@@ -94,6 +94,16 @@ typedef struct NBDStructuredError {
uint16_t message_length;
} QEMU_PACKED NBDStructuredError;
+typedef struct NBDStructuredMeta {
+ NBDStructuredReplyChunk h;
+ uint32_t context_id;
+} QEMU_PACKED NBDStructuredMeta;
+
+typedef struct NBDExtent {
+ uint32_t length;
+ uint32_t flags;
+} QEMU_PACKED NBDExtent;
+
/* Transmission (export) flags: sent from server to client during handshake,
but describe what will happen during transmission */
#define NBD_FLAG_HAS_FLAGS (1 << 0) /* Flags are there */
@@ -120,6 +130,7 @@ typedef struct NBDStructuredError {
#define NBD_REP_ACK (1) /* Data sending finished. */
#define NBD_REP_SERVER (2) /* Export description. */
+#define NBD_REP_META_CONTEXT (4)
#define NBD_REP_ERR_UNSUP NBD_REP_ERR(1) /* Unknown option */
#define NBD_REP_ERR_POLICY NBD_REP_ERR(2) /* Server denied */
@@ -127,6 +138,8 @@ typedef struct NBDStructuredError {
#define NBD_REP_ERR_PLATFORM NBD_REP_ERR(4) /* Not compiled in */
#define NBD_REP_ERR_TLS_REQD NBD_REP_ERR(5) /* TLS required */
#define NBD_REP_ERR_SHUTDOWN NBD_REP_ERR(7) /* Server shutting down */
+#define NBD_REP_ERR_TOO_BIG NBD_REP_ERR(9) /* The request or the reply is
+ too large to process */
/* Request flags, sent from client to server during transmission phase */
#define NBD_CMD_FLAG_FUA (1 << 0) /* 'force unit access' during write */
@@ -142,6 +155,7 @@ enum {
NBD_CMD_TRIM = 4,
/* 5 reserved for failed experiment NBD_CMD_CACHE */
NBD_CMD_WRITE_ZEROES = 6,
+ NBD_CMD_BLOCK_STATUS = 7
};
#define NBD_DEFAULT_PORT 10809
@@ -163,6 +177,7 @@ enum {
#define NBD_REPLY_TYPE_NONE 0
#define NBD_REPLY_TYPE_OFFSET_DATA 1
#define NBD_REPLY_TYPE_OFFSET_HOLE 2
+#define NBD_REPLY_TYPE_BLOCK_STATUS 5
#define NBD_REPLY_TYPE_ERROR ((1 << 15) + 1)
#define NBD_REPLY_TYPE_ERROR_OFFSET ((1 << 15) + 2)
@@ -83,6 +83,10 @@
#define NBD_OPT_PEEK_EXPORT (4)
#define NBD_OPT_STARTTLS (5)
#define NBD_OPT_STRUCTURED_REPLY (8)
+#define NBD_OPT_LIST_META_CONTEXT (9)
+#define NBD_OPT_SET_META_CONTEXT (10)
+
+#define NBD_META_NS_BITMAPS "qemu-dirty-bitmap"
/* NBD errors are based on errno numbers, so there is a 1:1 mapping,
* but only a limited set of errno values is specified in the protocol.
