libvirt/docs/internals/rpc.html.in
Daniel P. Berrange 977ba05973 Add some docs about the RPC protocol and APIs
* remote.html.in: Remove obsolete notes about internals of the
  RPC protocol
* internals/rpc.html.in: Extensive docs on RPC protocol/API
* sitemap.html.in: Add new page
2011-08-12 13:54:10 +01:00

876 lines
34 KiB
HTML

<html>
<body>
<h1>libvirt RPC infrastructure</h1>
<ul id="toc"></ul>
<p>
libvirt includes a basic protocol and code to implement
an extensible, secure client/server RPC service. This was
originally designed for communication between the libvirt
client library and the libvirtd daemon, but the code is
now isolated to allow reuse in other areas of libvirt code.
This document provides an overview of the protocol and
structure / operation of the internal RPC library APIs.
</p>
<h2><a name="protocol">RPC protocol</a></h2>
<p>
libvirt uses a simple, variable length, packet based RPC protocol.
All structured data within packets is encoded using the
<a href="http://en.wikipedia.org/wiki/External_Data_Representation">XDR standard</a>
as currently defined by <a href="https://tools.ietf.org/html/rfc4506">RFC 4506</a>.
On any connection running the RPC protocol, there can be multiple
programs active, each supporting one or more versions. A program
defines a set of procedures that it supports. The procedures can
support call+reply method invocation, asynchronous events,
and generic data streams. Method invocations can be overlapped,
so waiting for a reply to one will not block the receipt of the
reply to another outstanding method. The protocol was loosely
inspired by the design of SunRPC. The definition of the RPC
protocol is in the file <code>src/rpc/virnetprotocol.x</code>
in the libvirt source tree.
</p>
<h3><a href="protocolframing">Packet framing</a></h3>
<p>
On the wire, there is no explicit packet framing marker. Instead
each packet is preceded by an unsigned 32-bit integer giving
the total length of the packet in bytes. This length includes
the 4-bytes of the length word itself. Conceptually the framing
looks like this:
</p>
<pre>
|~~~ Packet 1 ~~~|~~~ Packet 2 ~~~|~~~ Packet 3 ~~~|~~~
+-------+------------+-------+------------+-------+------------+...
| n=U32 | (n-4) * U8 | n=U32 | (n-4) * U8 | n=U32 | (n-4) * U8 |
+-------+------------+-------+------------+-------+------------+...
|~ Len ~|~ Data ~|~ Len ~|~ Data ~|~ Len ~|~ Data ~|~
</pre>
<h3><a href="protocoldata">Packet data</a></h3>
<p>
The data in each packet is split into two parts, a short
fixed length header, followed by a variable length payload.
So a packet from the illustration above is more correctly
shown as
</p>
<pre>
+-------+-------------+---------------....---+
| n=U32 | 6*U32 | (n-(7*4))*U8 |
+-------+-------------+---------------....---+
|~ Len ~|~ Header ~|~ Payload .... ~|
</pre>
<h3><a href="protocolheader">Packet header</a></h3>
<p>
The header contains 6 fields, encoded as signed/unsigned 32-bit
integers.
</p>
<pre>
+---------------+
| program=U32 |
+---------------+
| version=U32 |
+---------------+
| procedure=S32 |
+---------------+
| type=S32 |
+---------------+
| serial=U32 |
+---------------+
| status=S32 |
+---------------+
</pre>
<dl>
<dt><code>program</code></dt>
<dd>
This is an arbitrarily chosen number that will uniquely
identify the "service" running over the stream.
</dd>
<dt><code>version</code></dt>
<dd>
This is the version number of the program, by convention
starting from '1'. When an incompatible change is made
to a program, the version number is incremented. Ideally
both versions will then be supported on the wire in
parallel for backwards compatibility.
