]> git.proxmox.com Git - mirror_qemu.git/blobdiff - block/nbd-client.c
nbd/client: Send NBD_CMD_DISC if open fails after connect
[mirror_qemu.git] / block / nbd-client.c
index c0683c3c8309e40d9ff5dc00cc1621d8c9b84c37..fc5b7eda8ee49ef7a0dc1dbe34d26b4731af9f0c 100644 (file)
@@ -78,7 +78,7 @@ static coroutine_fn void nbd_read_reply_entry(void *opaque)
     while (!s->quit) {
         assert(s->reply.handle == 0);
         ret = nbd_receive_reply(s->ioc, &s->reply, &local_err);
-        if (ret < 0) {
+        if (local_err) {
             error_report_err(local_err);
         }
         if (ret <= 0) {
@@ -92,7 +92,9 @@ static coroutine_fn void nbd_read_reply_entry(void *opaque)
         i = HANDLE_TO_INDEX(s, s->reply.handle);
         if (i >= MAX_NBD_REQUESTS ||
             !s->requests[i].coroutine ||
-            !s->requests[i].receiving) {
+            !s->requests[i].receiving ||
+            (nbd_reply_is_structured(&s->reply) && !s->info.structured_reply))
+        {
             break;
         }
 
@@ -139,6 +141,7 @@ static int nbd_co_send_request(BlockDriverState *bs,
     assert(i < MAX_NBD_REQUESTS);
 
     s->requests[i].coroutine = qemu_coroutine_self();
+    s->requests[i].offset = request->from;
     s->requests[i].receiving = false;
 
     request->handle = INDEX_TO_HANDLE(s, i);
@@ -179,75 +182,606 @@ err:
     return rc;
 }
 
