qcow2: Ignore reserved bits in get_cluster_offset

With this change, reading from a qcow2 image ignores all reserved bits
that are set in an L1 or L2 table entry.

Now get_cluster_offset() assigns *cluster_offset only the offset without
any other flags. The cluster type is not longer encoded in the offset,
but a positive return value in case of success.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

3 files changed, 60 insertions, 19 deletions

diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index dcf70a2..c8c85d2 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -367,11 +367,9 @@ out:
  *
  * on exit, *num is the number of contiguous sectors we can read.
  *
- * Return 0, if the offset is found
- * Return -errno, otherwise.
- *
+ * Returns the cluster type (QCOW2_CLUSTER_*) on success, -errno in error
+ * cases.
  */
-
 int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
     int *num, uint64_t *cluster_offset)
 {
@@ -407,19 +405,19 @@ int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
     /* seek the the l2 offset in the l1 table */
 
     l1_index = offset >> l1_bits;
-    if (l1_index >= s->l1_size)
+    if (l1_index >= s->l1_size) {
+        ret = QCOW2_CLUSTER_UNALLOCATED;
         goto out;
+    }
 
-    l2_offset = s->l1_table[l1_index];
-
-    /* seek the l2 table of the given l2 offset */
-
-    if (!l2_offset)
+    l2_offset = s->l1_table[l1_index] & L1E_OFFSET_MASK;
+    if (!l2_offset) {
+        ret = QCOW2_CLUSTER_UNALLOCATED;
         goto out;
+    }
 
     /* load the l2 table in memory */
 
-    l2_offset &= ~QCOW_OFLAG_COPIED;
     ret = l2_load(bs, l2_offset, &l2_table);
     if (ret < 0) {
         return ret;
@@ -431,26 +429,37 @@ int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
     *cluster_offset = be64_to_cpu(l2_table[l2_index]);
     nb_clusters = size_to_clusters(s, nb_needed << 9);
 
-    if (!*cluster_offset) {
+    ret = qcow2_get_cluster_type(*cluster_offset);
+    switch (ret) {
+    case QCOW2_CLUSTER_COMPRESSED:
+        /* Compressed clusters can only be processed one by one */
+        c = 1;
+        *cluster_offset &= L2E_COMPRESSED_OFFSET_SIZE_MASK;
+        break;
+    case QCOW2_CLUSTER_UNALLOCATED:
         /* how many empty clusters ? */
         c = count_contiguous_free_clusters(nb_clusters, &l2_table[l2_index]);
-    } else {
+        *cluster_offset = 0;
+        break;
+    case QCOW2_CLUSTER_NORMAL:
         /* how many allocated clusters ? */
         c = count_contiguous_clusters(nb_clusters, s->cluster_size,
                 &l2_table[l2_index], 0, QCOW_OFLAG_COPIED);
+        *cluster_offset &= L2E_OFFSET_MASK;
+        break;
     }
 
     qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
 
-   nb_available = (c * s->cluster_sectors);
+    nb_available = (c * s->cluster_sectors);
+
 out:
     if (nb_available > nb_needed)
         nb_available = nb_needed;
 
     *num = nb_available - index_in_cluster;
 
-    *cluster_offset &=~QCOW_OFLAG_COPIED;
-    return 0;
+    return ret;
 }
 
 /*
diff --git a/block/qcow2.c b/block/qcow2.c
index 70d3141..4c82adc 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -449,7 +449,8 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
         qemu_iovec_copy(&hd_qiov, qiov, bytes_done,
             cur_nr_sectors * 512);
 
-        if (!cluster_offset) {
+        switch (ret) {
+        case QCOW2_CLUSTER_UNALLOCATED:
 
             if (bs->backing_hd) {
                 /* read from the base image */
@@ -469,7 +470,9 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
                 /* Note: in this case, no need to wait */
                 qemu_iovec_memset(&hd_qiov, 0, 512 * cur_nr_sectors);
             }
-        } else if (cluster_offset & QCOW_OFLAG_COMPRESSED) {
+            break;
+
+        case QCOW2_CLUSTER_COMPRESSED:
             /* add AIO support for compressed blocks ? */
             ret = qcow2_decompress_cluster(bs, cluster_offset);
             if (ret < 0) {
@@ -479,7 +482,9 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
             qemu_iovec_from_buffer(&hd_qiov,
                 s->cluster_cache + index_in_cluster * 512,
                 512 * cur_nr_sectors);
-        } else {
+            break;
+
+        case QCOW2_CLUSTER_NORMAL:
             if ((cluster_offset & 511) != 0) {
                 ret = -EIO;
                 goto fail;
@@ -520,6 +525,12 @@ static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
                 qemu_iovec_from_buffer(&hd_qiov, cluster_data,
                     512 * cur_nr_sectors);
             }
+            break;
+
+        default:
+            g_assert_not_reached();
+            ret = -EIO;
+            goto fail;
         }
 
         remaining_sectors -= cur_nr_sectors;
diff --git a/block/qcow2.h b/block/qcow2.h
index ddb976a..a22d7fa 100644
--- a/block/qcow2.h
+++ b/block/qcow2.h
@@ -165,6 +165,16 @@ typedef struct QCowL2Meta
     QLIST_ENTRY(QCowL2Meta) next_in_flight;
 } QCowL2Meta;
 
+enum {
+    QCOW2_CLUSTER_UNALLOCATED,
+    QCOW2_CLUSTER_NORMAL,
+    QCOW2_CLUSTER_COMPRESSED,
+};
+
+#define L1E_OFFSET_MASK 0x00ffffffffffff00ULL
+#define L2E_OFFSET_MASK 0x00ffffffffffff00ULL
+#define L2E_COMPRESSED_OFFSET_SIZE_MASK 0x3fffffffffffffffULL
+
 static inline int size_to_clusters(BDRVQcowState *s, int64_t size)
 {
     return (size + (s->cluster_size - 1)) >> s->cluster_bits;
@@ -182,6 +192,17 @@ static inline int64_t align_offset(int64_t offset, int n)
     return offset;
 }
 
+static inline int qcow2_get_cluster_type(uint64_t l2_entry)
+{
+    if (l2_entry & QCOW_OFLAG_COMPRESSED) {
+        return QCOW2_CLUSTER_COMPRESSED;
+    } else if (!(l2_entry & L2E_OFFSET_MASK)) {
+        return QCOW2_CLUSTER_UNALLOCATED;
+    } else {
+        return QCOW2_CLUSTER_NORMAL;
+    }
+}
+
 
 // FIXME Need qcow2_ prefix to global functions