</dd>
<dt><code>procedure</code></dt>
<dd>
This is an arbitrarily chosen number that will uniquely
identify the method call, or event associated with the
packet. By convention, procedure numbers start from 1
and are assigned monotonically thereafter.
</dd>
<dt><code>type</code></dt>
<dd>
<p>
This can be one of the following enumeration values
</p>
<ol>
<li>call: invocation of a method call</li>
<li>reply: completion of a method call</li>
<li>event: an asynchronous event</li>
<li>stream: control info or data from a stream</li>
</ol>
</dd>
<dt><code>serial</code></dt>
<dd>
This is an number that starts from 1 and increases
each time a method call packet is sent. A reply or
stream packet will have a serial number matching the
original method call packet serial. Events always
have the serial number set to 0.
</dd>
<dt><code>status</code></dt>
<dd>
<p>
This can one of the following enumeration values
</p>
<ol>
<li>ok: a normal packet. this is always set for method calls or events.
For replies it indicates successful completion of the method. For
streams it indicates confirmation of the end of file on the stream.</li>
<li>error: for replies this indicates that the method call failed
and error information is being returned. For streams this indicates
that not all data was sent and the stream has aborted</li>
<li>continue: for streams this indicates that further data packets
will be following</li>
</ol>
</dl>
<h3><a href="protocolpayload">Packet payload</a></h3>
<p>
The payload of a packet will vary depending on the <code>type</code>
and <code>status</code> fields from the header.
</p>
<ul>
<li>type=call: the in parameters for the method call, XDR encoded</li>
<li>type=reply+status=ok: the return value and/or out parameters for the method call, XDR encoded</li>
<li>type=reply+status=error: the error information for the method, a virErrorPtr XDR encoded</li>
<li>type=event: the parameters for the event, XDR encoded</li>
<li>type=stream+status=ok: no payload</li>
<li>type=stream+status=error: the error information for the method, a virErrorPtr XDR encoded</li>
<li>type=stream+status=continue: the raw bytes of data for the stream. No XDR encoding</li>
</ul>
<p>
For the exact payload information for each procedure, consult the XDR protocol
definition for the program+version in question
</p>
<h3><a name="wireexamples">Wire examples</a></h3>
<p>
The following diagrams illustrate some example packet exchanges
between a client and server
</p>
<h4><a name="wireexamplescall">Method call</a></h4>
<p>
A single method call and successful
reply, for a program=8, version=1, procedure=3, which 10 bytes worth
of input args, and 4 bytes worth of return values. The overall input
packet length is 4 + 24 + 10 == 38, and output packet length 32
</p>
<pre>
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 1 | 0 | .o.oOo.o. | --&gt; S (call)
+--+-----------------------+-----------+
+--+-----------------------+--------+
C &lt;-- |32| 8 | 1 | 3 | 1 | 1 | 0 | .o.oOo | &lt;-- S (reply)
+--+-----------------------+--------+
</pre>
<h4><a name="wireexamplescallerr">Method call with error</a></h4>
<p>
An unsuccessful method call will instead return an error object
</p>
<pre>
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 1 | 0 | .o.oOo.o. | --&gt; S (call)
+--+-----------------------+-----------+
+--+-----------------------+--------------------------+
C &lt;-- |48| 8 | 1 | 3 | 2 | 1 | 0 | .o.oOo.o.oOo.o.oOo.o.oOo | &lt;-- S (error)
+--+-----------------------+--------------------------+
</pre>
<h4><a name="wireexamplescallup">Method call with upload stream</a></h4>
<p>
A method call which also involves uploading some data over
a stream will result in
</p>
<pre>
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 1 | 0 | .o.oOo.o. | --&gt; S (call)
+--+-----------------------+-----------+
+--+-----------------------+--------+
C &lt;-- |32| 8 | 1 | 3 | 1 | 1 | 0 | .o.oOo | &lt;-- S (reply)
+--+-----------------------+--------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
...
