On 04/20/2016 04:37 AM, Alex Bligh wrote: > Eric, > > In general I think we should probably apply this and see how it > looks though I suspect fixing things is going to be an iterative > matter. > >> @@ -1086,10 +1088,7 @@ The following request types exist: >> If structured replies were not negotiated, then a read request >> MUST always be answered by a simple reply, as documented above >> (using magic 0x67446698 `NBD_SIMPLE_REPLY_MAGIC`, and containing >> - length bytes of data according to the client's request, although >> - those bytes MAY be invalid if an error is returned, and a hard >> - disconnect MUST be initiated if an error occurs after a header >> - claiming no error). >> + length bytes of data according to the client's request). > > That's a nit that isn't a move. > > '*length*' bytes of data > > What's the reason for removing the fact that in simple replies you > can send invalid data (as the client will be expecting the data)? When I first wrote that paragraph, it lived in the experimental section, way below where the original description of NBD_CMD_READ simple replies lived, so it was a nice parenthetical refresher. But now that the text is alongside the rest of NBD_CMD_READ, it is both redundant (the very next paragraph says the same thing), and wrong (it says "as documented above" which applies to the "magic 0x67446698" but NOT to the "bytes MAY be invalid if an error", because there is no text above that states the bytes may be invalid until the next paragraph). -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org
Attachment:
signature.asc
Description: OpenPGP digital signature