diff mbox series

[v3,04/20] global: ensure that object IDs are always padded

Message ID 3947c0c04dafa470facb7c9dd03133d69d09a161.1718347699.git.ps@pks.im (mailing list archive)
State Accepted
Commit c98d762ed96ba9beead28608b3c7d1e477a8fe3d
Headers show
Series Introduce `USE_THE_REPOSITORY_VARIABLE` macro | expand

Commit Message

Patrick Steinhardt June 14, 2024, 6:49 a.m. UTC
The `oidcmp()` and `oideq()` functions only compare the prefix length as
specified by the given hash algorithm. This mandates that the object IDs
have a valid hash algorithm set, or otherwise we wouldn't be able to
figure out that prefix. As we do not have a hash algorithm in many
cases, for example when handling null object IDs, this assumption cannot
always be fulfilled. We thus have a fallback in place that instead uses
`the_repository` to derive the hash function. This implicit dependency
is hidden away from callers and can be quite surprising, especially in
contexts where there may be no repository.

In theory, we can adapt those functions to always memcmp(3P) the whole
length of their hash arrays. But there exist a couple of sites where we
populate `struct object_id`s such that only the prefix of its hash that
is actually used by the hash algorithm is populated. The remaining bytes
are left uninitialized. The fact that those bytes are uninitialized also
leads to warnings under Valgrind in some places where we copy those
bytes.

Refactor callsites where we populate object IDs to always initialize all
bytes. This also allows us to get rid of `oidcpy_with_padding()`, for
one because the input is now fully initialized, and because `oidcpy()`
will now always copy the whole hash array.

Signed-off-by: Patrick Steinhardt <ps@pks.im>
---
 hash-ll.h           |  2 ++
 hash.h              | 16 ----------------
 hex.c               |  6 +++++-
 http-push.c         |  1 +
 notes.c             |  2 ++
 object-file.c       |  2 ++
 oidtree.c           |  4 ++--
 parallel-checkout.c |  8 +-------
 8 files changed, 15 insertions(+), 26 deletions(-)
diff mbox series

Patch

diff --git a/hash-ll.h b/hash-ll.h
index dbb96369fc..b72f84f4ae 100644
--- a/hash-ll.h
+++ b/hash-ll.h
@@ -288,6 +288,8 @@  static inline void oidread(struct object_id *oid, const unsigned char *hash,
 			   const struct git_hash_algo *algop)
 {
 	memcpy(oid->hash, hash, algop->rawsz);
+	if (algop->rawsz < GIT_MAX_RAWSZ)
+		memset(oid->hash + algop->rawsz, 0, GIT_MAX_RAWSZ - algop->rawsz);
 	oid->algo = hash_algo_by_ptr(algop);
 }
 
diff --git a/hash.h b/hash.h
index 43623a0c86..e43e3d8b5a 100644
--- a/hash.h
+++ b/hash.h
@@ -31,22 +31,6 @@  static inline int is_null_oid(const struct object_id *oid)
 	return oideq(oid, null_oid());
 }
 
-/* Like oidcpy() but zero-pads the unused bytes in dst's hash array. */
-static inline void oidcpy_with_padding(struct object_id *dst,
-				       const struct object_id *src)
-{
-	size_t hashsz;
-
-	if (!src->algo)
-		hashsz = the_hash_algo->rawsz;
-	else
-		hashsz = hash_algos[src->algo].rawsz;
-
-	memcpy(dst->hash, src->hash, hashsz);
-	memset(dst->hash + hashsz, 0, GIT_MAX_RAWSZ - hashsz);
-	dst->algo = src->algo;
-}
-
 static inline int is_empty_blob_oid(const struct object_id *oid)
 {
 	return oideq(oid, the_hash_algo->empty_blob);
diff --git a/hex.c b/hex.c
index d42262bdca..bc9e86a978 100644
--- a/hex.c
+++ b/hex.c
@@ -25,8 +25,12 @@  int get_oid_hex_algop(const char *hex, struct object_id *oid,
 		      const struct git_hash_algo *algop)
 {
 	int ret = get_hash_hex_algop(hex, oid->hash, algop);
-	if (!ret)
+	if (!ret) {
 		oid_set_algo(oid, algop);
+		if (algop->rawsz != GIT_MAX_RAWSZ)
+			memset(oid->hash + algop->rawsz, 0,
+			       GIT_MAX_RAWSZ - algop->rawsz);
+	}
 	return ret;
 }
 
