git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH v5 18/27] refs: resolve symbolic refs first

From
David Turner <dturner@twopensource.com>
Date
Feb 18, 2016, 05:17 UTC
Message-ID
<1455772670-21142-19-git-send-email-dturner@twopensource.com>
In-Reply-To
<1455772670-21142-1-git-send-email-dturner@twopensource.com>

Before committing ref updates, split symbolic ref updates into two parts: an update to the underlying ref, and a log-only update to the symbolic ref. This ensures that both references are locked correctly while their reflogs are updated.

It is still possible to confuse git by concurrent updates, since the splitting of symbolic refs does not happen under lock. So a symbolic ref could be replaced by a plain ref in the middle of this operation, which would lead to reflog discontinuities and missed old-ref checks.

All callers to lock_ref_sha1_basic outside of ref_transaction_commit now set REF_NODEREF. And ref_transaction_commit only happens on already-derefernced refs. So we can assume REF_NODEREF when resolving inside this function.

Signed-off-by: David Turner <dturner@twopensource.com>
---
 refs.c               |  69 ++++++++++++++++++++++++++
 refs/files-backend.c | 137 ++++++++++++++++++++++++++-------------------------
 refs/refs-internal.h |   8 +++
 3 files changed, 148 insertions(+), 66 deletions(-)
diff --git a/refs.c b/refs.c
index c9fa34d..01cf6c3 100644
--- a/refs.c
+++ b/refs.c
@@ -1227,6 +1227,71 @@ int refs_init_db(int shared, struct strbuf *err)
 	return the_refs_backend->init_db(shared, err);
 }
 
+/*
+ * Special case for symbolic refs when REF_NODEREF is not turned on.
+ * Dereference them here, mark them REF_LOG_ONLY, and add an update
+ * for the underlying ref.
+ */
+static int dereference_symrefs(struct ref_transaction *transaction,
+			       struct strbuf *err)
+{
+	int i;
+	int nr = transaction->nr;
+
+	for (i = 0; i < nr; i++) {
+		struct ref_update *update = transaction->updates[i];
+		const char *resolved;
+		unsigned char sha1[20];
+		int resolve_flags = 0;
+		int mustexist = update->flags & REF_HAVE_OLD &&
+			!is_null_sha1(update->old_sha1);
+		int deleting = (update->flags & REF_HAVE_NEW) &&
+			is_null_sha1(update->new_sha1);
+
+		if (mustexist)
+			resolve_flags |= RESOLVE_REF_READING;
+		if (deleting)
+			resolve_flags |= RESOLVE_REF_ALLOW_BAD_NAME |
+				RESOLVE_REF_NO_RECURSE;
+
+		if (strcmp(update->refname, "HEAD"))
+			update->flags |= REF_IS_NOT_HEAD;
+
+		resolved = resolve_ref_unsafe(update->refname, resolve_flags,
+					      sha1, &update->type);
+		if (!resolved) {
+			/*
+			 * We may notice this breakage later and die
+			 * with a sensible error message
+			 */
+			update->type |= REF_ISBROKEN;
+			continue;
+		}
+
+		hashcpy(update->read_sha1, sha1);
+
+		if (update->flags & REF_NODEREF ||
+		    !(update->type & REF_ISSYMREF))
+			continue;
+
+		/* Create a new transaction for the underlying ref */
+		if (ref_transaction_update(transaction,
+					   resolved,
+					   update->new_sha1,
+					   (update->flags & REF_HAVE_OLD) ?
+					   update->old_sha1 : NULL,
+					   update->flags & ~REF_IS_NOT_HEAD,
+					   update->msg, err))
+			return -1;
+
+		/* Make the symbolic ref update non-recursive */
+		update->flags |= REF_LOG_ONLY | REF_NODEREF;
+		update->flags &= ~REF_HAVE_OLD;
+	}
+
+	return 0;
+}
+
 int ref_transaction_commit(struct ref_transaction *transaction,
 			   struct strbuf *err)
 {
@@ -1243,6 +1308,10 @@ int ref_transaction_commit(struct ref_transaction *transaction,
 		return 0;
 	}
 
