[PATCH v2 11/17] lock_ref_sha1_basic(): if locking fails with ENOENT, retry

From: Michael Haggerty <mhagger@alum.mit.edu>
Date: Jan 6, 2014, 13:45 UTC
Message-ID: <1389015935-21936-12-git-send-email-mhagger@alum.mit.edu>
In-Reply-To: <1389015935-21936-1-git-send-email-mhagger@alum.mit.edu>

If hold_lock_file_for_update() fails with errno==ENOENT, it might be because somebody else (for example, a pack-refs process) has just deleted one of the lockfile's ancestor directories. So if this condition is detected, try again (up to 3 times).

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
---
 refs.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/refs.c b/refs.c
index 6eb8a02..8a3d3ea 100644
--- a/refs.c
+++ b/refs.c
@@ -2081,7 +2081,7 @@ static struct ref_lock *lock_ref_sha1_basic(const char *refname,
 
 	lock->lk = xcalloc(1, sizeof(struct lock_file));
 
-	lflags = LOCK_DIE_ON_ERROR;
+	lflags = 0;
 	if (flags & REF_NODEREF) {
 		refname = orig_refname;
 		lflags |= LOCK_NODEREF;
@@ -2109,6 +2109,17 @@ static struct ref_lock *lock_ref_sha1_basic(const char *refname,
 	}
 
 	lock->lock_fd = hold_lock_file_for_update(lock->lk, ref_file, lflags);
+	if (lock->lock_fd < 0) {
+		if (errno == ENOENT && --attempts > 0)
+			/*
+			 * Maybe somebody just deleted one of the
+			 * directories leading to ref_file.  Try
+			 * again:
+			 */
+			goto retry;
+		else
+			unable_to_lock_index_die(ref_file, errno);
+	}
 	return old_sha1 ? verify_lock(lock, old_sha1, mustexist) : lock;
 
  error_return:

-- 
1.8.5.2