diff --git a/http-push.c b/http-push.c
index 86de238b84..a97df4a1fb 100644
--- a/http-push.c
+++ b/http-push.c
@@ -1016,6 +1016,7 @@  static void remote_ls(const char *path, int flags,
 /* extract hex from sharded "xx/x{38}" filename */
 static int get_oid_hex_from_objpath(const char *path, struct object_id *oid)
 {
+	memset(oid->hash, 0, GIT_MAX_RAWSZ);
 	oid->algo = hash_algo_by_ptr(the_hash_algo);
 
 	if (strlen(path) != the_hash_algo->hexsz + 1)
diff --git a/notes.c b/notes.c
index 3a8da92fb9..afe2e2882e 100644
--- a/notes.c
+++ b/notes.c
@@ -427,6 +427,8 @@  static void load_subtree(struct notes_tree *t, struct leaf_node *subtree,
 					 hashsz - prefix_len))
 				goto handle_non_note; /* entry.path is not a SHA1 */
 
+			memset(object_oid.hash + hashsz, 0, GIT_MAX_RAWSZ - hashsz);
+
 			type = PTR_TYPE_NOTE;
 		} else if (path_len == 2) {
 			/* This is potentially an internal node */
diff --git a/object-file.c b/object-file.c
index c161e3e2a5..bb97f8a809 100644
--- a/object-file.c
+++ b/object-file.c
@@ -2743,6 +2743,8 @@  int for_each_file_in_obj_subdir(unsigned int subdir_nr,
 		    !hex_to_bytes(oid.hash + 1, de->d_name,
 				  the_hash_algo->rawsz - 1)) {
 			oid_set_algo(&oid, the_hash_algo);
+			memset(oid.hash + the_hash_algo->rawsz, 0,
+			       GIT_MAX_RAWSZ - the_hash_algo->rawsz);
 			if (obj_cb) {
 				r = obj_cb(&oid, path->buf, data);
 				if (r)
diff --git a/oidtree.c b/oidtree.c
index daef175dc7..92d03b52db 100644
--- a/oidtree.c
+++ b/oidtree.c
@@ -42,7 +42,7 @@  void oidtree_insert(struct oidtree *ot, const struct object_id *oid)
 	 * Clear the padding and copy the result in separate steps to
 	 * respect the 4-byte alignment needed by struct object_id.
 	 */
-	oidcpy_with_padding(&k, oid);
+	oidcpy(&k, oid);
 	memcpy(on->k, &k, sizeof(k));
 
 	/*
@@ -60,7 +60,7 @@  int oidtree_contains(struct oidtree *ot, const struct object_id *oid)
 	struct object_id k;
 	size_t klen = sizeof(k);
 
-	oidcpy_with_padding(&k, oid);
+	oidcpy(&k, oid);
 
 	if (oid->algo == GIT_HASH_UNKNOWN)
 		klen -= sizeof(oid->algo);
diff --git a/parallel-checkout.c b/parallel-checkout.c
index b5a714c711..08b960aac8 100644
--- a/parallel-checkout.c
+++ b/parallel-checkout.c
@@ -429,13 +429,7 @@  static void send_one_item(int fd, struct parallel_checkout_item *pc_item)
 	fixed_portion->ident = pc_item->ca.ident;
 	fixed_portion->name_len = name_len;
 	fixed_portion->working_tree_encoding_len = working_tree_encoding_len;
-	/*
-	 * We pad the unused bytes in the hash array because, otherwise,
-	 * Valgrind would complain about passing uninitialized bytes to a
-	 * write() syscall. The warning doesn't represent any real risk here,
-	 * but it could hinder the detection of actual errors.
-	 */
-	oidcpy_with_padding(&fixed_portion->oid, &pc_item->ce->oid);
+	oidcpy(&fixed_portion->oid, &pc_item->ce->oid);
 
 	variant = data + sizeof(*fixed_portion);
 	if (working_tree_encoding_len) {