2 ======================================================================
4 General ideas for improvements
5 ------------------------------
7 * Listen on specific interfaces or protocols (eg. only IPv6).
9 * Performance - measure and improve it. Chart it over various buffer
10 sizes and threads, as that should make it easier to identify
13 * Exit on last connection (the default behaviour of qemu-nbd unless
16 * Limit number of incoming connections (like qemu-nbd -e).
18 * For parallel plugins, only create threads on demand from parallel
19 client requests, rather than pre-creating all threads at connection
20 time, up to the thread pool size limit. Of course, once created, a
21 thread is reused as possible until the connection closes.
23 * Async callbacks. The current parallel support requires one thread
24 per pending message; a solution with fewer threads would split
25 low-level code between request and response, where the callback has
26 to inform nbdkit when the response is ready:
27 https://www.redhat.com/archives/libguestfs/2018-January/msg00149.html
29 * More NBD protocol features. The currently missing features are
30 structured replies for sparse reads, block size constraints, and
33 * Add a callback to let plugins request minimum alignment for the
34 buffer to pread/pwrite; useful for a plugin utilizing O_DIRECT or
35 other situation where pre-aligned buffers are more efficient.
36 Ideally, a blocksize filter would honor strict alignment below and
37 advertise loose alignment above; all other filters (particularly
38 ones like offset) can fail to initialize if they can't guarantee
39 strict alignment and don't want to deal with bounce buffers.
41 * Test that zero-length read/write/extents requests behave sanely
42 (NBD protocol says they are unspecified).
44 * If a client negotiates structured replies, and issues a read/extents
45 call that exceeds EOF (qemu 3.1 is one such client, when nbdkit
46 serves non-sector-aligned images), return the valid answer for the
47 subset of the request in range and then NBD_REPLY_TYPE_ERROR_OFFSET
48 for the tail, rather than erroring the entire request.
50 * Test and document how to run nbdkit from inetd and xinetd in
53 * Audit the code base to get rid of strerror() usage (the function is
54 not thread-safe); however, using geterror_r() can be tricky as it
55 has a different signature in glibc than in POSIX.
57 * Teach nbdkit_error() to have smart newline appending (for existing
58 inconsistent clients), while fixing internal uses to omit the
59 newline. Commit ef4f72ef has some ideas on smart newlines, but that
60 should probably be factored into a utility function.
62 * Add a mode of operation where nbdkit is handed a pre-opened fd to be
63 used immediately in transmission phase (skipping handshake). There
64 are already third-party clients of the kernel's /dev/nbdX which rely
65 on their own protocol instead of NBD handshake, before calling
66 ioctl(NBD_SET_SOCK); this mode would let the third-party client
67 continue to keep their non-standard handshake while utilizing nbdkit
68 to prototype new behaviors in serving the kernel.
70 * Clients should be able to list export names supported by plugins.
71 Current behaviour is not really correct: We only list the -e
72 parameter from the command line, which is different from the export
73 name(s) that a plugin might want to support. Probably we should
74 deprecate the -e option entirely since it does nothing useful.
76 * Add plugin "connect" method. This would be called on a connection
77 before handshaking or TLS negotiation, and could be used (with
78 nbdkit_peer_name) to accept or reject connections based on IP
79 address, rather like a poor man's TCP wrappers. See also commit
82 * Background thread for filters. Some filters (readahead, cache and
83 proposed scan filter - see below) could be more effective if they
84 were able to defer work to a background thread. However it's not as
85 simple as just creating a background thread, because you only have
86 access to the connection / next_ops from the context of a filter
87 callback. (Also you can't "save" next_ops because it becomes
88 invalid outside the callback). The background thread would need to
89 have its own connection to the plugin, which would be independent of
90 any client connection, and this requires some care because it breaks
91 an existing assumption.
93 * Add scan filter. This would be placed on top of cache filters and
94 would scan (read) the whole disk in the background, ensuring it is
95 copied into the cache. Useful if you have a slow plugin, limited
96 size device, and lots of local disk space, especially if you know
97 that the NBD clients will eventually read all of the device. RWMJ
98 wrote an implementation of this but it doesn't work well without a
101 * "nbdkit.so": nbdkit as a loadable shared library. The aim of nbdkit
102 is to make it reusable from other programs (see nbdkit-captive(1)).
103 If it was a loadable shared library it would be even more reusable.
104 API would allow you to create an nbdkit instance, configure it (same
105 as the current command line), start it serving on a socket, etc.
106 However perhaps the current ability to work well as a subprocess is
107 good enough? Also allowing multiple instances of nbdkit to be
108 loaded in the same process is probably impossible.
110 * Examine other fuzzers: https://gitlab.com/akihe/radamsa
112 * Analyze code with Coverity.
114 Suggestions for plugins
115 -----------------------
117 Note: qemu supports other formats such as libnfs, iscsi, gluster and
118 ceph/rbd, and while similar plugins could be written for nbdkit there
119 is no compelling reason unless the result is better than qemu-nbd.