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+
C --&gt; |24| 8 | 1 | 3 | 3 | 1 | 0 | --&gt; S (stream finish)
+--+-----------------------+
+--+-----------------------+
C &lt;-- |24| 8 | 1 | 3 | 3 | 1 | 0 | &lt;-- S (stream finish)
+--+-----------------------+
</pre>
<h4><a name="wireexamplescallbi">Method call bidirectional stream</a></h4>
<p>
A method call which also involves a bi-directional stream will
result in
</p>
<pre>
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 1 | 0 | .o.oOo.o. | --&gt; S (call)
+--+-----------------------+-----------+
+--+-----------------------+--------+
C &lt;-- |32| 8 | 1 | 3 | 1 | 1 | 0 | .o.oOo | &lt;-- S (reply)
+--+-----------------------+--------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C &lt;-- |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | &lt;-- S (stream data down)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C &lt;-- |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | &lt;-- S (stream data down)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C &lt;-- |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | &lt;-- S (stream data down)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C &lt;-- |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | &lt;-- S (stream data down)
+--+-----------------------+-------------....-------+
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
..
+--+-----------------------+-------------....-------+
C --&gt; |38| 8 | 1 | 3 | 3 | 1 | 2 | .o.oOo.o.oOo....o.oOo. | --&gt; S (stream data up)
+--+-----------------------+-------------....-------+
+--+-----------------------+
C --&gt; |24| 8 | 1 | 3 | 3 | 1 | 0 | --&gt; S (stream finish)
+--+-----------------------+
+--+-----------------------+
C &lt;-- |24| 8 | 1 | 3 | 3 | 1 | 0 | &lt;-- S (stream finish)
+--+-----------------------+
</pre>
<h4><a name="wireexamplescallmany">Method calls overlapping</a></h4>
<pre>
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 1 | 0 | .o.oOo.o. | --&gt; S (call 1)
+--+-----------------------+-----------+
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 2 | 0 | .o.oOo.o. | --&gt; S (call 2)
+--+-----------------------+-----------+
+--+-----------------------+--------+
C &lt;-- |32| 8 | 1 | 3 | 1 | 2 | 0 | .o.oOo | &lt;-- S (reply 2)
+--+-----------------------+--------+
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 3 | 0 | .o.oOo.o. | --&gt; S (call 3)
+--+-----------------------+-----------+
+--+-----------------------+--------+
C &lt;-- |32| 8 | 1 | 3 | 1 | 3 | 0 | .o.oOo | &lt;-- S (reply 3)
+--+-----------------------+--------+
+--+-----------------------+-----------+
C --&gt; |38| 8 | 1 | 3 | 0 | 4 | 0 | .o.oOo.o. | --&gt; S (call 4)
+--+-----------------------+-----------+
+--+-----------------------+--------+
C &lt;-- |32| 8 | 1 | 3 | 1 | 1 | 0 | .o.oOo | &lt;-- S (reply 1)
+--+-----------------------+--------+
+--+-----------------------+--------+
C &lt;-- |32| 8 | 1 | 3 | 1 | 4 | 0 | .o.oOo | &lt;-- S (reply 4)
+--+-----------------------+--------+
</pre>
<h2><a name="security">RPC security</a></h2>
<p>
There are various things to consider to ensure an implementation
of the RPC protocol can be satisfactorily secured
</p>
<h3><a name="securitytls">Authentication/encryption</a></h3>
<p>
The basic RPC protocol does not define or require any specific
authentication/encryption capabilities. A generic solution to
providing encryption for the protocol is to run the protocol
over a TLS encrypted data stream. x509 certificate checks can
be done to form a crude authentication mechanism. It is also
possible for an RPC program to negotiate an encryption /
authentication capability, such as SASL, which may then also
provide per-packet data encryption. Finally the protocol data
stream can of course be tunnelled over transports such as SSH.