-static int nbd_co_receive_reply(NBDClientSession *s,
-                                uint64_t handle,
-                                QEMUIOVector *qiov)
+static inline uint16_t payload_advance16(uint8_t **payload)
+{
+    *payload += 2;
+    return lduw_be_p(*payload - 2);
+}
+
+static inline uint32_t payload_advance32(uint8_t **payload)
+{
+    *payload += 4;
+    return ldl_be_p(*payload - 4);
+}
+
+static inline uint64_t payload_advance64(uint8_t **payload)
+{
+    *payload += 8;
+    return ldq_be_p(*payload - 8);
+}
+
+static int nbd_parse_offset_hole_payload(NBDStructuredReplyChunk *chunk,
+                                         uint8_t *payload, uint64_t orig_offset,
+                                         QEMUIOVector *qiov, Error **errp)
+{
+    uint64_t offset;
+    uint32_t hole_size;
+
+    if (chunk->length != sizeof(offset) + sizeof(hole_size)) {
+        error_setg(errp, "Protocol error: invalid payload for "
+                         "NBD_REPLY_TYPE_OFFSET_HOLE");
+        return -EINVAL;
+    }
+
+    offset = payload_advance64(&payload);
+    hole_size = payload_advance32(&payload);
+
+    if (!hole_size || offset < orig_offset || hole_size > qiov->size ||
+        offset > orig_offset + qiov->size - hole_size) {
+        error_setg(errp, "Protocol error: server sent chunk exceeding requested"
+                         " region");
+        return -EINVAL;
+    }
+
+    qemu_iovec_memset(qiov, offset - orig_offset, 0, hole_size);
+
+    return 0;
+}
+
+/* nbd_parse_blockstatus_payload
+ * support only one extent in reply and only for
+ * base:allocation context
+ */
+static int nbd_parse_blockstatus_payload(NBDClientSession *client,
+                                         NBDStructuredReplyChunk *chunk,
+                                         uint8_t *payload, uint64_t orig_length,
+                                         NBDExtent *extent, Error **errp)
+{
+    uint32_t context_id;
+
+    if (chunk->length != sizeof(context_id) + sizeof(*extent)) {
+        error_setg(errp, "Protocol error: invalid payload for "
+                         "NBD_REPLY_TYPE_BLOCK_STATUS");
+        return -EINVAL;
+    }
+
+    context_id = payload_advance32(&payload);
+    if (client->info.meta_base_allocation_id != context_id) {
+        error_setg(errp, "Protocol error: unexpected context id %d for "
+                         "NBD_REPLY_TYPE_BLOCK_STATUS, when negotiated context "
+                         "id is %d", context_id,
+                         client->info.meta_base_allocation_id);
+        return -EINVAL;
+    }
+
+    extent->length = payload_advance32(&payload);
+    extent->flags = payload_advance32(&payload);
+
+    if (extent->length == 0 ||
+        (client->info.min_block && !QEMU_IS_ALIGNED(extent->length,
+                                                    client->info.min_block))) {
+        error_setg(errp, "Protocol error: server sent status chunk with "
+                   "invalid length");
+        return -EINVAL;
+    }
+
+    /* The server is allowed to send us extra information on the final
+     * extent; just clamp it to the length we requested. */
+    if (extent->length > orig_length) {
+        extent->length = orig_length;
+    }
+
+    return 0;
+}
+
+/* nbd_parse_error_payload
+ * on success @errp contains message describing nbd error reply
+ */
+static int nbd_parse_error_payload(NBDStructuredReplyChunk *chunk,
+                                   uint8_t *payload, int *request_ret,
+                                   Error **errp)
+{
+    uint32_t error;
+    uint16_t message_size;
+
+    assert(chunk->type & (1 << 15));
+
+    if (chunk->length < sizeof(error) + sizeof(message_size)) {
+        error_setg(errp,
+                   "Protocol error: invalid payload for structured error");
+        return -EINVAL;
+    }
+
+    error = nbd_errno_to_system_errno(payload_advance32(&payload));
+    if (error == 0) {
+        error_setg(errp, "Protocol error: server sent structured error chunk "
+                         "with error = 0");
+        return -EINVAL;
+    }
+
+    *request_ret = -error;
+    message_size = payload_advance16(&payload);
+
+    if (message_size > chunk->length - sizeof(error) - sizeof(message_size)) {
+        error_setg(errp, "Protocol error: server sent structured error chunk "
+                         "with incorrect message size");
+        return -EINVAL;
+    }
+
+    /* TODO: Add a trace point to mention the server complaint */
+
+    /* TODO handle ERROR_OFFSET */
+
+    return 0;
+}
+
+static int nbd_co_receive_offset_data_payload(NBDClientSession *s,
+                                              uint64_t orig_offset,
+                                              QEMUIOVector *qiov, Error **errp)
+{
+    QEMUIOVector sub_qiov;
+    uint64_t offset;
+    size_t data_size;
+    int ret;
+    NBDStructuredReplyChunk *chunk = &s->reply.structured;
+
+    assert(nbd_reply_is_structured(&s->reply));
+
+    /* The NBD spec requires at least one byte of payload */
+    if (chunk->length <= sizeof(offset)) {
+        error_setg(errp, "Protocol error: invalid payload for "
+                         "NBD_REPLY_TYPE_OFFSET_DATA");
+        return -EINVAL;
+    }
+
+    if (nbd_read(s->ioc, &offset, sizeof(offset), errp) < 0) {
+        return -EIO;
+    }
+    be64_to_cpus(&offset);
+
+    data_size = chunk->length - sizeof(offset);
+    assert(data_size);
+    if (offset < orig_offset || data_size > qiov->size ||
+        offset > orig_offset + qiov->size - data_size) {
+        error_setg(errp, "Protocol error: server sent chunk exceeding requested"
+                         " region");
+        return -EINVAL;
+    }
+
+    qemu_iovec_init(&sub_qiov, qiov->niov);
+    qemu_iovec_concat(&sub_qiov, qiov, offset - orig_offset, data_size);
+    ret = qio_channel_readv_all(s->ioc, sub_qiov.iov, sub_qiov.niov, errp);
+    qemu_iovec_destroy(&sub_qiov);
+
+    return ret < 0 ? -EIO : 0;
+}
+
+#define NBD_MAX_MALLOC_PAYLOAD 1000
+/* nbd_co_receive_structured_payload
+ */
+static coroutine_fn int nbd_co_receive_structured_payload(
+        NBDClientSession *s, void **payload, Error **errp)
+{
+    int ret;
+    uint32_t len;
+
+    assert(nbd_reply_is_structured(&s->reply));
+
+    len = s->reply.structured.length;
+
+    if (len == 0) {
+        return 0;
+    }
+
+    if (payload == NULL) {
+        error_setg(errp, "Unexpected structured payload");
+        return -EINVAL;
+    }
+
+    if (len > NBD_MAX_MALLOC_PAYLOAD) {
+        error_setg(errp, "Payload too large");
+        return -EINVAL;
+    }
+
+    *payload = g_new(char, len);
+    ret = nbd_read(s->ioc, *payload, len, errp);
+    if (ret < 0) {
+        g_free(*payload);
+        *payload = NULL;
+        return ret;
+    }
+
+    return 0;
+}
+
+/* nbd_co_do_receive_one_chunk
+ * for simple reply:
+ *   set request_ret to received reply error
+ *   if qiov is not NULL: read payload to @qiov
+ * for structured reply chunk:
+ *   if error chunk: read payload, set @request_ret, do not set @payload
+ *   else if offset_data chunk: read payload data to @qiov, do not set @payload
+ *   else: read payload to @payload
+ *
+ * If function fails, @errp contains corresponding error message, and the
+ * connection with the server is suspect.  If it returns 0, then the
+ * transaction succeeded (although @request_ret may be a negative errno
+ * corresponding to the server's error reply), and errp is unchanged.
+ */
+static coroutine_fn int nbd_co_do_receive_one_chunk(
+        NBDClientSession *s, uint64_t handle, bool only_structured,
+        int *request_ret, QEMUIOVector *qiov, void **payload, Error **errp)
 {
     int ret;
     int i = HANDLE_TO_INDEX(s, handle);
+    void *local_payload = NULL;
+    NBDStructuredReplyChunk *chunk;
+
+    if (payload) {
+        *payload = NULL;
+    }
+    *request_ret = 0;
 