@@ -105,6 +109,8 @@ static inline const char *nbd_opt_name(int opt)
case NBD_OPT_PEEK_EXPORT: return "peek_export";
case NBD_OPT_STARTTLS: return "tls";
case NBD_OPT_STRUCTURED_REPLY: return "structured_reply";
+ case NBD_OPT_LIST_META_CONTEXT: return "list_meta_context";
+ case NBD_OPT_SET_META_CONTEXT: return "set_meta_context";
}
return "<unknown option>";
@@ -21,6 +21,8 @@
#include "qapi/error.h"
#include "nbd-internal.h"
+#define NBD_MAX_BITMAP_EXTENTS (0x100000 / 8) /* 1 mb of extents data */
+
static int system_errno_to_nbd_errno(int err)
{
switch (err) {
@@ -102,6 +104,7 @@ struct NBDClient {
bool closing;
bool structured_reply;
+ BdrvDirtyBitmap *export_bitmap;
};
/* That's all folks */
@@ -421,7 +424,304 @@ static QIOChannel *nbd_negotiate_handle_starttls(NBDClient *client,
return QIO_CHANNEL(tioc);
}
+static int nbd_negotiate_read_size_string(QIOChannel *ioc, char **str,
+ uint32_t max_len)
+{
+ uint32_t len;
+
+ if (nbd_negotiate_read(ioc, &len, sizeof(len)) != sizeof(len)) {
+ LOG("read failed");
+ return -EIO;
+ }
+
+ cpu_to_be32s(&len);
+
+ if (max_len > 0 && len > max_len) {
+ LOG("Bad length received");
+ return -EINVAL;
+ }
+
+ *str = g_malloc(len + 1);
+
+ if (nbd_negotiate_read(ioc, *str, len) != len) {
+ LOG("read failed");
+ g_free(str);
+ return -EIO;
+ }
+ (*str)[len] = '\0';
+
+ return sizeof(len) + len;
+}
+
+static int nbd_negotiate_send_meta_context(QIOChannel *ioc,
+ const char *context,
+ uint32_t opt)
+{
+ int ret;
+ size_t len = strlen(context);
+ uint32_t context_id = cpu_to_be32(100);
+
+ ret = nbd_negotiate_send_rep_len(ioc, NBD_REP_META_CONTEXT, opt,
+ len + sizeof(context_id));
+ if (ret < 0) {
+ return ret;
+ }
+
+ if (nbd_negotiate_write(ioc, &context_id, sizeof(context_id)) !=
+ sizeof(context_id))
+ {
+ LOG("write failed");
+ return -EIO;
+ }
+
+ if (nbd_negotiate_write(ioc, context, len) != len) {
+ LOG("write failed");
+ return -EIO;
+ }
+
+ return 0;
+}
+
+static int nbd_negotiate_send_bitmap(QIOChannel *ioc, const char *bitmap_name,
+ uint32_t opt)
+{
+ char *context = g_strdup_printf("%s:%s", NBD_META_NS_BITMAPS, bitmap_name);
+ int ret = nbd_negotiate_send_meta_context(ioc, context, opt);
+
+ g_free(context);
+
+ return ret;
+}
+
+static int nbd_negotiate_one_bitmap_query(QIOChannel *ioc, BlockDriverState *bs,
+ uint32_t opt, const char *query,
+ BdrvDirtyBitmap **bitmap)
+{
+ BdrvDirtyBitmap *bm = bdrv_find_dirty_bitmap(bs, query);
+ if (bm != NULL) {
+ if (bitmap != NULL) {
+ *bitmap = bm;
+ }
+ return nbd_negotiate_send_bitmap(ioc, query, opt);
+ }
+
+ return 0;
+}
+
+static int nbd_negotiate_one_meta_query(QIOChannel *ioc, BlockDriverState *bs,
+ uint32_t opt, BdrvDirtyBitmap **bitmap)
+{
+ int ret = 0, nb_read;
+ char *query, *colon, *namespace, *subquery;
+
+ *bitmap = NULL;
+ nb_read = nbd_negotiate_read_size_string(ioc, &query, 0);
+ if (nb_read < 0) {
+ return nb_read;
+ }
+
+ colon = strchr(query, ':');
+ if (colon == NULL) {
+ ret = -EINVAL;
+ goto out;
+ }
+ *colon = '\0';
+ namespace = query;
+ subquery = colon + 1;
+
+ if (strcmp(namespace, NBD_META_NS_BITMAPS) == 0) {
+ ret = nbd_negotiate_one_bitmap_query(ioc, bs, opt, subquery, bitmap);
+ }
+
+out:
+ g_free(query);
+ return ret < 0 ? ret : nb_read;
+}
+
+/* start handle LIST_META_CONTEXT and SET_META_CONTEXT requests
+ * @opt should be NBD_OPT_LIST_META_CONTEXT or NBD_OPT_SET_META_CONTEXT
+ * @length related option data to read
+ * @nb_queries out parameter, number of queries specified by client
+ * @bs out parameter, bs for export, selected by client
+ * will be zero if some not critical error occured and error reply
+ * was sent.