</p>
<h3><a name="securitylimits">Data limits</a></h3>
<p>
Although the protocol itself defines many arbitrary sized data values in the
payloads, to avoid denial of service attack there are a number of size limit
checks prior to encoding or decoding data. There is a limit on the maximum
size of a single RPC message, limit on the maximum string length, and limits
on any other parameter which uses a variable length array. These limits can
be raised, subject to agreement between client/server, without otherwise
breaking compatibility of the RPC data on the wire.
</p>
<h3><a name="securityvalidate">Data validation</a></h3>
<p>
It is important that all data be fully validated before performing
any actions based on the data. When reading an RPC packet, the
first four bytes must be read and the max packet size limit validated,
before any attempt is made to read the variable length packet data.
After a complete packet has been read, the header must be decoded
and all 6 fields fully validated, before attempting to dispatch
the payload. Once dispatched, the payload can be decoded and passed
onto the appropriate API for execution. The RPC code must not take
any action based on the payload, since it has no way to validate
the semantics of the payload data. It must delegate this to the
execution API (e.g. corresponding libvirt public API).
</p>
<h2><a name="internals">RPC internal APIs</a></h2>
<p>
The generic internal RPC library code lives in the <code>src/rpc/</code>
directory of the libvirt source tree. Unless otherwise noted, the
objects are all threadsafe. The core object types and their
purposes are:
</p>
<h3><a name="apioverview">Overview of RPC objects</a></h3>
<p>
The following is a high level overview of the role of each
of the main RPC objects
</p>
<dl>
<dt><code>virNetSASLContextPtr</code> (virnetsaslcontext.h)</dt>
<dd>The virNetSASLContext APIs maintain SASL state for a network
service (server or client). This is primarily used on the server
to provide a whitelist of allowed SASL usernames for clients.
</dd>
<dt><code>virNetSASLSessionPtr</code> (virnetsaslcontext.h)</dt>
<dd>The virNetSASLSession APIs maintain SASL state for a single
network connection (socket). This is used to perform the multi-step
SASL handshake and perform encryption/decryption of data once
authenticated, via integration with virNetSocket.
</dd>
<dt><code>virNetTLSContextPtr</code> (virnettlscontext.h)</dt>
<dd>The virNetTLSContext APIs maintain TLS state for a network
service (server or client). This is primarily used on the server
to provide a whitelist of allowed x509 distinguished names, as
well as diffie-hellman keys. It can also do validation of
x509 certificates prior to initiating a connection, in order
to improve detection of configuration errors.
</dd>
<dt><code>virNetTLSSessionPtr</code> (virnettlscontext.h)</dt>
<dd>The virNetTLSSession APIs maintain TLS state for a single
network connection (socket). This is used to perform the multi-step
TLS handshake and perform encryption/decryption of data once
authenticated, via integration with virNetSocket.
</dd>
<dt><code>virNetSocketPtr</code> (virnetsocket.h)</dt>
<dd>The virNetSocket APIs provide a higher level wrapper around
the raw BSD sockets and getaddrinfo APIs. They allow for creation
of both server and client sockets. Data transports supported are
TCP, UNIX, SSH tunnel or external command tunnel. Internally the
TCP socket impl uses the getaddrinfo info APIs to ensure correct
protocol-independent behaviour, thus supporting both IPv4 and IPv6.
The socket APIs can be associated with a virNetSASLSessionPtr or
virNetTLSSessionPtr object to allow seamless encryption/decryption
of all writes and reads. For UNIX sockets it is possible to obtain
the remote client user ID and process ID. Integration with the
libvirt event loop also allows use of callbacks for notification
of various I/O conditions
</dd>
<dt><code>virNetMessagePtr</code> (virnetmessage.h)</dt>
<dd>The virNetMessage APIs provide a wrapper around the libxdr
API calls, to facilitate processing and creation of RPC
packets. There are convenience APIs for encoding/encoding the
packet headers, encoding/decoding the payload using an XDR
filter, encoding/decoding a raw payload (for streams), and
encoding a virErrorPtr object. There is also a means to
add to/serve from a linked-list queue of messages.</dd>
<dt><code>virNetClientPtr</code> (virnetclient.h)</dt>
<dd>The virNetClient APIs provide a way to connect to a
remote server and run one or more RPC protocols over
the connection. Connections can be made over TCP, UNIX
sockets, SSH tunnels, or external command tunnels. There
is support for both TLS and SASL session encryption.
