tdb: fix short write logic in tdb_new_database

Commit 207a213c/24fed55d purported to fix the problem of signals during
tdb_new_database (which could cause a spurious short write, hence a failure).
However, the code is wrong: newdb+written is not correct.

Fix this by introducing a general tdb_write_all() and using it here and in
the tracing code.

Cc: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

1 files changed, 1 insertions, 15 deletions

diff --git a/lib/tdb/common/open.c b/lib/tdb/common/open.c
index dfe780d21b..3ff6b17c03 100644
--- a/lib/tdb/common/open.c
+++ b/lib/tdb/common/open.c
@@ -83,22 +83,8 @@ static int tdb_new_database(struct tdb_context *tdb, int hash_size)
 	/* Don't endian-convert the magic food! */
 	memcpy(newdb->magic_food, TDB_MAGIC_FOOD, strlen(TDB_MAGIC_FOOD)+1);
 	/* we still have "ret == -1" here */
-	written = write(tdb->fd, newdb, size);
-	if (written == size) {
+	if (tdb_write_all(tdb->fd, newdb, size))
 		ret = 0;
-	} else if (written != -1) {
-		/* call write once again, this usually should return -1 and
-		 * set errno appropriately */
-		size -= written;
-		written = write(tdb->fd, newdb+written, size);
-		if (written == size) {
-			ret = 0;
-		} else if (written >= 0) {
-			/* a second incomplete write - we give up.
-			 * guessing the errno... */
-			errno = ENOSPC;
-		}
-	}
 
   fail:
 	SAFE_FREE(newdb);