+ *
+ * Returns:
+ * Err. code < 0 on critical error
+ * Number of bytes read otherwise (will be equal to length on non critical
+ * error or if there no queries in request)
+ */
+static int nbd_negotiate_opt_meta_context_start(NBDClient *client, uint32_t opt,
+ uint32_t length,
+ uint32_t *nb_queries,
+ BlockDriverState **bs)
+{
+ int ret;
+ NBDExport *exp;
+ char *export_name;
+ int nb_read = 0;
+
+ if (!client->structured_reply) {
+ uint32_t tail = length - nb_read;
+ LOG("Structured reply is not negotiated");
+
+ if (nbd_negotiate_drop_sync(client->ioc, tail) != tail) {
+ return -EIO;
+ }
+ ret = nbd_negotiate_send_rep_err(client->ioc, NBD_REP_ERR_INVALID, opt,
+ "Structured reply is not negotiated");
+ g_free(export_name);
+
+ if (ret < 0) {
+ return ret;
+ } else {
+ *bs = NULL;
+ *nb_queries = 0;
+ return length;
+ }
+ }
+
+ nb_read = nbd_negotiate_read_size_string(client->ioc, &export_name,
+ NBD_MAX_NAME_SIZE);
+ if (nb_read < 0) {
+ return nb_read;
+ }
+
+ exp = nbd_export_find(export_name);
+ if (exp == NULL) {
+ uint32_t tail = length - nb_read;
+ LOG("export '%s' is not found", export_name);
+
+ if (nbd_negotiate_drop_sync(client->ioc, tail) != tail) {
+ return -EIO;
+ }
+ ret = nbd_negotiate_send_rep_err(client->ioc, NBD_REP_ERR_INVALID, opt,
+ "export '%s' is not found",
+ export_name);
+ g_free(export_name);
+
+ if (ret < 0) {
+ return ret;
+ } else {
+ *bs = NULL;
+ *nb_queries = 0;
+ return length;
+ }
+ }
+ g_free(export_name);
+
+ *bs = blk_bs(exp->blk);
+ if (*bs == NULL) {
+ LOG("export without bs");
+ return -EINVAL;
+ }
+
+ if (nbd_negotiate_read(client->ioc, nb_queries,
+ sizeof(*nb_queries)) != sizeof(*nb_queries))
+ {
+ LOG("read failed");
+ return -EIO;
+ }
+ cpu_to_be32s(nb_queries);
+
+ nb_read += sizeof(*nb_queries);
+
+ return nb_read;
+}
+
+static int nbd_negotiate_list_meta_context(NBDClient *client, uint32_t length)
+{
+ int ret;
+ BlockDriverState *bs;
+ uint32_t nb_queries;
+ int i;
+ int nb_read;
+
+ nb_read = nbd_negotiate_opt_meta_context_start(client,
+ NBD_OPT_LIST_META_CONTEXT,
+ length, &nb_queries, &bs);
+ if (nb_read < 0) {
+ return nb_read;
+ }
+ if (bs == NULL) {
+ /* error reply was already sent by nbd_negotiate_opt_meta_context_start
+ * */
+ return 0;
+ }
+
+ if (nb_queries == 0) {
+ BdrvDirtyBitmap *bm = NULL;
+
+ if (nb_read != length) {
+ return -EINVAL;
+ }
+
+ while ((bm = bdrv_dirty_bitmap_next(bs, bm)) != 0) {
+ nbd_negotiate_send_bitmap(client->ioc, bdrv_dirty_bitmap_name(bm),
+ NBD_OPT_LIST_META_CONTEXT);
+ }
+ }
+
+ for (i = 0; i < nb_queries; ++i) {
+ ret = nbd_negotiate_one_meta_query(client->ioc, bs,
+ NBD_OPT_LIST_META_CONTEXT, NULL);
+ if (ret < 0) {
+ return ret;
+ }
+
+ nb_read += ret;
+ }
+
+ if (nb_read != length) {
+ return -EINVAL;
+ }
+
+ return nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK,
+ NBD_OPT_LIST_META_CONTEXT);
+}
+
+static int nbd_negotiate_set_meta_context(NBDClient *client, uint32_t length)