     /* Wait until we're woken up by nbd_read_reply_entry.  */
     s->requests[i].receiving = true;
     qemu_coroutine_yield();
     s->requests[i].receiving = false;
     if (!s->ioc || s->quit) {
-        ret = -EIO;
-    } else {
-        assert(s->reply.handle == handle);
-        ret = -s->reply.error;
-        if (qiov && s->reply.error == 0) {
-            if (qio_channel_readv_all(s->ioc, qiov->iov, qiov->niov,
-                                      NULL) < 0) {
-                ret = -EIO;
-                s->quit = true;
-            }
+        error_setg(errp, "Connection closed");
+        return -EIO;
+    }
+
+    assert(s->reply.handle == handle);
+
+    if (nbd_reply_is_simple(&s->reply)) {
+        if (only_structured) {
+            error_setg(errp, "Protocol error: simple reply when structured "
+                             "reply chunk was expected");
+            return -EINVAL;
         }
 
-        /* Tell the read handler to read another header.  */
-        s->reply.handle = 0;
+        *request_ret = -nbd_errno_to_system_errno(s->reply.simple.error);
+        if (*request_ret < 0 || !qiov) {
+            return 0;
+        }
+
+        return qio_channel_readv_all(s->ioc, qiov->iov, qiov->niov,
+                                     errp) < 0 ? -EIO : 0;
+    }
+
+    /* handle structured reply chunk */
+    assert(s->info.structured_reply);
+    chunk = &s->reply.structured;
+
+    if (chunk->type == NBD_REPLY_TYPE_NONE) {
+        if (!(chunk->flags & NBD_REPLY_FLAG_DONE)) {
+            error_setg(errp, "Protocol error: NBD_REPLY_TYPE_NONE chunk without"
+                       " NBD_REPLY_FLAG_DONE flag set");
+            return -EINVAL;
+        }
+        if (chunk->length) {
+            error_setg(errp, "Protocol error: NBD_REPLY_TYPE_NONE chunk with"
+                       " nonzero length");
+            return -EINVAL;
+        }
+        return 0;
     }
 
