit-graph throughout the test suite. The GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS environment variable was introduced to write the changed-path Bloom filters whenever "git commit-graph write" is run. However, the GIT_TEST_COMMIT_GRAPH trick doesn't launch a separate process and instead writes it directly. To expand the number of tests that have commits in the commit-graph file, add a helper method that computes the commit-graph and place that helper inside "git commit" and "git merge". In the helper method, check GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS to ensure we are writing changed-path Bloom filters whenever possible. Signed-off-by: Derrick Stolee Signed-off-by: Junio C Hamano diff --git a/builtin/commit.c b/builtin/commit.c index d3e7781..27d4ff6 100644 --- a/builtin/commit.c +++ b/builtin/commit.c @@ -1700,9 +1700,7 @@ int cmd_commit(int argc, const char **argv, const char *prefix) "new_index file. Check that disk is not full and quota is\n" "not exceeded, and then \"git restore --staged :/\" to recover.")); - if (git_env_bool(GIT_TEST_COMMIT_GRAPH, 0) && - write_commit_graph_reachable(the_repository->objects->odb, 0, NULL)) - return 1; + git_test_write_commit_graph_or_die(); repo_rerere(the_repository, 0); run_command_v_opt(argv_gc_auto, RUN_GIT_CMD); diff --git a/builtin/merge.c b/builtin/merge.c index d127d22..db11af5 100644 --- a/builtin/merge.c +++ b/builtin/merge.c @@ -40,6 +40,7 @@ #include "branch.h" #include "commit-reach.h" #include "wt-status.h" +#include "commit-graph.h" #define DEFAULT_TWOHEAD (1<<0) #define DEFAULT_OCTOPUS (1<<1) @@ -1673,9 +1674,11 @@ int cmd_merge(int argc, const char **argv, const char *prefix) head_commit); } - if (squash) + if (squash) { finish(head_commit, remoteheads, NULL, NULL); - else + + git_test_write_commit_graph_or_die(); + } else write_merge_state(remoteheads); if (merge_was_ok) diff --git a/commit-graph.c b/commit-graph.c index 7766862..196e817 100644 --- a/commit-graph.c +++ b/commit-graph.c @@ -19,6 +19,20 @@ #include "bloom.h" #include "commit-slab.h" +void git_test_write_commit_graph_or_die(void) +{ + int flags = 0; + if (!git_env_bool(GIT_TEST_COMMIT_GRAPH, 0)) + return; + + if (git_env_bool(GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS, 0)) + flags = COMMIT_GRAPH_WRITE_BLOOM_FILTERS; + + if (write_commit_graph_reachable(the_repository->objects->odb, + flags, NULL)) + die("failed to write commit-graph under GIT_TEST_COMMIT_GRAPH"); +} + #define GRAPH_SIGNATURE 0x43475048 /* "CGPH" */ #define GRAPH_CHUNKID_OIDFANOUT 0x4f494446 /* "OIDF" */ #define GRAPH_CHUNKID_OIDLOOKUP 0x4f49444c /* "OIDL" */ diff --git a/commit-graph.h b/commit-graph.h index 8655d06..3948448 100644 --- a/commit-graph.h +++ b/commit-graph.h @@ -11,6 +11,15 @@ #define GIT_TEST_COMMIT_GRAPH_DIE_ON_LOAD "GIT_TEST_COMMIT_GRAPH_DIE_ON_LOAD" #define GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS "GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS" +/* + * This method is only used to enhance coverage of the commit-graph + * feature in the test suite with the GIT_TEST_COMMIT_GRAPH and + * GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS environment variables. Do not + * call this method oustide of a builtin, and only if you know what + * you are doing! + */ +void git_test_write_commit_graph_or_die(void); + struct commit; struct bloom_filter_settings; -- cgit v0.10.2-6-g49f6 From 0906ac2b54b69eb98c448ad0072be1a1c355fbc4 Mon Sep 17 00:00:00 2001 From: Derrick Stolee Date: Thu, 16 Apr 2020 20:14:04 +0000 Subject: blame: use changed-path Bloom filters The changed-path Bloom filters help reduce the amount of tree parsing required during history queries. Before calculating a diff, we can ask the filter if a path changed between a commit and its first parent. If the filter