120 For the majority of users it would be better if they were directed to
121 qemu-nbd for these use cases.
125 https://lists.gnu.org/archive/html/qemu-devel/2017-11/msg02971.html
126 is a partial solution but it needs cleaning up.
128 nbdkit-floppy-plugin:
130 * Add boot sector support. In theory this is easy (eg. using
131 SYSLINUX), but the practical reality of making a fully bootable
132 floppy is rather more complex.
134 * Add multiple dir merging.
136 nbdkit-linuxdisk-plugin:
138 * Add multiple dir merging (in e2fsprogs mke2fs).
140 Suggestions for language plugins
141 --------------------------------
145 * Get the __docstring__ from the module and print it in --help output.
146 This requires changes to the core API so that config_help is a
147 function rather than a variable (see V3 suggestions below).
151 * Consider supporting a more idiomatic style for writing Rust plugins.
153 * Better documentation.
157 * There is no attempt to ‘make install’ or otherwise package the
158 crate. Since it looks as if Rust code is normally distributed as
159 source it's not clear what that would even mean.
161 Suggestions for filters
162 -----------------------
164 * tar plugin should really be a filter
166 * gzip plugin should really be a filter
168 * libarchive could be used to implement a general tar/zip filter
170 * LUKS encrypt/decrypt filter, bonus points if compatible with qemu
171 LUKS-encrypted disk images
173 * masking plugin features for testing clients (see 'nozero' and 'fua'
174 filters for examples)
176 * nbdkit-cache-filter should handle ENOSPC errors automatically by
177 reclaiming blocks from the cache
179 * zstd filter was requested as a way to do what we currently do with
180 xz but saving many hours on compression (at the cost of hundreds of
182 https://github.com/facebook/zstd/issues/395#issuecomment-535875379
186 * allow other kinds of traffic shaping such as VBR
188 * limit traffic per client (ie. per IP address)
190 * split large requests to avoid long, lumpy sleeps when request size
191 is much larger than rate limit
195 * allow user to specify which errors cause a retry and which ones are
196 passed through; for example there's probably no point retrying on
199 * implement a softer mode (retry-reopen=no) where we don't reopen the
200 plugin, we just retry the data command that failed
202 * there are all kinds of extra complications possible here,
203 eg. specifying a pattern of retrying and reopening:
204 retry-method=RRRORRRRRORRRRR meaning to retry the data command 3
205 times, reopen, retry 5 times, etc.
212 Things like blacklisting or whitelisting IP addresses can be done
213 using external wrappers (TCP wrappers, systemd).
215 However it might be nice to have a configurable filter for preventing
216 valid but not sensible requests. The server already filters invalid
217 requests. This would be like seccomp, and could be implemented using
218 an eBPF-based parser. Unfortunately actual eBPF is difficult to use
219 for userspace processes. The "standard" isn't solidly defined - the
220 Linux kernel implementation is the standard - and Linux has by far the
221 best implementation, particularly around bytecode verification and
222 JITting. There is a userspace VM (ubpf) but it has very limited
223 capabilities compared to Linux.
228 Filters allow certain types of composition, but others would not be
229 possible, for example RAIDing over multiple nbd sources. Because the
230 plugin API limits us to loading a single plugin to the server, the
231 best way to do this (and the most robust) is to compose multiple
232 nbdkit processes. Perhaps libnbd will prove useful for this purpose.
237 * Figure out how to get 'make distcheck' working. VPATH builds are
238 working, but various pkg-config results that try to stick
239 bash-completion and ocaml add-ons into their system-wide home do
240 not play nicely with --prefix builds for a non-root user.
247 From time to time we may update the plugin protocol. This section
248 collects ideas for things which might be fixed in the next version of
251 Note that we keep the old protocol(s) around so that source
252 compatibility is retained. Plugins must opt in to the new protocol
253 using ‘#define NBDKIT_API_VERSION <version>’.
255 * All methods taking a ‘count’ field should be uint64_t (instead of
256 uint32_t). Although the NBD protocol does not support 64 bit
257 lengths, it might do in future.
259 * pread could be changed to allow it to support Structured Replies
260 (SRs). This could mean allowing it to return partial data, holes,
261 zeroes, etc. For a client that negotiates SR coupled with a plugin
262 that supports .extents, the v2 protocol would allow us to at least
263 synthesize NBD_REPLY_TYPE_OFFSET_HOLE for less network traffic, even
264 though the plugin will still have to fully populate the .pread
265 buffer; the v3 protocol should make sparse reads more direct.
267 * Parameters should be systematized so that they aren't just (key,
268 value) strings. nbdkit should know the possible keys for the plugin
269 and filters, and the type of the values, and both check and parse
272 * Modify open() API so it takes an export name parameter.
274 * Change config_help from a variable to a function.