-    s->requests[i].coroutine = NULL;
+    if (chunk->type == NBD_REPLY_TYPE_OFFSET_DATA) {
+        if (!qiov) {
+            error_setg(errp, "Unexpected NBD_REPLY_TYPE_OFFSET_DATA chunk");
+            return -EINVAL;
+        }
+
+        return nbd_co_receive_offset_data_payload(s, s->requests[i].offset,
+                                                  qiov, errp);
+    }
+
+    if (nbd_reply_type_is_error(chunk->type)) {
+        payload = &local_payload;
+    }
+
+    ret = nbd_co_receive_structured_payload(s, payload, errp);
+    if (ret < 0) {
+        return ret;
+    }
+
+    if (nbd_reply_type_is_error(chunk->type)) {
+        ret = nbd_parse_error_payload(chunk, local_payload, request_ret, errp);
+        g_free(local_payload);
+        return ret;
+    }
+
+    return 0;
+}
+
+/* nbd_co_receive_one_chunk
+ * Read reply, wake up read_reply_co and set s->quit if needed.
+ * Return value is a fatal error code or normal nbd reply error code
+ */
+static coroutine_fn int nbd_co_receive_one_chunk(
+        NBDClientSession *s, uint64_t handle, bool only_structured,
+        QEMUIOVector *qiov, NBDReply *reply, void **payload, Error **errp)
+{
+    int request_ret;
+    int ret = nbd_co_do_receive_one_chunk(s, handle, only_structured,
+                                          &request_ret, qiov, payload, errp);
+
+    if (ret < 0) {
+        s->quit = true;
+    } else {
+        /* For assert at loop start in nbd_read_reply_entry */
+        if (reply) {
+            *reply = s->reply;
+        }
+        s->reply.handle = 0;
+        ret = request_ret;
+    }
 
-    /* Kick the read_reply_co to get the next reply.  */
     if (s->read_reply_co) {
         aio_co_wake(s->read_reply_co);
     }
 
