Re: [libvirt] [Qemu-devel] QMP; unsigned 64-bit ints; JSON standards compliance

> The Golang JSON parser decodes JSON numbers to float64 by default so > will have this precision limitation too, though at least they provide > a backdoor for custom parsing from the original serialized representation. > >> > QEMU, and indeed many applications, want to handle 64-bit integers. >> > The C JSON library impls have traditionally mapped integers to the >> > data type 'long long int' which gives a min/max of -(2^63) / 2^63-1. >> > >> > QEMU however /really/ needs 64-bit unsigned integers, ie a max 2^64-1. Correct. Support for integers 2^63..2^64-1 is relatively recent: commit 2bc7cfea095 (v2.10, 2017). Since we really needed these, the QObject input visitor silently casts negative integers to uint64_t. It still does for backward compatibility. Commit 5923f85fb82 (right after 2bc7cfea095) explains The input visitor will cast i64 input to u64 for compatibility reasons (existing json QMP client already use negative i64 for large u64, and expect an implicit cast in qemu). Note: before the patch, uint64_t values above INT64_MAX are sent over json QMP as negative values, e.g. UINT64_MAX is sent as -1. After the patch, they are sent unmodified. Clearly a bug fix, but we have to consider compatibility issues anyway. libvirt should cope fine, because its parsing of unsigned integers accepts negative values modulo 2^64. There's hope that other clients will, too.

>> > Libvirt has historically used the YAJL library which uses 'long long int' >> > and thus can't officially go beyond 2^63-1 values. Fortunately it lets >> > libvirt get at the raw json string, so libvirt can re-parse the value >> > to get an 'unsigned long long'. >> > >> > We recently tried to switch to Jansson because YAJL has a dead upstream >> > for many years and countless unanswered bugs & patches. Unfortunately we >> > forgot about this need for 2^64-1 max, and Jansson also uses 'long long int' >> > and raises a fatal parse error for unsigned 64-bit values above 2^63-1. It >> > also provides no backdoor for libvirt todo its own integer parsing. Thus >> > we had to abort our switch to jansson as it broke parsing QEMU's JSON: >> > >> > https://bugzilla.redhat.com/show_bug.cgi?id=1614569 >> > >> > Other JSON libraries we've investigated have similar problems. I imagine >> > the same may well be true of non-C based JOSN impls, though I've not >> > investigated in any detail. >> > >> > Essentially libvirt is stuck with either using the dead YAJL library >> > forever, or writing its own JSON parser (most likely copying QEMU's >> > JSON code into libvirt's git). >> > >> > This feels like a very unappealing situation to be in as not being >> > able to use a JSON library of our choice is loosing one of the key >> > benefits of using a standard data format. >> > >> > Thus I'd like to see a solution to this to allow QMP to be reliably >> > consumed by any JSON library that exists. JSON is terrible at interoperability, so good luck with that. If you reduce your order to "the commonly used JSON libraries we know", we can talk.

>> > I can think of some options: >> > >> > 1. Encode unsigned 64-bit integers as signed 64-bit integers. >> > >> > This follows the example that most C libraries map JSON ints >> > to 'long long int'. This is still relying on undefined >> > behaviour as apps don't need to support > 2^53-1. >> > >> > Apps would need to cast back to 'unsigned long long' for >> > those QMP fields they know are supposed to be unsigned. Ugly. It's also what we did until v2.10, August 2017. QMP's input direction still does it, for backward compatibility. >> > >> > >> > 2. Encode all 64-bit integers as a pair of 32-bit integers. >> > >> > This is fully compliant with the JSON spec as each half >> > is fully within the declared limits. App has to split or >> > assemble the 2 pieces from/to a signed/unsigned 64-bit >> > int as needed. Differently ugly. >> > >> > >> > 3. Encode all 64-bit integers as strings >> > >> > The application has todo all parsing/formatting client >> > side. Yet another ugly. >> > >> > >> > None of these changes are backwards compatible, so I doubt we could make >> > the change transparently in QMP. Instead we would have to have a >> > QMP greeting message capability where the client can request enablement >> > of the enhanced integer handling. We might be able to do option 1 without capability negotiation. v2.10's change from option 1 to what we have now produced zero complaints. On the other hand, we made that change for a reason, so we may want a "send large integers as negative integers" capability regardless. >> > >> > Any of the three options above would likely work for libvirt, but I >> > would have a slight preference for either 2 or 3, so that we become >> > 100% standards compliant. There's no such thing. You mean "we maximize interoperability with common implementations of JSON".

Let's talk implementation for a bit. Encoding and decoding integers in funny ways should be fairly easy in the QObject visitors. The generated QMP marshallers all use them. Trouble is a few commands still bypass the generated marshallers, and mess with the QObject themselves: * query-qmp-schema: minor hack explained in qmp_query_qmp_schema()'s comment. Should be harmless. * netdev_add: not QAPIfied. Eric's patches to QAPIfy it got stuck because they reject some abuses like passing numbers and bools as strings. * device_add: not QAPIfied. We're not sure QAPIfication is feasible. netdev_add and device_add both use qemu_opts_from_qdict(). Perhaps we could hack that to mirror what the QObject visitor do. Else, we might have to do it in the JSON parser. Should be possible, but I'd rather not. >> My preference would be 3 with the strings defined as being >> %x lower case hex formated with a 0x prefix and no longer than 18 characters >> ("0x" + 16 nybbles). Zero padding allowed but not required. >> It's readable and unambiguous when dealing with addresses; I don't want >> to have to start decoding (2) by hand when debugging. > > Yep, that's a good point about readability. QMP sending all integers in decimal is inconvenient for some values, such as addresses. QMP sending all (large) integers in hexadecimal would be inconvenient for other values. Let's keep it simple & stupid. If you want sophistication, JSON is the wrong choice. Option 1 feels simplest.

Option 2 feels ugliest. Less simple, more interoperable than option 1.

Option 3 is like option 2, just not quite as ugly.

Can we agree to eliminate option 2 from the race?