From 394b84bb70b09ca6c9eebfc850c0023edae64ff8 Mon Sep 17 00:00:00 2001 From: Matthew John Cheetham Date: Sat, 27 Dec 2025 15:12:07 -0800 Subject: [PATCH] fsck: snapshot default refs before object walk Fsck has a race when operating on live repositories; consider the following simple script that writes new commits as fsck runs: #!/bin/bash git fsck & PID=$! while ps -p $PID >/dev/null; do sleep 3 git commit -q --allow-empty -m "Another commit" done Since fsck reads refs at the beginning, walks those for connectivity, and then reads the refs again at the end to check, this can cause fsck to get confused and think that the new refs refer to missing commits and that new reflog entries are invalid. Running the above script in a clone of git.git results in the following (output ellipsized to remove additional errors of the same type): $ ./fsck-while-writing.sh Checking ref database: 100% (1/1), done. Checking object directories: 100% (256/256), done. warning in tag d6602ec5194c87b0fc87103ca4d67251c76f233a: missingTaggerEntry: invalid format - expected 'tagger' line Checking objects: 100% (835091/835091), done. error: HEAD: invalid reflog entry 2aac9f9286e2164fbf8e4f1d1df53044ace2b310 error: HEAD: invalid reflog entry 2aac9f9286e2164fbf8e4f1d1df53044ace2b310 error: HEAD: invalid reflog entry da0f5b80d61844a6f0ad2ddfd57e4fdfa246ea68 error: HEAD: invalid reflog entry da0f5b80d61844a6f0ad2ddfd57e4fdfa246ea68 [...] error: HEAD: invalid reflog entry 87c8a5c2f6b79d9afa9e941590b9a097b6f7ac09 error: HEAD: invalid reflog entry d80887a48865e6ad165274b152cbbbed29f8a55a error: HEAD: invalid reflog entry d80887a48865e6ad165274b152cbbbed29f8a55a error: HEAD: invalid reflog entry 6724f2dfede88bfa9445a333e06e78536c0c6c0d error: refs/heads/mybranch invalid reflog entry 2aac9f9286e2164fbf8e4f1d1df53044ace2b310 error: refs/heads/mybranch: invalid reflog entry 2aac9f9286e2164fbf8e4f1d1df53044ace2b310 error: refs/heads/mybranch: invalid reflog entry da0f5b80d61844a6f0ad2ddfd57e4fdfa246ea68 error: refs/heads/mybranch: invalid reflog entry da0f5b80d61844a6f0ad2ddfd57e4fdfa246ea68 [...] error: refs/heads/mybranch: invalid reflog entry 87c8a5c2f6b79d9afa9e941590b9a097b6f7ac09 error: refs/heads/mybranch: invalid reflog entry d80887a48865e6ad165274b152cbbbed29f8a55a error: refs/heads/mybranch: invalid reflog entry d80887a48865e6ad165274b152cbbbed29f8a55a error: refs/heads/mybranch: invalid reflog entry 6724f2dfede88bfa9445a333e06e78536c0c6c0d Checking connectivity: 833846, done. missing commit 6724f2dfede88bfa9445a333e06e78536c0c6c0d Verifying commits in commit graph: 100% (242243/242243), done. This problem doesn't occur when refs are specified on the command line for us to check, since we use those specified refs for both walking and checking. Using the same refs for walking and checking seems to just make sense, so modify the existing code to do the same when refs aren't specified. Snapshot the refs at the beginning, and also ignore all reflog entries since the time of our snapshot (while this technically means we could ignore a reflog entry created before the fsck process if the local clock is weird, since reflogs are local-only there are not concerns about differences between clocks on different machines). This combination of changes modifies the output of running the above script to: $ ./fsck-while-writing.sh Checking ref database: 100% (1/1), done. Checking object directories: 100% (256/256), done. warning in tag d6602ec5194c87b0fc87103ca4d67251c76f233a: missingTaggerEntry: invalid format - expected 'tagger' line Checking objects: 100% (835091/835091), done. Checking connectivity: 833846, done. Verifying commits in commit graph: 100% (242243/242243), done. While worries about live updates while running fsck is likely of most interest for forge operators, it will likely also benefit those with automated jobs (such as git maintenance) or even casual users who want to do other work in their clone while fsck is running. Signed-off-by: Matthew John Cheetham Co-authored-by: Elijah Newren [en: several changes: * adjusted for upstream refactorings to refs callback call signatures * handle reflogs as well * free recorded snapshot of refs when done * default to snapshotting instead of making it a non-default option * provide reproducible testcase in commit message and rewrite commit message around it ] Signed-off-by: Elijah Newren --- builtin/fsck.c | 74 +++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 70 insertions(+), 4 deletions(-) diff --git a/builtin/fsck.c b/builtin/fsck.c index c489582faa6650..8d20505f5d4e1c 100644 --- a/builtin/fsck.c +++ b/builtin/fsck.c @@ -51,6 +51,7 @@ static int show_progress = -1; static int show_dangling = 1; static int name_objects; static int check_references = 1; +static timestamp_t now; #define ERROR_OBJECT 01 #define ERROR_REACHABLE 02 #define ERROR_PACK 04 @@ -509,6 +510,9 @@ static int fsck_handle_reflog_ent(const char *refname, timestamp_t timestamp, int tz UNUSED, const char *message UNUSED, void *cb_data UNUSED) { + if (now && timestamp > now) + return 0; + if (verbose) fprintf_ln(stderr, _("Checking reflog %s->%s"), oid_to_hex(ooid), oid_to_hex(noid)); @@ -567,14 +571,53 @@ static int fsck_head_link(const char *head_ref_name, const char **head_points_at, struct object_id *head_oid); -static void get_default_heads(void) +struct ref_snapshot { + size_t nr; + size_t name_alloc; + size_t oid_alloc; + char **refname; + struct object_id *oid; +}; + +static int snapshot_refs(const struct reference *ref, void *cb_data) +{ + struct ref_snapshot *refs = cb_data; + + ALLOC_GROW(refs->refname, refs->nr + 1, refs->name_alloc); + ALLOC_GROW(refs->oid, refs->nr + 1, refs->oid_alloc); + + refs->refname[refs->nr] = xstrdup(ref->name); + oidcpy(&refs->oid[refs->nr], ref->oid); + refs->nr++; + + return 0; +} + +static void free_snapshot_refs(struct ref_snapshot *snapshot) +{ + for (size_t i = 0; i < snapshot->nr; i++) + free(snapshot->refname[i]); + free(snapshot->refname); + free(snapshot->oid); +} + +static void get_default_heads(struct ref_snapshot *the_refs) { struct worktree **worktrees, **p; const char *head_points_at; struct object_id head_oid; - refs_for_each_rawref(get_main_ref_store(the_repository), - fsck_handle_ref, NULL); + if (the_refs) + for (size_t i = 0; i < the_refs->nr; i++) { + struct reference ref = { + .name = the_refs->refname[i], + .oid = &the_refs->oid[i], + }; + fsck_handle_ref(&ref, NULL); + } + else + refs_for_each_rawref(get_main_ref_store(the_repository), + fsck_handle_ref, NULL); worktrees = get_worktrees(); for (p = worktrees; *p; p++) { @@ -964,6 +1007,14 @@ int cmd_fsck(int argc, { int i; struct odb_source *source; + struct ref_snapshot default_refs_snapshot = { + .nr = 0, + .name_alloc = 0, + .oid_alloc = 0, + .refname = NULL, + .oid = NULL + }; + bool use_snapshot; /* fsck knows how to handle missing promisor objects */ fetch_if_missing = 0; @@ -999,6 +1050,19 @@ int cmd_fsck(int argc, if (check_references) fsck_refs(the_repository); + /* + * Take a snapshot of the refs before walking objects to avoid looking + * at a set of refs that may be changed by the user while we are walking + * objects. We can still walk over new objects that are added during the + * execution of fsck but won't miss any objects that were reachable. + */ + use_snapshot = !argc; + if (use_snapshot) { + now = time(NULL); + refs_for_each_rawref(get_main_ref_store(the_repository), + snapshot_refs, &default_refs_snapshot); + } + if (connectivity_only) { for_each_loose_object(the_repository->objects, mark_loose_for_connectivity, NULL, 0); @@ -1071,7 +1135,7 @@ int cmd_fsck(int argc, * in this case (ie this implies --cache). */ if (!argc) { - get_default_heads(); + get_default_heads(use_snapshot ? &default_refs_snapshot : NULL); keep_cache_objects = 1; } @@ -1148,5 +1212,7 @@ int cmd_fsck(int argc, } } + if (use_snapshot) + free_snapshot_refs(&default_refs_snapshot); return errors_found; }