+    return ret;
+}
+
+typedef struct NBDReplyChunkIter {
+    int ret;
+    bool fatal;
+    Error *err;
+    bool done, only_structured;
+} NBDReplyChunkIter;
+
+static void nbd_iter_error(NBDReplyChunkIter *iter, bool fatal,
+                           int ret, Error **local_err)
+{
+    assert(ret < 0);
+
+    if ((fatal && !iter->fatal) || iter->ret == 0) {
+        if (iter->ret != 0) {
+            error_free(iter->err);
+            iter->err = NULL;
+        }
+        iter->fatal = fatal;
+        iter->ret = ret;
+        error_propagate(&iter->err, *local_err);
+    } else {
+        error_free(*local_err);
+    }
+
+    *local_err = NULL;
+}
+
+/* NBD_FOREACH_REPLY_CHUNK
+ */
+#define NBD_FOREACH_REPLY_CHUNK(s, iter, handle, structured, \
+                                qiov, reply, payload) \
+    for (iter = (NBDReplyChunkIter) { .only_structured = structured }; \
+         nbd_reply_chunk_iter_receive(s, &iter, handle, qiov, reply, payload);)
+
+/* nbd_reply_chunk_iter_receive
+ */
+static bool nbd_reply_chunk_iter_receive(NBDClientSession *s,
+                                         NBDReplyChunkIter *iter,
+                                         uint64_t handle,
+                                         QEMUIOVector *qiov, NBDReply *reply,
+                                         void **payload)
+{
+    int ret;
+    NBDReply local_reply;
+    NBDStructuredReplyChunk *chunk;
+    Error *local_err = NULL;
+    if (s->quit) {
+        error_setg(&local_err, "Connection closed");
+        nbd_iter_error(iter, true, -EIO, &local_err);
+        goto break_loop;
+    }
+
+    if (iter->done) {
+        /* Previous iteration was last. */
+        goto break_loop;
+    }
+
+    if (reply == NULL) {
+        reply = &local_reply;
+    }
+
+    ret = nbd_co_receive_one_chunk(s, handle, iter->only_structured,
+                                   qiov, reply, payload, &local_err);
+    if (ret < 0) {
+        /* If it is a fatal error s->quit is set by nbd_co_receive_one_chunk */
+        nbd_iter_error(iter, s->quit, ret, &local_err);
+    }
+
+    /* Do not execute the body of NBD_FOREACH_REPLY_CHUNK for simple reply. */
+    if (nbd_reply_is_simple(&s->reply) || s->quit) {
+        goto break_loop;
+    }
+
+    chunk = &reply->structured;
+    iter->only_structured = true;
+
+    if (chunk->type == NBD_REPLY_TYPE_NONE) {
+        /* NBD_REPLY_FLAG_DONE is already checked in nbd_co_receive_one_chunk */
+        assert(chunk->flags & NBD_REPLY_FLAG_DONE);
+        goto break_loop;
+    }
+
+    if (chunk->flags & NBD_REPLY_FLAG_DONE) {
+        /* This iteration is last. */
+        iter->done = true;
+    }
+
+    /* Execute the loop body */
+    return true;
+
+break_loop:
+    s->requests[HANDLE_TO_INDEX(s, handle)].coroutine = NULL;
+
     qemu_co_mutex_lock(&s->send_mutex);
     s->in_flight--;
     qemu_co_queue_next(&s->free_sema);
     qemu_co_mutex_unlock(&s->send_mutex);
 