The client also supports management of multiple data streams
over each connection. Each client object can be used from
multiple threads concurrently, with method calls/replies
being interleaved on the wire as required.
</dd>
<dt><code>virNetClientProgramPtr</code> (virnetclientprogram.h)</dt>
<dd>The virNetClientProgram APIs are used to register a
program+version with the connection. This then enables
invocation of method calls, receipt of asynchronous
events and use of data streams, within that program+version.
When created a set of callbacks must be supplied to take
care of dispatching any incoming asynchronous events.
</dd>
<dt><code>virNetClientStreamPtr</code> (virnetclientstream.h)</dt>
<dd>The virNetClientStream APIs are used to control transmission and
receipt of data over a stream active on a client. Streams provide
a low latency, unlimited length, bi-directional raw data exchange
mechanism layered over the RPC connection
</dd>
<dt><code>virNetServerPtr</code> (virnetserver.h)</dt>
<dd>The virNetServer APIs are used to manage a network server. A
server exposed one or more programs, over one or more services.
It manages multiple client connections invoking multiple RPC
calls in parallel, with dispatch across multiple worker threads.
</dd>
<dt><code>virNetServerMDNSPtr</code> (virnetservermdns.h)</dt>
<dd>The virNetServerMDNS APIs are used to advertise a server
across the local network, enabling clients to automatically
detect the existence of remote services. This is done by
interfacing with the Avahi mDNS advertisement service.
</dd>
<dt><code>virNetServerClientPtr</code> (virnetserverclient.h)</dt>
<dd>The virNetServerClient APIs are used to manage I/O related
to a single client network connection. It handles initial
validation and routing of incoming RPC packets, and transmission
of outgoing packets.
</dd>
<dt><code>virNetServerProgramPtr</code> (virnetserverprogram.h)</dt>
<dd>The virNetServerProgram APIs are used to provide the implementation
of a single program/version set. Primarily this includes a set of
callbacks used to actually invoke the APIs corresponding to
program procedure numbers. It is responsible for all the serialization
of payloads to/from XDR.</dd>
<dt><code>virNetServerServicePtr</code> (virnetserverservice.h)</dt>
<dd>The virNetServerService APIs are used to connect the server to
one or more network protocols. A single service may involve multiple
sockets (ie both IPv4 and IPv6). A service also has an associated
authentication policy for incoming clients.
</dd>
</dl>
<h3><a name="apiclientdispatch">Client RPC dispatch</a></h3>
<p>
The client RPC code must allow for multiple overlapping RPC method
calls to be invoked, transmission and receipt of data for multiple
streams and receipt of asynchronous events. Understandably this
involves coordination of multiple threads.
</p>
<p>
The core requirement in the client dispatch code is that only
one thread is allowed to be performing I/O on the socket at
any time. This thread is said to be "holding the buck". When
any other thread comes along and needs to do I/O it must place
its packets on a queue and delegate processing of them to the
thread that has the buck. This thread will send out the method
call, and if it sees a reply will pass it back to the waiting
thread. If the other thread's reply hasn't arrived, by the time
the main thread has got its own reply, then it will transfer
responsibility for I/O to the thread that has been waiting the
longest. It is said to be "passing the buck" for I/O.
</p>
<p>
When no thread is performing any RPC method call, or sending
stream data there is still a need to monitor the socket for
incoming I/O related to asynchronous events, or stream data
receipt. For this task, a watch is registered with the event
loop which triggers whenever the socket is readable. This
watch is automatically disabled whenever any other thread
grabs the buck, and re-enabled when the buck is released.