+	ret = dereference_symrefs(transaction, err);
+	if (ret)
+		goto done;
+
 	if (get_affected_refnames(transaction, &affected_refnames, err)) {
 		ret = TRANSACTION_GENERIC_ERROR;
 		goto done;
diff --git a/refs/files-backend.c b/refs/files-backend.c
index 42f2d5b..b3cca8e 100644
--- a/refs/files-backend.c
+++ b/refs/files-backend.c
@@ -7,7 +7,6 @@
 
 struct ref_lock {
 	char *ref_name;
-	char *orig_ref_name;
 	struct lock_file *lk;
 	struct object_id old_oid;
 };
@@ -1779,7 +1778,6 @@ static void unlock_ref(struct ref_lock *lock)
 	if (lock->lk)
 		rollback_lock_file(lock->lk);
 	free(lock->ref_name);
-	free(lock->orig_ref_name);
 	free(lock);
 }
 
@@ -1835,6 +1833,7 @@ static int remove_empty_directories(struct strbuf *path)
  */
 static struct ref_lock *lock_ref_sha1_basic(const char *refname,
 					    const unsigned char *old_sha1,
+					    const unsigned char *read_sha1,
 					    const struct string_list *extras,
 					    const struct string_list *skip,
 					    unsigned int flags, int *type_p,
@@ -1842,14 +1841,14 @@ static struct ref_lock *lock_ref_sha1_basic(const char *refname,
 {
 	struct strbuf ref_file = STRBUF_INIT;
 	struct strbuf orig_ref_file = STRBUF_INIT;
-	const char *orig_refname = refname;
 	struct ref_lock *lock;
 	int last_errno = 0;
 	int type;
 	int lflags = 0;
 	int mustexist = (old_sha1 && !is_null_sha1(old_sha1));
-	int resolve_flags = 0;
+	int resolve_flags = RESOLVE_REF_NO_RECURSE;
 	int attempts_remaining = 3;
+	int resolved;
 
 	assert(err);
 
@@ -1859,65 +1858,65 @@ static struct ref_lock *lock_ref_sha1_basic(const char *refname,
 		resolve_flags |= RESOLVE_REF_READING;
 	if (flags & REF_DELETING)
 		resolve_flags |= RESOLVE_REF_ALLOW_BAD_NAME;
-	if (flags & REF_NODEREF) {
-		resolve_flags |= RESOLVE_REF_NO_RECURSE;
+	if (flags & REF_NODEREF)
 		lflags |= LOCK_NO_DEREF;
-	}
 