-    return ret;
+    return false;
 }
 
-static int nbd_co_request(BlockDriverState *bs,
-                          NBDRequest *request,
-                          QEMUIOVector *qiov)
+static int nbd_co_receive_return_code(NBDClientSession *s, uint64_t handle,
+                                      Error **errp)
+{
+    NBDReplyChunkIter iter;
+
+    NBD_FOREACH_REPLY_CHUNK(s, iter, handle, false, NULL, NULL, NULL) {
+        /* nbd_reply_chunk_iter_receive does all the work */
+    }
+
+    error_propagate(errp, iter.err);
+    return iter.ret;
+}
+
+static int nbd_co_receive_cmdread_reply(NBDClientSession *s, uint64_t handle,
+                                        uint64_t offset, QEMUIOVector *qiov,
+                                        Error **errp)
+{
+    NBDReplyChunkIter iter;
+    NBDReply reply;
+    void *payload = NULL;
+    Error *local_err = NULL;
+
+    NBD_FOREACH_REPLY_CHUNK(s, iter, handle, s->info.structured_reply,
+                            qiov, &reply, &payload)
+    {
+        int ret;
+        NBDStructuredReplyChunk *chunk = &reply.structured;
+
+        assert(nbd_reply_is_structured(&reply));
+
+        switch (chunk->type) {
+        case NBD_REPLY_TYPE_OFFSET_DATA:
+            /* special cased in nbd_co_receive_one_chunk, data is already
+             * in qiov */
+            break;
+        case NBD_REPLY_TYPE_OFFSET_HOLE:
+            ret = nbd_parse_offset_hole_payload(&reply.structured, payload,
+                                                offset, qiov, &local_err);
+            if (ret < 0) {
+                s->quit = true;
+                nbd_iter_error(&iter, true, ret, &local_err);
+            }
+            break;
+        default:
+            if (!nbd_reply_type_is_error(chunk->type)) {
+                /* not allowed reply type */
+                s->quit = true;
+                error_setg(&local_err,
+                           "Unexpected reply type: %d (%s) for CMD_READ",
+                           chunk->type, nbd_reply_type_lookup(chunk->type));
+                nbd_iter_error(&iter, true, -EINVAL, &local_err);
+            }
+        }
+
+        g_free(payload);
+        payload = NULL;
+    }
+
+    error_propagate(errp, iter.err);
+    return iter.ret;
+}
+
+static int nbd_co_receive_blockstatus_reply(NBDClientSession *s,
+                                            uint64_t handle, uint64_t length,
+                                            NBDExtent *extent, Error **errp)
+{
+    NBDReplyChunkIter iter;
+    NBDReply reply;
+    void *payload = NULL;
+    Error *local_err = NULL;
+    bool received = false;
+
+    assert(!extent->length);
+    NBD_FOREACH_REPLY_CHUNK(s, iter, handle, s->info.structured_reply,
+                            NULL, &reply, &payload)
+    {
+        int ret;
+        NBDStructuredReplyChunk *chunk = &reply.structured;
+
+        assert(nbd_reply_is_structured(&reply));
+
+        switch (chunk->type) {
+        case NBD_REPLY_TYPE_BLOCK_STATUS:
+            if (received) {
+                s->quit = true;
+                error_setg(&local_err, "Several BLOCK_STATUS chunks in reply");
+                nbd_iter_error(&iter, true, -EINVAL, &local_err);
+            }
+            received = true;
+
+            ret = nbd_parse_blockstatus_payload(s, &reply.structured,
+                                                payload, length, extent,
+                                                &local_err);
+            if (ret < 0) {
+                s->quit = true;
+                nbd_iter_error(&iter, true, ret, &local_err);
+            }
+            break;
+        default:
+            if (!nbd_reply_type_is_error(chunk->type)) {
+                s->quit = true;
+                error_setg(&local_err,
+                           "Unexpected reply type: %d (%s) "
+                           "for CMD_BLOCK_STATUS",
+                           chunk->type, nbd_reply_type_lookup(chunk->type));
+                nbd_iter_error(&iter, true, -EINVAL, &local_err);
+            }
+        }
+
+        g_free(payload);
+        payload = NULL;
+    }
+
+    if (!extent->length && !iter.err) {
+        error_setg(&iter.err,
+                   "Server did not reply with any status extents");
+        if (!iter.ret) {
+            iter.ret = -EIO;
+        }
+    }
+    error_propagate(errp, iter.err);
+    return iter.ret;
+}
+
+static int nbd_co_request(BlockDriverState *bs, NBDRequest *request,
+                          QEMUIOVector *write_qiov)
 {
-    NBDClientSession *client = nbd_get_client_session(bs);
     int ret;
+    Error *local_err = NULL;
+    NBDClientSession *client = nbd_get_client_session(bs);
 
-    if (qiov) {
-        assert(request->type == NBD_CMD_WRITE || request->type == NBD_CMD_READ);
-        assert(request->len == iov_size(qiov->iov, qiov->niov));
+    assert(request->type != NBD_CMD_READ);
+    if (write_qiov) {
+        assert(request->type == NBD_CMD_WRITE);
+        assert(request->len == iov_size(write_qiov->iov, write_qiov->niov));
     } else {
-        assert(request->type != NBD_CMD_WRITE && request->type != NBD_CMD_READ);
+        assert(request->type != NBD_CMD_WRITE);
     }
-    ret = nbd_co_send_request(bs, request,
-                              request->type == NBD_CMD_WRITE ? qiov : NULL);
+    ret = nbd_co_send_request(bs, request, write_qiov);
     if (ret < 0) {
         return ret;
     }
 
