search for: libnbd_handshake_flag_fixed_newstyl

...duce the set of flags supported by the client to test that +a particular server can handle various clients that were compliant to +older versions of the NBD specification. + +The C<flags> argument is a bitmask, including zero or more of the +following handshake flags: + +=over 4 + +=item C<LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE> = 1 + +The server gracefully handles unknown option requests from the +client, rather than disconnecting. Without this flag, a client +cannot safely request to use extensions such as TLS encryption or +structured replies, as the request may cause an older server to +drop the connection. + +=i...

...ply.c b/generator/states-newstyle-opt-structured-reply.c index 71a4952..8b689a6 100644 --- a/generator/states-newstyle-opt-structured-reply.c +++ b/generator/states-newstyle-opt-structured-reply.c @@ -22,7 +22,7 @@ STATE_MACHINE { NEWSTYLE.OPT_STRUCTURED_REPLY.START: assert (h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE); if (!h->request_sr) { - SET_NEXT_STATE (%^OPT_SET_META_CONTEXT.START); + SET_NEXT_STATE (%^OPT_META_CONTEXT.START); return 0; } @@ -87,7 +87,7 @@ STATE_MACHINE { if (h->opt_mode) SET_NEXT_STATE (%.NEGOTIATING); else - SET_NEXT_STATE (%^OPT_SET_META_CONTEXT.S...

.../* NEWSTYLE can be entered multiple times, from MAGIC.CHECK_MAGIC and + * during various nbd_opt_* calls during NEGOTIATION. Each previous + * state has informed us what we still need to do. + */ + switch (h->current_opt) { + case NBD_OPT_ABORT: + if ((h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE) == 0) { + SET_NEXT_STATE (%.DEAD); + set_error (ENOTSUP, "handshake: server is not using fixed newstyle"); + return 0; + } + SET_NEXT_STATE (%PREPARE_OPT_ABORT); + return 0; + case 0: + break; + default: + abort (); + } + } h-&g...

Well, I'm not quite done (I still want to get nbdinfo to work on a single nbd connection for all cases when reading the heads of the file is not required), but I'm happy with patches 1-11, and 12-13 show where I'm headed for getting NBD_OPT_INFO to work. Posting now to see if some of the earlier patches are ready to commit while I continue working on the latter half. Eric Blake (13):

...(%SEND_OPT_ABORT); + return 0; + case 0: + break; + default: + abort (); + } + } h->rbuf = &h->sbuf; h->rlen = sizeof h->sbuf.gflags; SET_NEXT_STATE (%RECV_GFLAGS); @@ -136,6 +155,11 @@ STATE_MACHINE { return 0; } + if ((h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE) == 0) + h->protocol = "newstyle"; + else + h->protocol = "newstyle-fixed"; + cflags = h->gflags; h->sbuf.cflags = htobe32 (cflags); h->wbuf = &h->sbuf; @@ -155,12 +179,23 @@ STATE_MACHINE { } return 0; + NEWSTYLE.SEND_OPT_ABORT: +...

I'm posting this now, as I'm at the end of a workday and I got things working for manual experimentation. Still to do: - write interop tests for qemu-nbd and nbdkit (including my proposed patch addition of qemu-nbd -A to show qemu:allocation-depth) - figure out if we can make 'nbdinfo --map' use the new API to automatically select all contexts advertised by the server Eric Blake

...+65,11 @@ nbd_create (void) h->tls_verify_peer = true; h->request_sr = true; - h->uri_allow_transports = (uint32_t) -1; + h->uri_allow_transports = LIBNBD_ALLOW_TRANSPORT_MASK; h->uri_allow_tls = LIBNBD_TLS_ALLOW; h->uri_allow_local_file = false; - h->gflags = (LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE | - LIBNBD_HANDSHAKE_FLAG_NO_ZEROES); + h->gflags = LIBNBD_HANDSHAKE_FLAG_MASK; s = getenv ("LIBNBD_DEBUG"); h->debug = s && strcmp (s, "1") == 0; diff --git a/python/t/110-defaults.py b/python/t/110-defaults.py index 47c1d02..fb961cf 100644...

...uint32_t len; + uint32_t len, opt; /* If the server doesn't support SRs then we must skip this group. * Also we skip the group if the client didn't request any metadata - * contexts. + * contexts, when doing SET (but an empty LIST is okay). */ assert (h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE); nbd_internal_reset_size_and_flags (h); - if (!h->structured_replies || - nbd_internal_string_list_length (h->request_meta_contexts) == 0) { - SET_NEXT_STATE (%^OPT_GO.START); - return 0; + if (h->opt_current == NBD_OPT_LIST_META_CONTEXT) { + assert (h->opt_mode); +...

...nbd_unlocked_aio_opt_go (struct nbd_handle *h, /* Issue NBD_OPT_INFO without waiting. */ int nbd_unlocked_aio_opt_info (struct nbd_handle *h, - nbd_completion_callback complete) + nbd_completion_callback *complete) { if ((h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE) == 0) { set_error (ENOTSUP, "server is not using fixed newstyle protocol"); @@ -189,7 +189,7 @@ nbd_unlocked_aio_opt_info (struct nbd_handle *h, } h->opt_current = NBD_OPT_INFO; - h->opt_cb.completion = complete; + h->opt_cb.completion = *complete; if (nbd_int...