-	refname = resolve_ref_unsafe(refname, resolve_flags,
-				     lock->old_oid.hash, &type);
-	if (!refname && errno == EISDIR) {
-		/*
-		 * we are trying to lock foo but we used to
-		 * have foo/bar which now does not exist;
-		 * it is normal for the empty directory 'foo'
-		 * to remain.
-		 */
-		strbuf_git_path(&orig_ref_file, "%s", orig_refname);
-		if (remove_empty_directories(&orig_ref_file)) {
+	if (type_p && *type_p & REF_ISSYMREF) {
+		hashcpy(lock->old_oid.hash, read_sha1);
+	} else {
+		resolved = !!resolve_ref_unsafe(refname, resolve_flags,
+						lock->old_oid.hash, &type);
+		if (!resolved && errno == EISDIR) {
+			/*
+			 * we are trying to lock foo but we used to
+			 * have foo/bar which now does not exist;
+			 * it is normal for the empty directory 'foo'
+			 * to remain.
+			 */
+			strbuf_git_path(&orig_ref_file, "%s", refname);
+			if (remove_empty_directories(&orig_ref_file)) {
+				struct ref_dir *loose_refs;
+				loose_refs = get_loose_refs(&ref_cache);
+				last_errno = errno;
+				if (!verify_refname_available_dir(refname, extras, skip,
+								  loose_refs, err))
+					strbuf_addf(err, "there are still refs under '%s'",
+						    refname);
+				goto error_return;
+			}
+			resolved = !!resolve_ref_unsafe(refname, resolve_flags,
+							lock->old_oid.hash, &type);
+		}
+
+		if (type_p)
+			*type_p = type;
+		if (!resolved) {
 			last_errno = errno;
-			if (!verify_refname_available_dir(orig_refname, extras, skip,
+			if (last_errno != ENOTDIR ||
+			    !verify_refname_available_dir(refname, extras, skip,
 							  get_loose_refs(&ref_cache), err))
-				strbuf_addf(err, "there are still refs under '%s'",
-					    orig_refname);
+				strbuf_addf(err,
+					    "unable to resolve reference %s: %s",
+					    refname, strerror(last_errno));
+
+			goto error_return;
+		}
+		/*
+		 * If the ref did not exist and we are creating it, make sure
+		 * there is no existing packed ref whose name begins with our
+		 * refname, nor a packed ref whose name is a proper prefix of
+		 * our refname.
+		 */
+		if (is_null_oid(&lock->old_oid) &&
+		    verify_refname_available_dir(refname, extras, skip,
+						 get_packed_refs(&ref_cache), err)) {
+			last_errno = ENOTDIR;
 			goto error_return;
 		}
-		refname = resolve_ref_unsafe(orig_refname, resolve_flags,
-					     lock->old_oid.hash, &type);
-	}
-	if (type_p)
-	    *type_p = type;
-	if (!refname) {
-		last_errno = errno;
-		if (last_errno != ENOTDIR ||
-		    !verify_refname_available_dir(orig_refname, extras, skip,
-						  get_loose_refs(&ref_cache), err))
-			strbuf_addf(err, "unable to resolve reference %s: %s",
-				    orig_refname, strerror(last_errno));
-
-		goto error_return;
-	}
-
-	if (flags & REF_NODEREF)
-		refname = orig_refname;
-
-	/*
-	 * If the ref did not exist and we are creating it, make sure
-	 * there is no existing packed ref whose name begins with our
-	 * refname, nor a packed ref whose name is a proper prefix of
-	 * our refname.
-	 */
-	if (is_null_oid(&lock->old_oid) &&
-	    verify_refname_available_dir(refname, extras, skip,
-					 get_packed_refs(&ref_cache), err)) {
-		last_errno = ENOTDIR;
-		goto error_return;
 	}
-
 	lock->lk = xcalloc(1, sizeof(struct lock_file));
 
 	lock->ref_name = xstrdup(refname);
-	lock->orig_ref_name = xstrdup(orig_refname);
 	strbuf_git_path(&ref_file, "%s", refname);
 
  retry:
@@ -1949,7 +1948,13 @@ static struct ref_lock *lock_ref_sha1_basic(const char *refname,
 			goto error_return;
 		}
 	}
-	if (verify_lock(lock, old_sha1, mustexist, err)) {
+
+	if (type_p && *type_p & REF_ISSYMREF && !(*type_p & REF_ISBROKEN)) {
+		/*
+		 * Old hash verification for symrefs happens on their
+		 * base ref.
+		 */
+	} else if (verify_lock(lock, old_sha1, mustexist, err)) {
 		last_errno = errno;
 		goto error_return;
 	}
@@ -2484,8 +2489,8 @@ static int files_rename_ref(const char *oldrefname, const char *newrefname,
 
 	logmoved = log;
 
-	lock = lock_ref_sha1_basic(newrefname, NULL, NULL, NULL, REF_NODEREF,
-				   NULL, &err);
+	lock = lock_ref_sha1_basic(newrefname, NULL, NULL, NULL, NULL,
+				   REF_NODEREF, NULL, &err);
 	if (!lock) {
 		error("unable to rename '%s' to '%s': %s", oldrefname, newrefname, err.buf);
 		strbuf_release(&err);
@@ -2503,8 +2508,9 @@ static int files_rename_ref(const char *oldrefname, const char *newrefname,
 	return 0;
 
