If the write request spans more than one L2 table, qcow2_alloc_cluster_offset cannot handle the required allocations atomically. This results in leaks if it allocated new clusters in any but the last L2 table touched and an error occurs in qcow2_co_writev before having established the L2 link. These non-atomic allocations were, however, indeed successful and are therefore given to the caller in the L2Meta list.
If an error occurs in qcow2_co_writev and the L2Meta list is unwound, all its remaining entries are clusters whose L2 links were not yet established. Thus, all allocations in that list should be undone. Signed-off-by: Max Reitz <mre...@redhat.com> --- block/qcow2.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/block/qcow2.c b/block/qcow2.c index 13e34f0..69d8270 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -1078,6 +1078,14 @@ static coroutine_fn int qcow2_co_writev(BlockDriverState *bs, goto fail; } + /* WARNING: If adding a "goto fail" anywhere from here until the + * current QCowL2Meta element is removed from the l2meta list (that + * is, until the end of this loop), mind the qcow2_free_clusters() + * call after the fail label. Always make sure that every list + * element with a non-zero nb_clusters value remaining when reaching + * the while loop after that label indeed corresponds to a leaked + * allocation which has to be freed. */ + /* Take the request off the list of running requests */ if (l2meta->nb_clusters != 0) { QLIST_REMOVE(l2meta, next_in_flight); @@ -1103,6 +1111,13 @@ fail: while (l2meta != NULL) { QCowL2Meta *next; + /* Undo all leaked allocations */ + if (l2meta->nb_clusters != 0) { + qcow2_free_clusters(bs, l2meta->alloc_offset, + l2meta->nb_clusters << s->cluster_bits, + QCOW2_DISCARD_OTHER); + } + if (l2meta->nb_clusters != 0) { QLIST_REMOVE(l2meta, next_in_flight); } -- 1.8.3.1