...h->rlen = sizeof h->sbuf.gflags; > SET_NEXT_STATE (%RECV_GFLAGS); And this hunk which I guess is related to above but I couldn't quite work out why it's needed either. > @@ -136,6 +155,11 @@ STATE_MACHINE { > return 0; > } > > + if ((h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE) == 0) > + h->protocol = "newstyle"; > + else > + h->protocol = "newstyle-fixed"; > + > cflags = h->gflags; > h->sbuf.cflags = htobe32 (cflags); > h->wbuf = &h->sbuf; This makes me think maybe we should just derive thi...

...r. diff --git a/generator/states-newstyle-opt-list.c b/generator/states-newstyle-opt-list.c index 19b353e..b8d8025 100644 --- a/generator/states-newstyle-opt-list.c +++ b/generator/states-newstyle-opt-list.c @@ -24,7 +24,8 @@ STATE_MACHINE { NEWSTYLE.OPT_LIST.START: assert (h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE); - assert (h->opt_mode && h->exports && !h->nr_exports); + assert (h->opt_mode && h->current_opt == NBD_OPT_LIST); + assert (CALLBACK_IS_NOT_NULL (h->opt_cb.fn.list)); h->sbuf.option.version = htobe64 (NBD_NEW_VERSION); h->sbuf.option.optio...

...; SET_NEXT_STATE (%RECV_GFLAGS); > > And this hunk which I guess is related to above but I couldn't quite > work out why it's needed either. > >> @@ -136,6 +155,11 @@ STATE_MACHINE { >> return 0; >> } >> >> + if ((h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE) == 0) >> + h->protocol = "newstyle"; >> + else >> + h->protocol = "newstyle-fixed"; >> + >> cflags = h->gflags; >> h->sbuf.cflags = htobe32 (cflags); >> h->wbuf = &h->sbuf; > > This makes...

This is a subset of my v2 posting, but limited to just the NBD_OPT_LIST handling. The biggest change since v2 is the addition of added unit testing in all four language bindings (C, python, ocaml, golang). The tests require nbdkit built from git on PATH, and may not be entirely idiomatic, but I at least validated that they catch issues (for example, adding an exit statement near the end of the

.../handle.c +++ b/lib/handle.c @@ -64,6 +64,11 @@ nbd_create (void) h->unique = 1; h->tls_verify_peer = true; h->request_sr = true; + + h->uri_allow_transports = (uint32_t) -1; + h->uri_allow_tls = LIBNBD_TLS_ALLOW; + h->uri_allow_local_file = false; + h->gflags = (LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE | LIBNBD_HANDSHAKE_FLAG_NO_ZEROES); @@ -360,3 +365,24 @@ nbd_unlocked_get_protocol (struct nbd_handle *h) return h->protocol; } + +int +nbd_unlocked_set_uri_allow_transports (struct nbd_handle *h, uint32_t mask) +{ + h->uri_allow_transports = mask; + return 0; +} +...

...r. diff --git a/generator/states-newstyle-opt-list.c b/generator/states-newstyle-opt-list.c index 19b353e..a549bdc 100644 --- a/generator/states-newstyle-opt-list.c +++ b/generator/states-newstyle-opt-list.c @@ -24,7 +24,8 @@ STATE_MACHINE { NEWSTYLE.OPT_LIST.START: assert (h->gflags & LIBNBD_HANDSHAKE_FLAG_FIXED_NEWSTYLE); - assert (h->opt_mode && h->exports && !h->nr_exports); + assert (h->opt_mode && h->opt_current == NBD_OPT_LIST); + assert (CALLBACK_IS_NOT_NULL (h->opt_cb.fn.list)); h->sbuf.option.version = htobe64 (NBD_NEW_VERSION); h->sbuf.option.optio...

Previous discussion: https://www.redhat.com/archives/libguestfs/2019-August/msg00102.html Last night I experimentally added support for URIs that contain the query parameter tls-psk-file, as part of rewriting the tests to cover more of the URI code. So you can now have a URI like: nbds://alice@localhost/?tls-psk-file=keys.psk However there's an obvious security problem here because now

This replaces 13/13 of my v2 series; and now that it has pretty good testsuite coverage and demonstrable performance improvement to nbdinfo, I'm going ahead and pushing this now. We may still want to add further nbd_opt_* commands for other fine-grained tuning of negotiation, but for now, I think things have stabilized on this end, and I can return to polishing .list_exports on the nbdkit

As promised in my earlier thread on libnbd completion callback question. Eric Blake (2): generator: Refactor handling of closures in unlocked functions generator: Free closures on failure docs/libnbd.pod | 2 +- generator/C.ml | 48 +++++++++++------ generator/C.mli | 1 + lib/debug.c | 7 +-- lib/opt.c | 31 ++++++-----

In v2: ack'ed preliminary patches have been pushed, and I've added a lot of testsuite coverage as well as putting the new API to use in nbdinfo. Eric Blake (2): api: Add nbd_opt_list_meta_context info: List available meta-contexts lib/internal.h | 1 + generator/API.ml | 84 ++++++++-

Well, the golang changes (patch 1 and 2/3 of v1) were already committed, all that was left was the OCaml changes. I'm a lot happier with how things turned out with an UNKNOWN constructor in the OCaml variants. Eric Blake (3): tests: Enhance coverage of enum/flag range checking ocaml: Support unknown values for Enum/Flags ocaml: Typesafe returns for REnum/RFlags generator/OCaml.ml