tdb2: catch errors in tdb1_needs_recovery()

The tdb1 backend simply returns "true" if there's an error determining
if a tdb needs recovery.  But this leads failtest down a rabbit hole;
it's better to return the error at this case (and makes for better for
diagnostics, since they will come from the first fault, not later in
tdb1_transaction_recover().

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
(Imported from CCAN commit 332d0c29baa6896e67c439aeb47f58a104fbc781)

3 files changed, 23 insertions, 9 deletions

diff --git a/lib/tdb2/tdb1_lock.c b/lib/tdb2/tdb1_lock.c
index f21a22cacd..495e1d1bf6 100644
--- a/lib/tdb2/tdb1_lock.c
+++ b/lib/tdb2/tdb1_lock.c
@@ -162,13 +162,21 @@ static int tdb1_lock_list(struct tdb_context *tdb, int list, int ltype,
 		check = !have_data_locks(tdb);
 		ret = tdb1_nest_lock(tdb, lock_offset(list), ltype, waitflag);
 
-		if (ret == 0 && check && tdb1_needs_recovery(tdb)) {
-			tdb1_nest_unlock(tdb, lock_offset(list), ltype);
+		if (ret == 0 && check) {
+			tdb_bool_err berr = tdb1_needs_recovery(tdb);
 
-			if (tdb1_lock_and_recover(tdb) == -1) {
+			if (berr < 0) {
 				return -1;
 			}
-			return tdb1_lock_list(tdb, list, ltype, waitflag);
+			if (berr == true) {
+				tdb1_nest_unlock(tdb, lock_offset(list), ltype);
+
+				if (tdb1_lock_and_recover(tdb) == -1) {
+					return -1;
+				}
+				return tdb1_lock_list(tdb, list, ltype,
+						      waitflag);
+			}
 		}
 	}
 	return ret;
@@ -249,6 +257,7 @@ int tdb1_allrecord_lock(struct tdb_context *tdb, int ltype,
 		       enum tdb_lock_flags flags, bool upgradable)
 {
 	enum TDB_ERROR ecode;
+	tdb_bool_err berr;
 
 	/* tdb_lock_gradual() doesn't know about tdb->tdb1.traverse_read. */
 	if (tdb->tdb1.traverse_read && !(tdb->flags & TDB_NOLOCK)) {
@@ -313,7 +322,12 @@ int tdb1_allrecord_lock(struct tdb_context *tdb, int ltype,
 	tdb->file->allrecord_lock.ltype = upgradable ? F_WRLCK : ltype;
 	tdb->file->allrecord_lock.off = upgradable;
 
-	if (tdb1_needs_recovery(tdb)) {
+	berr = tdb1_needs_recovery(tdb);
+	if (berr < 0) {
+		return -1;
+	}
+
+	if (berr == true) {
 		tdb1_allrecord_unlock(tdb, ltype);
 		if (tdb1_lock_and_recover(tdb) == -1) {
 			return -1;
diff --git a/lib/tdb2/tdb1_private.h b/lib/tdb2/tdb1_private.h
index ea147a8967..68dc39f6f8 100644
--- a/lib/tdb2/tdb1_private.h
+++ b/lib/tdb2/tdb1_private.h
@@ -153,7 +153,7 @@ int tdb1_ofs_read(struct tdb_context *tdb, tdb1_off_t offset, tdb1_off_t *d);
 int tdb1_ofs_write(struct tdb_context *tdb, tdb1_off_t offset, tdb1_off_t *d);
 int tdb1_lock_record(struct tdb_context *tdb, tdb1_off_t off);
 int tdb1_unlock_record(struct tdb_context *tdb, tdb1_off_t off);
-bool tdb1_needs_recovery(struct tdb_context *tdb);
+tdb_bool_err tdb1_needs_recovery(struct tdb_context *tdb);
 int tdb1_rec_read(struct tdb_context *tdb, tdb1_off_t offset, struct tdb1_record *rec);
 int tdb1_rec_write(struct tdb_context *tdb, tdb1_off_t offset, struct tdb1_record *rec);
 int tdb1_do_delete(struct tdb_context *tdb, tdb1_off_t rec_ptr, struct tdb1_record *rec);
diff --git a/lib/tdb2/tdb1_transaction.c b/lib/tdb2/tdb1_transaction.c
index 08eac1df34..ecd7d2628e 100644
--- a/lib/tdb2/tdb1_transaction.c
+++ b/lib/tdb2/tdb1_transaction.c
@@ -1289,14 +1289,14 @@ int tdb1_transaction_recover(struct tdb_context *tdb)
 }
 
 /* Any I/O failures we say "needs recovery". */
-bool tdb1_needs_recovery(struct tdb_context *tdb)
+tdb_bool_err tdb1_needs_recovery(struct tdb_context *tdb)
 {
 	tdb1_off_t recovery_head;
 	struct tdb1_record rec;
 
 	/* find the recovery area */
 	if (tdb1_ofs_read(tdb, TDB1_RECOVERY_HEAD, &recovery_head) == -1) {
-		return true;
+		return tdb->last_error;
 	}
 
 	if (recovery_head == 0) {
@@ -1307,7 +1307,7 @@ bool tdb1_needs_recovery(struct tdb_context *tdb)
 	/* read the recovery record */
 	if (tdb->tdb1.io->tdb1_read(tdb, recovery_head, &rec,
 				   sizeof(rec), TDB1_DOCONV()) == -1) {
-		return true;
+		return tdb->last_error;
 	}
 
 	return (rec.magic == TDB1_RECOVERY_MAGIC);