  rollback:
-	lock = lock_ref_sha1_basic(oldrefname, NULL, NULL, NULL, REF_NODEREF,
-				   NULL, &err);
+	lock = lock_ref_sha1_basic(oldrefname, NULL, NULL, NULL, NULL,
+				   REF_NODEREF, NULL, &err);
+
 	if (!lock) {
 		error("unable to lock %s for rollback: %s", oldrefname, err.buf);
 		strbuf_release(&err);
@@ -2732,9 +2738,7 @@ static int commit_ref_update(struct ref_lock *lock,
 			     int flags, struct strbuf *err)
 {
 	clear_loose_ref_cache(&ref_cache);
-	if (log_ref_write(lock->ref_name, lock->old_oid.hash, sha1, logmsg, flags, err) < 0 ||
-	    (strcmp(lock->ref_name, lock->orig_ref_name) &&
-	     log_ref_write(lock->orig_ref_name, lock->old_oid.hash, sha1, logmsg, flags, err) < 0)) {
+	if (log_ref_write(lock->ref_name, lock->old_oid.hash, sha1, logmsg, flags, err) < 0) {
 		char *old_msg = strbuf_detach(err, NULL);
 		strbuf_addf(err, "Cannot update the ref '%s': %s",
 			    lock->ref_name, old_msg);
@@ -2742,7 +2746,7 @@ static int commit_ref_update(struct ref_lock *lock,
 		unlock_ref(lock);
 		return -1;
 	}
-	if (strcmp(lock->orig_ref_name, "HEAD") != 0) {
+	if (flags & REF_IS_NOT_HEAD) {
 		/*
 		 * Special hack: If a branch is updated directly and HEAD
 		 * points to it (may happen on the remote side of a push
@@ -2837,8 +2841,8 @@ static int files_create_symref(const char *refname,
 	struct ref_lock *lock;
 	int ret;
 
-	lock = lock_ref_sha1_basic(refname, NULL, NULL, NULL, REF_NODEREF, NULL,
-				   &err);
+	lock = lock_ref_sha1_basic(refname, NULL, NULL, NULL, NULL, REF_NODEREF,
+				   NULL, &err);
 	if (!lock) {
 		error("%s", err.buf);
 		strbuf_release(&err);
@@ -3106,6 +3110,7 @@ static int files_transaction_commit(struct ref_transaction *transaction,
 				update->refname,
 				((update->flags & REF_HAVE_OLD) ?
 				 update->old_sha1 : NULL),
+				update->read_sha1,
 				affected_refnames, NULL,
 				update->flags,
 				&update->type,
@@ -3352,7 +3357,7 @@ static int files_reflog_expire(const char *refname, const unsigned char *sha1,
 	struct ref_lock *lock;
 	char *log_file;
 	int status = 0;
-	int type;
+	int type = 0;
 	struct strbuf err = STRBUF_INIT;
 
 	memset(&cb, 0, sizeof(cb));
@@ -3365,7 +3370,7 @@ static int files_reflog_expire(const char *refname, const unsigned char *sha1,
 	 * reference itself, plus we might need to update the
 	 * reference if --updateref was specified:
 	 */
-	lock = lock_ref_sha1_basic(refname, sha1, NULL, NULL, REF_NODEREF,
+	lock = lock_ref_sha1_basic(refname, sha1, NULL, NULL, NULL, REF_NODEREF,
 				   &type, &err);
 	if (!lock) {
 		error("cannot lock ref '%s': %s", refname, err.buf);
diff --git a/refs/refs-internal.h b/refs/refs-internal.h
index 283f71f..0337d2e 100644
--- a/refs/refs-internal.h
+++ b/refs/refs-internal.h
@@ -49,6 +49,8 @@
  */
 #define REF_LOG_ONLY 0x80
 
+#define REF_IS_NOT_HEAD 0x100
+
 /*
  * Return true iff refname is minimally safe. "Safe" here means that
  * deleting a loose reference by this name will not do any damage, for
@@ -147,6 +149,12 @@ struct ref_update {
 	 */
 	unsigned char old_sha1[20];
 	/*
+	 * During the symbolic ref split stage, we resolve refs.
+	 * We'll re-resolve non-symbolic refs once they are locked,
+	 * but we store this to avoid re-resolving symbolic refs.
+	 */
+	unsigned char read_sha1[20];
+	/*
 	 * One or more of REF_HAVE_NEW, REF_HAVE_OLD, REF_NODEREF,
 	 * REF_DELETING, and REF_ISPRUNING:
 	 */

