Libguestfs - Mar 2022 - [PATCH v2 3/3] spec: Clarify BLOCK_STATUS reply details

Widening the cc list this time.  Patch 1 and 2 are unchanged, but I
spent more time rewording the changes in patch 3.

Eric Blake (3):
  docs: Clarify NBD_REPLY_TYPE_ERROR lengths
  docs: Clarify structured reads vs. block size constraints
  spec: Clarify BLOCK_STATUS reply details

 doc/proto.md | 57 ++++++++++++++++++++++++++++++++--------------------
 1 file changed, 35 insertions(+), 22 deletions(-)

-- 
2.35.1

Add explicit mention that human-readable error message strings must
comply with the overall NBD string limits.

Signed-off-by: Eric Blake <eblake at redhat.com>
---
 doc/proto.md | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/doc/proto.md b/doc/proto.md
index 81ac755..bfebb5a 100644
--- a/doc/proto.md
+++ b/doc/proto.md
@@ -1824,7 +1824,9 @@ remaining structured fields at the end.
   be at least 6.  This chunk represents that an error occurred,
   and the client MAY NOT make any assumptions about partial
   success. This type SHOULD NOT be used more than once in a
-  structured reply.  Valid as a reply to any request.
+  structured reply.  Valid as a reply to any request.  Note that
+  *message length* MUST NOT exceed the 4096 bytes string length
+  limit.

   The payload is structured as:

-- 
2.35.1

The text for structured reads mentioned a mandatory split of certain
large reads, without also mentioning that large reads are generally
not possible when block constraints are in play.

Signed-off-by: Eric Blake <eblake at redhat.com>
---
 doc/proto.md | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/doc/proto.md b/doc/proto.md
index bfebb5a..8a817d2 100644
--- a/doc/proto.md
+++ b/doc/proto.md
@@ -1911,13 +1911,14 @@ The following request types exist:
     chunks that describe data outside the offset and length of the
     request, but MAY send the content chunks in any order (the client
     MUST reassemble content chunks into the correct order), and MAY
-    send additional content chunks even after reporting an error chunk.
-    Note that a request for more than 2^32 - 8 bytes MUST be split
-    into at least two chunks, so as not to overflow the length field
-    of a reply while still allowing space for the offset of each
-    chunk.  When no error is detected, the server MUST send enough
-    data chunks to cover the entire region described by the offset and
-    length of the client's request.
+    send additional content chunks even after reporting an error
+    chunk.  Note that a request for more than 2^32 - 8 bytes (if
+    permitted by block size constraints) MUST be split into at least
+    two chunks, so as not to overflow the length field of a reply
+    while still allowing space for the offset of each chunk.  When no
+    error is detected, the server MUST send enough data chunks to
+    cover the entire region described by the offset and length of the
+    client's request.

     To minimize traffic, the server MAY use a content or error chunk
     as the final chunk by setting the `NBD_REPLY_FLAG_DONE` flag, but
-- 
2.35.1

Our docs were inconsistent on whether a NBD_REPLY_TYPE_BLOCK_STATUS
reply chunk can exceed the client's requested length, and silent on
whether the lengths must be consistent whem multiple contexts were
negotiated.  Clarify this to match existing practice as implemented in
qemu-nbd.  Clean up some nearby grammatical errors while at it.
---
 doc/proto.md | 38 ++++++++++++++++++++++++--------------
 1 file changed, 24 insertions(+), 14 deletions(-)

diff --git a/doc/proto.md b/doc/proto.md
index 8a817d2..8276201 100644
--- a/doc/proto.md
+++ b/doc/proto.md
@@ -882,15 +882,22 @@ The procedure works as follows:
   server supports.
 - During transmission, a client can then indicate interest in metadata
   for a given region by way of the `NBD_CMD_BLOCK_STATUS` command,
-  where *offset* and *length* indicate the area of interest. The
-  server MUST then respond with the requested information, for all
-  contexts which were selected during negotiation. For every metadata
-  context, the server sends one set of extent chunks, where the sizes
-  of the extents MUST be less than or equal to the length as specified
-  in the request. Each extent comes with a *flags* field, the
-  semantics of which are defined by the metadata context.
-- A server MUST reply to `NBD_CMD_BLOCK_STATUS` with a structured
-  reply of type `NBD_REPLY_TYPE_BLOCK_STATUS`.
+  where *offset* and *length* indicate the area of interest. On
+  success, the server MUST respond with one structured reply chunk of
+  type `NBD_REPLY_TYPE_BLOCK_STATUS` per metadata context selected
+  during negotiation, where each reply chunk is a list of one or more
+  extents for that context.
+
+The client's requested *length* is only a hint to the server, so the
+cumulative size of the extents in each chunk of the server's reply may
+be shorter or longer the original request; and when more than one
+metadata context was negotiated, the cumulative length per context may
+differ within a single block status request.  Each extent comes with a
+*flags* field, the semantics of which are defined by the metadata
+context.  The client may use the `NBD_CMD_FLAG_REQ_ONE` flag to
+further constrains the server's reply so that each chunk contains
+exactly one extent whose length does not exceed the client's original
+*length*.

 A client MUST NOT use `NBD_CMD_BLOCK_STATUS` unless it selected a
 nonzero number of metadata contexts during negotiation, and used the
@@ -1778,8 +1785,8 @@ MUST initiate a hard disconnect.
   *length* MUST be 4 + (a positive integer multiple of 8).  This reply
   represents a series of consecutive block descriptors where the sum
   of the length fields within the descriptors is subject to further
-  constraints documented below. This chunk type MUST appear
-  exactly once per metadata ID in a structured reply.
+  constraints documented below.  A successful block status request MUST
+  have exactly one status chunk per negotiated metadata context ID.

   The payload starts with:

@@ -1801,15 +1808,18 @@ MUST initiate a hard disconnect.
   *length* of the final extent MAY result in a sum larger than the
   original requested length, if the server has that information anyway
   as a side effect of reporting the status of the requested region.
+  When multiple metadata contexts are negotiated, the cumulative
+  lengths in each chunk reply need not be identical.

   Even if the client did not use the `NBD_CMD_FLAG_REQ_ONE` flag in
   its request, the server MAY return fewer descriptors in the reply
   than would be required to fully specify the whole range of requested
   information to the client, if looking up the information would be
   too resource-intensive for the server, so long as at least one
-  extent is returned. Servers should however be aware that most
-  clients implementations will then simply ask for the next extent
-  instead.
+  extent is returned.  Servers should however be aware that most
+  client implementations will likely follow up with a request for
+  extent information at the first offset not covered by a
+  reduced-length reply.

 All error chunk types have bit 15 set, and begin with the same
 *error*, *message length*, and optional *message* fields as
-- 
2.35.1