mirror of
https://git.yoctoproject.org/poky
synced 2026-02-12 03:33:02 +01:00
d08e8d726d
CVE-2024-32002: Git is a revision control system. Prior to versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4, repositories with submodules can be crafted in a way that exploits a bug in Git whereby it can be fooled into writing files not into the submodule's worktree but into a `.git/` directory. This allows writing a hook that will be executed while the clone operation is still running, giving the user no opportunity to inspect the code that is being executed. The problem has been patched in versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4. If symbolic link support is disabled in Git (e.g. via `git config --global core.symlinks false`), the described attack won't work. As always, it is best to avoid cloning repositories from untrusted sources. CVE-2024-32004: Git is a revision control system. Prior to versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4, an attacker can prepare a local repository in such a way that, when cloned, will execute arbitrary code during the operation. The problem has been patched in versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4. As a workaround, avoid cloning repositories from untrusted sources. CVE-2024-32020: Git is a revision control system. Prior to versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4, local clones may end up hardlinking files into the target repository's object database when source and target repository reside on the same disk. If the source repository is owned by a different user, then those hardlinked files may be rewritten at any point in time by the untrusted user. Cloning local repositories will cause Git to either copy or hardlink files of the source repository into the target repository. This significantly speeds up such local clones compared to doing a "proper" clone and saves both disk space and compute time. When cloning a repository located on the same disk that is owned by a different user than the current user we also end up creating such hardlinks. These files will continue to be owned and controlled by the potentially-untrusted user and can be rewritten by them at will in the future. The problem has been patched in versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4. CVE-2024-32021: Git is a revision control system. Prior to versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4, when cloning a local source repository that contains symlinks via the filesystem, Git may create hardlinks to arbitrary user-readable files on the same filesystem as the target repository in the `objects/` directory. Cloning a local repository over the filesystem may creating hardlinks to arbitrary user-owned files on the same filesystem in the target Git repository's `objects/` directory. When cloning a repository over the filesystem (without explicitly specifying the `file://` protocol or `--no-local`), the optimizations for local cloning will be used, which include attempting to hard link the object files instead of copying them. While the code includes checks against symbolic links in the source repository, which were added during the fix for CVE-2022-39253, these checks can still be raced because the hard link operation ultimately follows symlinks. If the object on the filesystem appears as a file during the check, and then a symlink during the operation, this will allow the adversary to bypass the check and create hardlinks in the destination objects directory to arbitrary, user-readable files. The problem has been patched in versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4. CVE-2024-32465: Git is a revision control system. The Git project recommends to avoid working in untrusted repositories, and instead to clone it first with `git clone --no-local` to obtain a clean copy. Git has specific protections to make that a safe operation even with an untrusted source repository, but vulnerabilities allow those protections to be bypassed. In the context of cloning local repositories owned by other users, this vulnerability has been covered in CVE-2024-32004. But there are circumstances where the fixes for CVE-2024-32004 are not enough: For example, when obtaining a `.zip` file containing a full copy of a Git repository, it should not be trusted by default to be safe, as e.g. hooks could be configured to run within the context of that repository. The problem has been patched in versions 2.45.1, 2.44.1, 2.43.4, 2.42.2, 2.41.1, 2.40.2, and 2.39.4. As a workaround, avoid using Git in repositories that have been obtained via archives from untrusted sources. References: https://nvd.nist.gov/vuln/detail/CVE-2024-32002 https://nvd.nist.gov/vuln/detail/CVE-2024-32004 https://nvd.nist.gov/vuln/detail/CVE-2024-32020 https://nvd.nist.gov/vuln/detail/CVE-2024-32021 https://nvd.nist.gov/vuln/detail/CVE-2024-32465 (From OE-Core rev: 209c41377abf6853455b00af3923f1b244a3766b) Signed-off-by: Soumya Sambu <soumya.sambu@windriver.com> Signed-off-by: Steve Sakoman <steve@sakoman.com>
115 lines
4.4 KiB
Diff
115 lines
4.4 KiB
Diff
From 1204e1a824c34071019fe106348eaa6d88f9528d Mon Sep 17 00:00:00 2001
|
|
From: Patrick Steinhardt <ps@pks.im>
|
|
Date: Mon, 15 Apr 2024 13:30:41 +0200
|
|
Subject: [PATCH] builtin/clone: refuse local clones of unsafe repositories
|
|
|
|
When performing a local clone of a repository we end up either copying
|
|
or hardlinking the source repository into the target repository. This is
|
|
significantly more performant than if we were to use git-upload-pack(1)
|
|
and git-fetch-pack(1) to create the new repository and preserves both
|
|
disk space and compute time.