-- 
2.4.2.767.g62658d5-twtrsrc
Previous: David TurnerNext: David Turner
Message 21 of 52 in “refs backends”
  1. 00/27 refs backendsDavid Turner, Feb 18, 2016
  2. 01/27 refs: Move head_ref{,_submodule} to the common codeDavid Turner, Feb 18, 2016
  3. 02/27 refs: move for_each_*ref* functions into common codeDavid Turner, Feb 18, 2016
  4. 03/27 refs: add a backend method structure with transaction functionsDavid Turner, Feb 18, 2016
  5. 04/27 refs: add methods for misc ref operationsDavid Turner, Feb 18, 2016
  6. 05/27 refs: add method for do_for_each_refDavid Turner, Feb 18, 2016
  7. 06/27 refs: add do_for_each_per_worktree_refDavid Turner, Feb 18, 2016
  8. 07/27 refs: add methods for reflogDavid Turner, Feb 18, 2016
  9. 08/27 refs: add method for initial ref transaction commitDavid Turner, Feb 18, 2016
  10. 09/27 refs: add method for delete_refsDavid Turner, Feb 18, 2016
  11. 10/27 refs: add methods to init refs dbDavid Turner, Feb 18, 2016
  12. 11/27 refs: add method to rename refsDavid Turner, Feb 18, 2016
  13. 12/27 refs: forbid cross-backend ref renamesDavid Turner, Feb 18, 2016
  14. Duy NguyenFeb 20, 2016
  15. David TurnerFeb 24, 2016
  16. 13/27 refs: make lock genericDavid Turner, Feb 18, 2016
  17. 14/27 refs: move duplicate check to common codeDavid Turner, Feb 18, 2016
  18. 15/27 refs: allow log-only updatesDavid Turner, Feb 18, 2016
  19. 16/27 refs: don't dereference on renameDavid Turner, Feb 18, 2016
  20. 17/27 refs: on symref reflog expire, lock symref not referrentDavid Turner, Feb 18, 2016
  21. 18/27 refs: resolve symbolic refs firstDavid Turner, Feb 18, 2016
  22. 19/27 refs: always handle non-normal refs in files backendDavid Turner, Feb 18, 2016
  23. 20/27 init: allow alternate ref strorage to be set for new reposDavid Turner, Feb 18, 2016
  24. 21/27 refs: check submodules' ref storage configDavid Turner, Feb 18, 2016
  25. 22/27 clone: allow ref storage backend to be set for cloneDavid Turner, Feb 18, 2016
  26. 23/27 svn: learn ref-storage argumentDavid Turner, Feb 18, 2016
  27. Eric WongFeb 20, 2016
  28. David TurnerFeb 23, 2016
  29. 24/27 refs: add register_ref_storage_backends()David Turner, Feb 18, 2016
  30. 25/27 refs: add LMDB refs storage backendDavid Turner, Feb 18, 2016
  31. Duy NguyenFeb 18, 2016
  32. David TurnerFeb 18, 2016
  33. Junio C HamanoFeb 18, 2016
  34. Duy NguyenFeb 19, 2016
  35. David TurnerFeb 19, 2016
  36. Duy NguyenFeb 20, 2016
  37. David TurnerFeb 24, 2016
  38. Junio C HamanoFeb 20, 2016
  39. David TurnerFeb 19, 2016
  40. Junio C HamanoFeb 19, 2016
  41. Duy NguyenFeb 20, 2016
  42. David TurnerFeb 24, 2016
  43. Duy NguyenFeb 25, 2016
  44. Duy NguyenFeb 20, 2016
  45. David TurnerFeb 24, 2016
  46. Duy NguyenFeb 25, 2016
  47. refs: document transaction semanticsDavid Turner, Feb 25, 2016
  48. David TurnerFeb 25, 2016
  49. Junio C HamanoFeb 25, 2016
  50. David TurnerFeb 25, 2016
  51. 26/27 refs: tests for lmdb backendDavid Turner, Feb 18, 2016
  52. 27/27 tests: add ref-storage argumentDavid Turner, Feb 18, 2016

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.