+{
+ int ret;
+ BlockDriverState *bs;
+ uint32_t nb_queries;
+ int nb_read;
+
+ nb_read = nbd_negotiate_opt_meta_context_start(client,
+ NBD_OPT_SET_META_CONTEXT,
+ length, &nb_queries, &bs);
+ if (nb_read < 0) {
+ return nb_read;
+ }
+ if (bs == NULL) {
+ /* error reply was already sent by nbd_negotiate_opt_meta_context_start
+ * */
+ return 0;
+ }
+
+ if (nb_queries == 0) {
+ return nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK,
+ NBD_OPT_SET_META_CONTEXT);
+ }
+
+ if (nb_queries > 1) {
+ return nbd_negotiate_send_rep_err(client->ioc, NBD_REP_ERR_TOO_BIG,
+ NBD_OPT_SET_META_CONTEXT,
+ "Only one exporting context is"
+ "supported");
+ }
+
+ ret = nbd_negotiate_one_meta_query(client->ioc, bs,
+ NBD_OPT_SET_META_CONTEXT,
+ &client->export_bitmap);
+ if (ret < 0) {
+ return ret;
+ }
+
+ return nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK,
+ NBD_OPT_SET_META_CONTEXT);
+}
/* Process all NBD_OPT_* client option commands.
* Return -errno on error, 0 on success. */
static int nbd_negotiate_options(NBDClient *client)
@@ -585,6 +885,20 @@ static int nbd_negotiate_options(NBDClient *client)
}
break;
+ case NBD_OPT_LIST_META_CONTEXT:
+ ret = nbd_negotiate_list_meta_context(client, length);
+ if (ret < 0) {
+ return ret;
+ }
+ break;
+
+ case NBD_OPT_SET_META_CONTEXT:
+ ret = nbd_negotiate_set_meta_context(client, length);
+ if (ret < 0) {
+ return ret;
+ }
+ break;
+
default:
if (nbd_negotiate_drop_sync(client->ioc, length) != length) {
return -EIO;
@@ -1159,6 +1473,124 @@ static int nbd_co_send_structured_none(NBDClient *client, uint64_t handle)
return nbd_co_send_buf(client, &chunk, sizeof(chunk));
}
+#define MAX_EXTENT_LENGTH UINT32_MAX
+
+static unsigned add_extents(NBDExtent *extents, unsigned nb_extents,
+ uint64_t length, uint32_t flags)
+{
+ unsigned i = 0;
+ uint32_t big_chunk = (MAX_EXTENT_LENGTH >> 9) << 9;
+ uint32_t big_chunk_be = cpu_to_be32(big_chunk);
+ uint32_t flags_be = cpu_to_be32(flags);
+
+ for (i = 0; i < nb_extents && length > MAX_EXTENT_LENGTH;
+ i++, length -= big_chunk)
+ {
+ extents[i].length = big_chunk_be;
+ extents[i].flags = flags_be;
+ }
+
+ if (length > 0 && i < nb_extents) {
+ extents[i].length = cpu_to_be32(length);
+ extents[i].flags = flags_be;
+ i++;
+ }
+
+ return i;
+}
+
+static unsigned bitmap_to_extents(BdrvDirtyBitmap *bitmap, uint64_t offset,
+ uint64_t length, NBDExtent *extents,
+ unsigned nb_extents)
+{
+ uint64_t begin, end; /* dirty region */
+ uint64_t start_sector = offset >> BDRV_SECTOR_BITS;
+ uint64_t last_sector = (offset + length - 1) >> BDRV_SECTOR_BITS;
+ unsigned i = 0;
+ uint64_t len;
+ uint32_t ma = -1;
+ ma = (ma / bdrv_dirty_bitmap_granularity(bitmap)) *
+ bdrv_dirty_bitmap_granularity(bitmap);
+
+ BdrvDirtyBitmapIter *it = bdrv_dirty_iter_new(bitmap, start_sector);
+
+ assert(nb_extents > 0);
+
+ begin = bdrv_dirty_iter_next(it);
+ if (begin == -1) {
+ begin = last_sector + 1;
