[PATCH] fsck: detect bare repos in trees and warn
- From
Glen Choo <chooglen@google.com>
- Date
- Apr 6, 2022, 23:22 UTC
- Message-ID
- <20220406232231.47714-1-chooglen@google.com>
- In-Reply-To
- <kl6lsfqpygsj.fsf@chooglen-macbookpro.roam.corp.google.com>
Git tries not to distribute configs in-repo because they are a security risk. However, an attacker can do exactly this if they embed a bare repo inside of another repo.
Teach fsck to detect whether a tree object contains a bare repo (as determined by setup.c) and warn. This will help hosting sites detect and prevent transmission of such malicious repos.
See [1] for a more in-depth discussion, including future steps and alternatives.
[1] https://lore.kernel.org/git/kl6lsfqpygsj.fsf@chooglen-macbookpro.roam.corp.google.com/
Signed-off-by: Glen Choo <chooglen@google.com> --- fsck.c | 19 +++++++++++++++++++ fsck.h | 1 + setup.c | 4 ++++ t/t1450-fsck.sh | 36 ++++++++++++++++++++++++++++++++++++ 4 files changed, 60 insertions(+)
diff --git a/fsck.c b/fsck.c index 3ec500d707..11c11c348a 100644 --- a/fsck.c +++ b/fsck.c @@ -573,6 +573,9 @@ static int fsck_tree(const struct object_id *tree_oid, int has_bad_modes = 0; int has_dup_entries = 0; int not_properly_sorted = 0; + int has_head = 0; + int has_refs_entry = 0; + int has_objects_entry = 0; struct tree_desc desc; unsigned o_mode; const char *o_name; @@ -602,6 +605,12 @@ static int fsck_tree(const struct object_id *tree_oid, has_dotdot |= !strcmp(name, ".."); has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name); has_zero_pad |= *(char *)desc.buffer == '0'; + has_head |= !strcasecmp(name, "HEAD") + && (S_ISLNK(mode) || S_ISREG(mode)); + has_refs_entry |= !strcasecmp(name, "refs") + && (S_ISLNK(mode) || S_ISDIR(mode)); + has_objects_entry |= !strcasecmp(name, "objects") + && (S_ISLNK(mode) || S_ISDIR(mode)); if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) { if (!S_ISLNK(mode)) @@ -739,6 +748,16 @@ static int fsck_tree(const struct object_id *tree_oid, retval += report(options, tree_oid, OBJ_TREE, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted"); + /* + * Determine if this tree looks like a bare repository according + * to the rules of setup.c. If those are changed, this should be + * changed too. + */ + if (has_head && has_refs_entry && has_objects_entry) + retval += report(options, tree_oid, OBJ_TREE, + FSCK_MSG_EMBEDDED_BARE_REPO, + "contains bare repository"); + return retval; } diff --git a/fsck.h b/fsck.h index d07f7a2459..3f0f73b0f3 100644 --- a/fsck.h +++ b/fsck.h @@ -65,6 +65,7 @@ enum fsck_msg_type { FUNC(NULL_SHA1, WARN) \ FUNC(ZERO_PADDED_FILEMODE, WARN) \ FUNC(NUL_IN_COMMIT, WARN) \ + FUNC(EMBEDDED_BARE_REPO, WARN) \ /* infos (reported as warnings, but ignored by default) */ \ FUNC(GITMODULES_PARSE, INFO) \ FUNC(GITIGNORE_SYMLINK, INFO) \ diff --git a/setup.c b/setup.c index 04ce33cdcd..2600548776 100644 --- a/setup.c +++ b/setup.c @@ -336,6 +336,10 @@ int get_common_dir_noenv(struct strbuf *sb, const char *gitdir) * - either a HEAD symlink or a HEAD file that is formatted as * a proper "ref:", or a regular file HEAD that has a properly * formatted sha1 object name. + * + * fsck.c checks for bare repositories in trees using similar rules, but a + * duplicated implementation. If these are changed, the correspnding code in + * fsck.c should change too. */ int is_git_directory(const char *suspect) { diff --git a/t/t1450-fsck.sh b/t/t1450-fsck.sh index de50c0ea01..a65827bc03 100755 --- a/t/t1450-fsck.sh +++ b/t/t1450-fsck.sh @@ -563,6 +563,42 @@ dot-backslash-case .\\\\.GIT\\\\foobar dotgit-case-backslash .git\\\\foobar EOF +test_expect_success "fsck notices bare repo" ' +( + mkdir -p embedded-bare-repo/bare && + git init embedded-bare-repo && + ( + cd embedded-bare-repo/bare && + echo content >HEAD && + mkdir refs/ objects/ && + echo content >refs/foo && + echo content >objects/foo && + git add . && + git commit -m base && + bad_tree=$(git rev-parse HEAD:bare) && + git fsck 2>out && + test_i18ngrep "warning.*tree $bad_tree: embeddedBareRepo: contains bare repository" out + ) +)' + +test_expect_success "fsck notices bare repo with odd casing" ' +( + mkdir -p embedded-bare-repo-case/bare && + git init embedded-bare-repo-case && + ( + cd embedded-bare-repo-case/bare && + echo content >heAD && + mkdir Refs/ objectS/ && + echo content >Refs/foo && + echo content >objectS/foo && + git add . && + git commit -m base && + bad_tree=$(git rev-parse HEAD:bare) && + git fsck 2>out && + test_i18ngrep "warning.*tree $bad_tree: embeddedBareRepo: contains bare repository" out + ) +)' + test_expect_success 'fsck allows .Ňit' ' ( git init not-dotgit && base-commit: 805e0a68082a217f0112db9ee86a022227a9c81b
-- 2.33.GIT