says "no" then we can move on without parsing trees. If the filter says "maybe" then we parse trees to discover if the answer is actually "yes" or "no". When computing a blame, there is a section in find_origin() that computes a diff between a commit and one of its parents. When this is the first parent, we can check the Bloom filters before calling diff_tree_oid(). In order to make this work with the blame machinery, we need to initialize a struct bloom_key with the initial path. But also, we need to add more keys to a list if a rename is detected. We then check to see if _any_ of these keys answer "maybe" in the diff. During development, I purposefully left out this "add a new key when a rename is detected" to see if the test suite would catch my error. That is how I discovered the issues with GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS from the previous change. With that change, we can feel some confidence in the coverage of this change. If a user requests copy detection using "git blame -C", then there are more places where the set of "important" files can expand. I do not know enough about how this happens in the blame machinery. Thus, the Bloom filter integration is explicitly disabled in this mode. A later change could expand the bloom_key data with an appropriate call (or calls) to add_bloom_key(). If we did not disable this mode, then the following tests would fail: t8003-blame-corner-cases.sh t8011-blame-split-file.sh Generally, this is a performance enhancement and should not change the behavior of 'git blame' in any way. If a repo has a commit-graph file with computed changed-path Bloom filters, then they should notice improved performance for their 'git blame' commands. Here are some example timings that I found by blaming some paths in the Linux kernel repository: git blame arch/x86/kernel/topology.c >/dev/null Before: 0.83s After: 0.24s git blame kernel/time/time.c >/dev/null Before: 0.72s After: 0.24s git blame tools/perf/ui/stdio/hist.c >/dev/null Before: 0.27s After: 0.11s I specifically looked for "deep" paths that were also edited many times. As a counterpoint, the MAINTAINERS file was edited many times but is located in the root tree. This means that the cost of computing a diff relative to the pathspec is very small. Here are the timings for that command: git blame MAINTAINERS >/dev/null Before: 20.1s After: 18.0s These timings are the best of five. The worst-case runs were on the order of 2.5 minutes for both cases. Note that the MAINTAINERS file has 18,740 lines across 17,000+ commits. This happens to be one of the cases where this change provides the least improvement. The lack of improvement for the MAINTAINERS file and the relatively modest improvement for the other examples can be easily explained. The blame machinery needs to compute line-level diffs to determine which lines were changed by each commit. That makes up a large proportion of the computation time, and this change does not attempt to improve on that section of the algorithm. The MAINTAINERS file is large and changed often, so it takes time to determine which lines were updated by which commit. In contrast, the code files are much smaller, and it takes longer to comute the line-by-line diff for a single patch on the Linux mailing lists. Outside of the "-C" integration, I believe there is little more to gain from the changed-path Bloom filters for 'git blame' after this patch. Signed-off-by: Derrick Stolee Signed-off-by: Junio C Hamano diff --git a/blame.c b/blame.c index 29770e5..9fbf79e 100644 --- a/blame.c +++ b/blame.c @@ -9,6 +9,8 @@ #include "blame.h" #include "alloc.h" #include "commit-slab.h" +#include "bloom.h" +#include "commit-graph.h" define_commit_slab(blame_suspects, struct blame_origin *); static struct blame_suspects