|
|
|
|
Unfortunately though, performing such a local clone of a repository that
|
|
is not owned by the current user is inherently unsafe:
|
|
|
|
- It is possible that source files get swapped out underneath us while
|
|
we are copying or hardlinking them. While we do perform some checks
|
|
here to assert that we hardlinked the expected file, they cannot
|
|
reliably thwart time-of-check-time-of-use (TOCTOU) style races. It
|
|
is thus possible for an adversary to make us copy or hardlink
|
|
unexpected files into the target directory.
|
|
|
|
Ideally, we would address this by starting to use openat(3P),
|
|
fstatat(3P) and friends. Due to platform compatibility with Windows
|
|
we cannot easily do that though. Furthermore, the scope of these
|
|
fixes would likely be quite broad and thus not fit for an embargoed
|
|
security release.
|
|
|
|
- Even if we handled TOCTOU-style races perfectly, hardlinking files
|
|
owned by a different user into the target repository is not a good
|
|
idea in general. It is possible for an adversary to rewrite those
|
|
files to contain whatever data they want even after the clone has
|
|
completed.
|
|
|
|
Address these issues by completely refusing local clones of a repository
|
|
that is not owned by the current user. This reuses our existing infra we
|
|
have in place via `ensure_valid_ownership()` and thus allows a user to
|
|
override the safety guard by adding the source repository path to the
|
|
"safe.directory" configuration.
|
|
|
|
This addresses CVE-2024-32020.
|
|
|
|
Signed-off-by: Patrick Steinhardt <ps@pks.im>
|
|
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
|
|
|
|
CVE: CVE-2024-32020
|
|
|
|
Upstream-Status: Backport [https://github.com/git/git/commit/1204e1a824c34071019fe106348eaa6d88f9528d]
|
|
|
|
Signed-off-by: Soumya Sambu <soumya.sambu@windriver.com>
|
|
---
|
|
builtin/clone.c | 14 ++++++++++++++
|
|
t/t0033-safe-directory.sh | 24 ++++++++++++++++++++++++
|
|
2 files changed, 38 insertions(+)
|
|
|
|
diff --git a/builtin/clone.c b/builtin/clone.c
|
|
index 4541a55..11f6b4b 100644
|
|
--- a/builtin/clone.c
|
|
+++ b/builtin/clone.c
|
|
@@ -312,6 +312,20 @@ static void copy_or_link_directory(struct strbuf *src, struct strbuf *dest,
|
|
int iter_status;
|
|
struct strbuf realpath = STRBUF_INIT;
|
|
|
|
+ /*
|
|
+ * Refuse copying directories by default which aren't owned by us. The
|
|
+ * code that performs either the copying or hardlinking is not prepared
|
|
+ * to handle various edge cases where an adversary may for example
|
|
+ * racily swap out files for symlinks. This can cause us to
|
|
+ * inadvertently use the wrong source file.
|
|
+ *
|
|
+ * Furthermore, even if we were prepared to handle such races safely,
|
|
+ * creating hardlinks across user boundaries is an inherently unsafe
|
|
+ * operation as the hardlinked files can be rewritten at will by the
|
|
+ * potentially-untrusted user. We thus refuse to do so by default.
|
|
+ */
|
|
+ die_upon_dubious_ownership(NULL, NULL, src_repo);
|
|
+
|
|
mkdir_if_missing(dest->buf, 0777);
|
|
|
|
iter = dir_iterator_begin(src->buf, DIR_ITERATOR_PEDANTIC);
|
|
diff --git a/t/t0033-safe-directory.sh b/t/t0033-safe-directory.sh
|
|
index 239d93f..751cba5 100755
|
|
--- a/t/t0033-safe-directory.sh
|
|
+++ b/t/t0033-safe-directory.sh
|
|
@@ -46,4 +46,28 @@ test_expect_success 'safe.directory=*, but is reset' '
|
|
expect_rejected_dir
|
|
'
|
|
|
|
+test_expect_success 'local clone of unowned repo refused in unsafe directory' '
|
|
+ test_when_finished "rm -rf source" &&
|
|
+ git init source &&
|
|
+ (
|
|
+ sane_unset GIT_TEST_ASSUME_DIFFERENT_OWNER &&
|
|
+ test_commit -C source initial
|
|
+ ) &&
|
|
+ test_must_fail git clone --local source target &&
|
|
+ test_path_is_missing target
|
|
+'
|
|
+
|
|
+test_expect_success 'local clone of unowned repo accepted in safe directory' '
|
|
+ test_when_finished "rm -rf source" &&
|
|
+ git init source &&
|
|
+ (
|
|
+ sane_unset GIT_TEST_ASSUME_DIFFERENT_OWNER &&
|
|
+ test_commit -C source initial
|
|
+ ) &&
|
|
+ test_must_fail git clone --local source target &&
|
|
+ git config --global --add safe.directory "$(pwd)/source/.git" &&
|
|
+ git clone --local source target &&
|
|
+ test_path_is_dir target
|
|
+'
|
|
+
|
|
test_done
|
|
--
|
|
2.40.0
|