-    return nbd_co_receive_reply(client, request->handle,
-                                request->type == NBD_CMD_READ ? qiov : NULL);
+    ret = nbd_co_receive_return_code(client, request->handle, &local_err);
+    if (local_err) {
+        error_report_err(local_err);
+    }
+    return ret;
 }
 
 int nbd_client_co_preadv(BlockDriverState *bs, uint64_t offset,
                          uint64_t bytes, QEMUIOVector *qiov, int flags)
 {
+    int ret;
+    Error *local_err = NULL;
+    NBDClientSession *client = nbd_get_client_session(bs);
     NBDRequest request = {
         .type = NBD_CMD_READ,
         .from = offset,
@@ -257,7 +791,20 @@ int nbd_client_co_preadv(BlockDriverState *bs, uint64_t offset,
     assert(bytes <= NBD_MAX_BUFFER_SIZE);
     assert(!flags);
 
-    return nbd_co_request(bs, &request, qiov);
+    if (!bytes) {
+        return 0;
+    }
+    ret = nbd_co_send_request(bs, &request, NULL);
+    if (ret < 0) {
+        return ret;
+    }
+
+    ret = nbd_co_receive_cmdread_reply(client, request.handle, offset, qiov,
+                                       &local_err);
+    if (local_err) {
+        error_report_err(local_err);
+    }
+    return ret;
 }
 
 int nbd_client_co_pwritev(BlockDriverState *bs, uint64_t offset,
@@ -270,6 +817,7 @@ int nbd_client_co_pwritev(BlockDriverState *bs, uint64_t offset,
         .len = bytes,
     };
 
+    assert(!(client->info.flags & NBD_FLAG_READ_ONLY));
     if (flags & BDRV_REQ_FUA) {
         assert(client->info.flags & NBD_FLAG_SEND_FUA);
         request.flags |= NBD_CMD_FLAG_FUA;
@@ -277,6 +825,9 @@ int nbd_client_co_pwritev(BlockDriverState *bs, uint64_t offset,
 
     assert(bytes <= NBD_MAX_BUFFER_SIZE);
 
+    if (!bytes) {
+        return 0;
+    }
     return nbd_co_request(bs, &request, qiov);
 }
 
@@ -290,6 +841,7 @@ int nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
         .len = bytes,
     };
 
+    assert(!(client->info.flags & NBD_FLAG_READ_ONLY));
     if (!(client->info.flags & NBD_FLAG_SEND_WRITE_ZEROES)) {
         return -ENOTSUP;
     }
@@ -302,6 +854,9 @@ int nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset,
         request.flags |= NBD_CMD_FLAG_NO_HOLE;
     }
 
+    if (!bytes) {
+        return 0;
+    }
     return nbd_co_request(bs, &request, NULL);
 }
 
@@ -329,13 +884,59 @@ int nbd_client_co_pdiscard(BlockDriverState *bs, int64_t offset, int bytes)
         .len = bytes,
     };
 
-    if (!(client->info.flags & NBD_FLAG_SEND_TRIM)) {
+    assert(!(client->info.flags & NBD_FLAG_READ_ONLY));
+    if (!(client->info.flags & NBD_FLAG_SEND_TRIM) || !bytes) {
         return 0;
     }
 