blame_suspects; @@ -1246,13 +1248,75 @@ static int fill_blob_sha1_and_mode(struct repository *r, return -1; } +struct blame_bloom_data { + /* + * Changed-path Bloom filter keys. These can help prevent + * computing diffs against first parents, but we need to + * expand the list as code is moved or files are renamed. + */ + struct bloom_filter_settings *settings; + struct bloom_key **keys; + int nr; + int alloc; +}; + +static int bloom_count_queries = 0; +static int bloom_count_no = 0; +static int maybe_changed_path(struct repository *r, + struct commit *parent, + struct blame_origin *origin, + struct blame_bloom_data *bd) +{ + int i; + struct bloom_filter *filter; + + if (!bd) + return 1; + + if (origin->commit->generation == GENERATION_NUMBER_INFINITY) + return 1; + + filter = get_bloom_filter(r, origin->commit, 0); + + if (!filter) + return 1; + + bloom_count_queries++; + for (i = 0; i < bd->nr; i++) { + if (bloom_filter_contains(filter, + bd->keys[i], + bd->settings)) + return 1; + } + + bloom_count_no++; + return 0; +} + +static void add_bloom_key(struct blame_bloom_data *bd, + const char *path) +{ + if (!bd) + return; + + if (bd->nr >= bd->alloc) { + bd->alloc *= 2; + REALLOC_ARRAY(bd->keys, bd->alloc); + } + + bd->keys[bd->nr] = xmalloc(sizeof(struct bloom_key)); + fill_bloom_key(path, strlen(path), bd->keys[bd->nr], bd->settings); + bd->nr++; +} + /* * We have an origin -- check if the same path exists in the * parent and return an origin structure to represent it. */ static struct blame_origin *find_origin(struct repository *r, struct commit *parent, - struct blame_origin *origin) + struct blame_origin *origin, + struct blame_bloom_data *bd) { struct blame_origin *porigin; struct diff_options diff_opts; @@ -1286,10 +1350,19 @@ static struct blame_origin *find_origin(struct repository *r, if (is_null_oid(&origin->commit->object.oid)) do_diff_cache(get_commit_tree_oid(parent), &diff_opts); - else - diff_tree_oid(get_commit_tree_oid(parent), - get_commit_tree_oid(origin->commit), - "", &diff_opts); + else { + int compute_diff = 1; + if (origin->commit->parents && + !oidcmp(&parent->object.oid, + &origin->commit->parents->item->object.oid)) + compute_diff = maybe_changed_path(r, parent, + origin, bd); + + if (compute_diff) + diff_tree_oid(get_commit_tree_oid(parent), + get_commit_tree_oid(origin->commit), + "", &diff_opts); + } diffcore_std(&diff_opts); if (!diff_queued_diff.nr) { @@ -1341,7 +1414,8 @@ static struct blame_origin *find_origin(struct repository *r, */ static struct blame_origin *find_rename(struct repository *r, struct commit *parent, - struct blame_origin *origin) + struct blame_origin *origin, + struct blame_bloom_data *bd) { struct blame_origin *porigin = NULL; struct diff_options diff_opts; @@ -1366,6 +1440,7 @@ static struct blame_origin *find_rename(struct repository *r, struct diff_filepair *p = diff_queued_diff.queue[i]; if ((p->status == 'R' || p->status == 'C') && !strcmp(p->two->path, origin->path)) { + add_bloom_key(bd, p->one->path); porigin = get_origin(parent, p->one->path); oidcpy(&porigin->blob_oid, &p->one->oid); porigin->mode = p->one->mode; @@ -2332,6 +2407,11 @@ static void distribute_blame(struct blame_scoreboard *sb, struct blame_entry *bl #define MAXSG 16 +typedef struct blame_origin *(*blame_find_alg)(struct repository *, + struct commit *, + struct blame_origin *, + struct blame_bloom_data *); + static void pass_blame(struct blame_scoreboard *sb, struct blame_origin *origin, int opt) { struct rev_info *revs = sb->revs; @@ -2356,8 +2436,7 @@ static void pass_blame(struct blame_scoreboard *sb, struct blame_origin *origin, * common cases, then we look for renames in the second pass. */ for (pass = 0; pass < 2 - sb->no_whole_file_rename; pass++) { - struct blame_origin *(*find)(struct repository *, struct commit *, struct blame_origin *); - find = pass ? find_rename : find_origin; + blame_find_alg find = pass ? find_rename : find_origin; for (i = 0, sg = first_scapegoat(revs, commit, sb->reverse); i < num_sg && sg; @@ -2369,7 +2448,7 @@ static void pass_blame(struct blame_scoreboard *sb, struct blame_origin *origin, continue; if (parse_commit(p)) continue; - porigin = find(sb->repo, p, origin); + porigin = find(sb->repo, p, origin, sb->bloom_data); if (!porigin) continue; if (oideq(&porigin->blob_oid, &origin->blob_oid)) { @@ -2809,3 +2888,45 @@ struct blame_entry *blame_entry_prepend(struct blame_entry *head, blame_origin_incref(o); return new_head; } + +void setup_blame_bloom_data(struct blame_scoreboard *sb, + const char *path) +{ + struct blame_bloom_data *bd; + + if (!sb->repo->objects->commit_graph) + return; + + if (!sb->repo->objects->commit_graph->bloom_filter_settings) + return; + + bd = xmalloc(sizeof(struct blame_bloom_data)); + + bd->settings = sb->repo->objects->commit_graph->bloom_filter_settings; + + bd->alloc = 4; + bd->nr = 0; + ALLOC_ARRAY(bd->keys, bd->alloc); + + add_bloom_key(bd, path); + + sb->bloom_data = bd; +} + +void cleanup_scoreboard(struct blame_scoreboard *sb) +{ + if (sb->bloom_data) { + int i; + for (i = 0; i < sb->bloom_data->nr; i++) { + free(sb->bloom_data->keys[i]->hashes); + free(sb->bloom_data->keys[i]); + } + free(sb->bloom_data->keys); + FREE_AND_NULL(sb->bloom_data); + + trace2_data_intmax("blame", sb->repo, + "bloom/queries", bloom_count_queries); + trace2_data_intmax("blame", sb->repo, + "bloom/response-no", bloom_count_no); + } +} diff --git a/blame.h b/blame.h index 089b181..b6bbee4 100644 --- a/blame.h +++ b/blame.h @@ -100,6 +100,8 @@ struct blame_entry { int unblamable; }; +struct blame_bloom_data; + /* * The current state of the blame assignment. */ @@ -156,6 +158,7 @@ struct blame_scoreboard { void(*found_guilty_entry)(struct blame_entry *, void *); void *found_guilty_entry_data; + struct blame_bloom_data *bloom_data; }; /* @@ -180,6 +183,9 @@ void init_scoreboard(struct blame_scoreboard *sb); void setup_scoreboard(struct blame_scoreboard *sb, const char *path, struct blame_origin **orig); +void setup_blame_bloom_data(struct blame_scoreboard *sb, + const char *path); +void cleanup_scoreboard(struct blame_scoreboard *sb); struct blame_entry *blame_entry_prepend(struct blame_entry *head, long start, long end, diff --git a/builtin/blame.c b/builtin/blame.c index bf1cecd..3c13634 100644 --- a/builtin/blame.c +++ b/builtin/blame.c @@ -1061,6 +1061,14 @@ parse_done: string_list_clear(&ignore_revs_file_list, 0); string_list_clear(&ignore_rev_list, 0); setup_scoreboard(&sb, path, &o); + + /* + * Changed-path Bloom filters are disabled when looking + * for copies. + */ + if (!(opt & PICKAXE_BLAME_COPY)) + setup_blame_bloom_data(&sb, path); + lno = sb.num_lines; if (lno && !range_list.nr) @@ -1164,5 +1172,7 @@ parse_done: printf("num get patch: %d\n", sb.num_get_patch); printf("num commits: %d\n", sb.num_commits); } + + cleanup_scoreboard(&sb); return 0; } -- cgit v0.10.2-6-g49f6 From fe88f9f91ffcb7a4dbb83b17e0b62c12409a6047 Mon Sep 17 00:00:00 2001 From: Jeff King Date: Thu, 23 Apr 2020 17:03:03 -0400 Subject: blame: drop unused parameter from maybe_changed_path We don't use the "parent" parameter at all (probably because the bloom filter for a commit is always defined against a single parent anyway). Signed-off-by: Jeff King Signed-off-by: Junio C Hamano diff --git a/blame.c b/blame.c index 9fbf79e..da7e288 100644 --- a/blame.c +++ b/blame.c @@ -1263,7 +1263,6 @@ struct blame_bloom_data { static int bloom_count_queries = 0; static int bloom_count_no = 0; static int maybe_changed_path(struct repository *r, - struct commit *parent, struct blame_origin *origin, struct blame_bloom_data *bd) { @@ -1355,8 +1354,7 @@ static struct blame_origin *find_origin(struct repository *r, if (origin->commit->parents && !oidcmp(&parent->object.oid, &origin->commit->parents->item->object.oid)) - compute_diff = maybe_changed_path(r, parent, - origin, bd); + compute_diff = maybe_changed_path(r, origin, bd); if (compute_diff) diff_tree_oid(get_commit_tree_oid(parent), -- cgit v0.10.2-6-g49f6 From 24b7d1e7b07aa1329d0d75f7922af8ef3e0782d2 Mon Sep 17 00:00:00 2001 From: Jeff King Date: Thu, 23 Apr 2020 16:59:07 -0400 Subject: test-bloom: fix some whitespace issues Signed-off-by: Jeff King Reviewed-by: Taylor Blau Signed-off-by: Junio C Hamano diff --git a/t/helper/test-bloom.c b/t/helper/test-bloom.c index ce41266..f9c0ce2 100644 --- a/t/helper/test-bloom.c +++ b/t/helper/test-bloom.c @@ -27,7 +27,7 @@ static void print_bloom_filter(struct bloom_filter *filter) { } printf("Filter_Length:%d\n", (int)filter->len); printf("Filter_Data:"); - for (i = 0; i < filter->len; i++){ + for (i = 0; i < filter->len; i++) { printf("%02x|", filter->data[i]); } printf("\n"); @@ -50,13 +50,13 @@ int cmd__bloom(int argc, const char **argv) printf("Murmur3 Hash with seed=0:0x%08x\n", hashed); } - if (!strcmp(argv[1], "generate_filter")) { + if (!strcmp(argv[1], "generate_filter")) { struct bloom_filter filter; int i = 2; filter.len = (settings.bits_per_entry + BITS_PER_WORD - 1) / BITS_PER_WORD; filter.data = xcalloc(filter.len, sizeof(unsigned char)); - if (!argv[2]){ + if (!argv[2]) { die("at least one input string expected"); } @@ -68,7 +68,7 @@ int cmd__bloom(int argc, const char **argv) print_bloom_filter(&filter); } - if (!strcmp(argv[1], "get_filter_for_commit")) { + if (!strcmp(argv[1], "get_filter_for_commit")) { struct object_id oid; const char *end; if (parse_oid_hex(argv[2], &oid, &end)) @@ -78,4 +78,4 @@ int cmd__bloom(int argc, const char **argv) } return 0; -} \ No newline at end of file +} -- cgit v0.10.2-6-g49f6 From 1b4c57fa87e121f155863f898dc39d06cf4a1d99 Mon Sep 17 00:00:00 2001 From: Jeff King Date: Thu, 23 Apr 2020 16:59:14 -0400 Subject: test-bloom: check that we have expected arguments If "test-tool bloom" is not fed a command, or if arguments are missing for some commands, it will just segfault. Let's check argc and write a friendlier usage message. Signed-off-by: Jeff King Reviewed-by: Taylor Blau Signed-off-by: Junio C Hamano diff --git a/t/helper/test-bloom.c b/t/helper/test-bloom.c index f9c0ce2..77eb27a 100644 --- a/t/helper/test-bloom.c +++ b/t/helper/test-bloom.c @@ -43,10 +43,21 @@ static void get_bloom_filter_for_commit(const struct object_id *commit_oid) print_bloom_filter(filter); } +static const char *bloom_usage = "\n" +" test-tool bloom get_murmer3 \n" +" test-tool bloom generate_filter [...]\n" +" test-tool get_filter_for_commit \n"; + int cmd__bloom(int argc, const char **argv) { + if (argc < 2) + usage(bloom_usage); + if (!strcmp(argv[1], "get_murmur3")) { - uint32_t hashed = murmur3_seeded(0, argv[2], strlen(argv[2])); + uint32_t hashed; + if (argc < 3) + usage(bloom_usage); + hashed = murmur3_seeded(0, argv[2], strlen(argv[2])); printf("Murmur3 Hash with seed=0:0x%08x\n", hashed); } @@ -56,9 +67,8 @@ int cmd__bloom(int argc, const char **argv) filter.len = (settings.bits_per_entry + BITS_PER_WORD - 1) / BITS_PER_WORD; filter.data = xcalloc(filter.len, sizeof(unsigned char)); - if (!argv[2]) { - die("at least one input string expected"); - } + if (argc - 1 < i) + usage(bloom_usage); while (argv[i]) { add_string_to_filter(argv[i], &filter); @@ -71,6 +81,8 @@ int cmd__bloom(int argc, const char **argv) if (!strcmp(argv[1], "get_filter_for_commit")) { struct object_id oid; const char *end; + if (argc < 3) + usage(bloom_usage); if (parse_oid_hex(argv[2], &oid, &end)) die("cannot parse oid '%s'", argv[2]); init_bloom_filters(); -- cgit v0.10.2-6-g49f6