+ }
+ if (begin > start_sector) {
+ len = (begin - start_sector) << BDRV_SECTOR_BITS;
+ i += add_extents(extents + i, nb_extents - i, len, 0);
+ }
+
+ while (begin != -1 && begin <= last_sector && i < nb_extents) {
+ end = bdrv_dirty_bitmap_next_zero(bitmap, begin + 1);
+
+ i += add_extents(extents + i, nb_extents - i,
+ (end - begin) << BDRV_SECTOR_BITS, 1);
+
+ if (end > last_sector || i >= nb_extents) {
+ break;
+ }
+
+ bdrv_set_dirty_iter(it, end);
+ begin = bdrv_dirty_iter_next(it);
+ if (begin == -1) {
+ begin = last_sector + 1;
+ }
+ if (begin > end) {
+ i += add_extents(extents + i, nb_extents - i,
+ (begin - end) << BDRV_SECTOR_BITS, 0);
+ }
+ }
+
+ bdrv_dirty_iter_free(it);
+
+ extents[0].length =
+ cpu_to_be32(be32_to_cpu(extents[0].length) -
+ (offset - (start_sector << BDRV_SECTOR_BITS)));
+
+ return i;
+}
+
+static int nbd_co_send_extents(NBDClient *client, uint64_t handle,
+ NBDExtent *extents, unsigned nb_extents,
+ uint32_t context_id)
+{
+ NBDStructuredMeta chunk;
+
+ struct iovec iov[] = {
+ {.iov_base = &chunk, .iov_len = sizeof(chunk)},
+ {.iov_base = extents, .iov_len = nb_extents * sizeof(extents[0])}
+ };
+
+ set_be_chunk(&chunk.h, NBD_REPLY_FLAG_DONE, NBD_REPLY_TYPE_BLOCK_STATUS,
+ handle, sizeof(chunk) - sizeof(chunk.h) + iov[1].iov_len);
+ stl_be_p(&chunk.context_id, context_id);
+
+ return nbd_co_send_iov(client, iov, 2);
+}
+
+static int nbd_co_send_bitmap(NBDClient *client, uint64_t handle,
+ BdrvDirtyBitmap *bitmap, uint64_t offset,
+ uint64_t length, uint32_t context_id)
+{
+ int ret;
+ unsigned nb_extents;
+ NBDExtent *extents = g_new(NBDExtent, NBD_MAX_BITMAP_EXTENTS);
+
+ nb_extents = bitmap_to_extents(bitmap, offset, length, extents,
+ NBD_MAX_BITMAP_EXTENTS);
+
+ ret = nbd_co_send_extents(client, handle, extents, nb_extents, context_id);
+
+ g_free(extents);
+
+ return ret;
+}
+
/* Collect a client request. Return 0 if request looks valid, -EAGAIN
* to keep trying the collection, -EIO to drop connection right away,
* and any other negative value to report an error to the client
@@ -1437,6 +1869,19 @@ static void nbd_trip(void *opaque)
goto out;
}
break;
+ case NBD_CMD_BLOCK_STATUS:
+ TRACE("Request type is BLOCK_STATUS");
+ if (client->export_bitmap == NULL) {
+ reply.error = EINVAL;
+ goto error_reply;
+ }
+ ret = nbd_co_send_bitmap(req->client, request.handle,
+ client->export_bitmap, request.from,
+ request.len, 0);
+ if (ret < 0) {
+ goto out;
+ }
+ break;
default:
LOG("invalid request type (%" PRIu32 ") received", request.type);
reply.error = EINVAL;
Only one meta context type is defined: qemu-bitmap:<bitmap-name>. Maximum one query is allowed for NBD_OPT_{SET,LIST}_META_CONTEXT, NBD_REP_ERR_TOO_BIG is returned otherwise. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> --- include/block/nbd.h | 15 ++ nbd/nbd-internal.h | 6 + nbd/server.c | 445 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 466 insertions(+)