     return nbd_co_request(bs, &request, NULL);
 }
 
+int coroutine_fn nbd_client_co_block_status(BlockDriverState *bs,
+                                            bool want_zero,
+                                            int64_t offset, int64_t bytes,
+                                            int64_t *pnum, int64_t *map,
+                                            BlockDriverState **file)
+{
+    int64_t ret;
+    NBDExtent extent = { 0 };
+    NBDClientSession *client = nbd_get_client_session(bs);
+    Error *local_err = NULL;
+
+    NBDRequest request = {
+        .type = NBD_CMD_BLOCK_STATUS,
+        .from = offset,
+        .len = MIN(MIN_NON_ZERO(QEMU_ALIGN_DOWN(INT_MAX,
+                                                bs->bl.request_alignment),
+                                client->info.max_block), bytes),
+        .flags = NBD_CMD_FLAG_REQ_ONE,
+    };
+
+    if (!client->info.base_allocation) {
+        *pnum = bytes;
+        return BDRV_BLOCK_DATA;
+    }
+
+    ret = nbd_co_send_request(bs, &request, NULL);
+    if (ret < 0) {
+        return ret;
+    }
+
+    ret = nbd_co_receive_blockstatus_reply(client, request.handle, bytes,
+                                           &extent, &local_err);
+    if (local_err) {
+        error_report_err(local_err);
+    }
+    if (ret < 0) {
+        return ret;
+    }
+
+    assert(extent.length);
+    *pnum = extent.length;
+    return (extent.flags & NBD_STATE_HOLE ? 0 : BDRV_BLOCK_DATA) |
+           (extent.flags & NBD_STATE_ZERO ? BDRV_BLOCK_ZERO : 0);
+}
+
 void nbd_client_detach_aio_context(BlockDriverState *bs)
 {
     NBDClientSession *client = nbd_get_client_session(bs);
@@ -369,6 +970,7 @@ int nbd_client_init(BlockDriverState *bs,
                     const char *export,
                     QCryptoTLSCreds *tlscreds,
                     const char *hostname,
+                    const char *x_dirty_bitmap,
                     Error **errp)
 {
     NBDClientSession *client = nbd_get_client_session(bs);
@@ -379,13 +981,29 @@ int nbd_client_init(BlockDriverState *bs,
     qio_channel_set_blocking(QIO_CHANNEL(sioc), true, NULL);
 
     client->info.request_sizes = true;
+    client->info.structured_reply = true;
+    client->info.base_allocation = true;
+    client->info.x_dirty_bitmap = g_strdup(x_dirty_bitmap);
     ret = nbd_receive_negotiate(QIO_CHANNEL(sioc), export,
                                 tlscreds, hostname,
                                 &client->ioc, &client->info, errp);
+    g_free(client->info.x_dirty_bitmap);
     if (ret < 0) {
         logout("Failed to negotiate with the NBD server\n");
         return ret;
     }
+    if (x_dirty_bitmap && !client->info.base_allocation) {
+        error_setg(errp, "requested x-dirty-bitmap %s not found",
+                   x_dirty_bitmap);
+        ret = -EINVAL;
+        goto fail;
+    }
+    if (client->info.flags & NBD_FLAG_READ_ONLY) {
+        ret = bdrv_apply_auto_read_only(bs, "NBD export is read-only", errp);
+        if (ret < 0) {
+            goto fail;
+        }
+    }
     if (client->info.flags & NBD_FLAG_SEND_FUA) {
         bs->supported_write_flags = BDRV_REQ_FUA;
         bs->supported_zero_flags |= BDRV_REQ_FUA;
@@ -393,9 +1011,6 @@ int nbd_client_init(BlockDriverState *bs,
     if (client->info.flags & NBD_FLAG_SEND_WRITE_ZEROES) {
         bs->supported_zero_flags |= BDRV_REQ_MAY_UNMAP;
     }
-    if (client->info.min_block > bs->bl.request_alignment) {
-        bs->bl.request_alignment = client->info.min_block;
-    }
 
     qemu_co_mutex_init(&client->send_mutex);
     qemu_co_queue_init(&client->free_sema);
@@ -415,4 +1030,17 @@ int nbd_client_init(BlockDriverState *bs,
 
     logout("Established connection with NBD server\n");
     return 0;
+
+ fail:
+    /*
+     * We have connected, but must fail for other reasons. The
+     * connection is still blocking; send NBD_CMD_DISC as a courtesy
+     * to the server.
+     */
+    {
+        NBDRequest request = { .type = NBD_CMD_DISC };
+
+        nbd_send_request(client->ioc ?: QIO_CHANNEL(sioc), &request);
+        return ret;
+    }
 }