</p>
<h4><a name="apiclientdispatchex1">Example with buck passing</a></h4>
<p>
In the first example, a second thread issues a API call
while the first thread holds the buck. The reply to the
first call arrives first, so the buck is passed to the
second thread.
</p>
<pre>
Thread-1
|
V
Call API1()
|
V
Grab Buck
| Thread-2
V |
Send method1 V
| Call API2()
V |
Wait I/O V
|&lt;--------Queue method2
V |
Send method2 V
| Wait for buck
V |
Wait I/O |
| |
V |
Recv reply1 |
| |
V |
Pass the buck-----&gt;|
| V
V Wait I/O
Return API1() |
V
Recv reply2
|
V
Release the buck
|
V
Return API2()
</pre>
<h4><a name="apiclientdispatchex2">Example without buck passing</a></h4>
<p>
In this second example, a second thread issues an API call
which is sent and replied to, before the first thread's
API call has completed. The first thread thus notifies
the second that its reply is ready, and there is no need
to pass the buck
</p>
<pre>
Thread-1
|
V
Call API1()
|
V
Grab Buck
| Thread-2
V |
Send method1 V
| Call API2()
V |
Wait I/O V
|&lt;--------Queue method2
V |
Send method2 V
| Wait for buck
V |
Wait I/O |
| |
V |
Recv reply2 |
| |
V |
Notify reply2------&gt;|
| V
V Return API2()
Wait I/O
|
V
Recv reply1
|
V
Release the buck
|
V
Return API1()
</pre>
<h4><a name="apiclientdispatchex3">Example with async events</a></h4>
<p>
In this example, only one thread is present and it has to
deal with some async events arriving. The events are actually
dispatched to the application from the event loop thread
</p>
<pre>
Thread-1
|
V
Call API1()
|
V
Grab Buck
|
V
Send method1
|
V
Wait I/O
| Event thread
V ...
Recv event1 |
| V
V Wait for timer/fd
Queue event1 |
| V
V Timer fires
Wait I/O |
| V
V Emit event1
Recv reply1 |
| V
V Wait for timer/fd
Return API1() |
...
</pre>
<h3><a name="apiserverdispatch">Server RPC dispatch</a></h3>
<p>
The RPC server code must support receipt of incoming RPC requests from
multiple client connections, and parallel processing of all RPC
requests, even many from a single client. This goal is achieved through
a combination of event driven I/O, and multiple processing threads.
</p>
<p>
The main libvirt event loop thread is responsible for performing all
socket I/O. It will read incoming packets from clients and willl
transmit outgoing packets to clients. It will handle the I/O to/from
streams associated with client API calls. When doing client I/O it
will also pass the data through any applicable encryption layer
(through use of the virNetSocket / virNetTLSSession and virNetSASLSession
integration). What is paramount is that the event loop thread never
do any task that can take a non-trivial amount of time.
</p>
<p>
When reading packets, the event loop will first read the 4 byte length
word. This is validated to make sure it does not exceed the maximum
permissible packet size, and the client is set to allow receipt of the
rest of the packet data. Once a complete packet has been received, the
next step is to decode the RPC header. The header is validated to
ensure the request is sensible, ie the server should not receive a
method reply from a client. If the client has not yet authenticated,
a security check is also applied to make sure the procedure is on the
whitelist of those allowed prior to auth. If the packet is a method
call, it will be placed on a global processing queue. The event loop
thread is now done with the packet for the time being.
</p>
<p>
The server has a pool of worker threads, which wait for method call
packets to be queued. One of them will grab the new method call off
the queue for processing. The first step is to decode the payload of
the packet to extract the method call arguments. The worker does not
attempt to do any semantic validation of the arguments, except to make
sure the size of any variable length fields is below defined limits.
</p>
<p>
The worker now invokes the libvirt API call that corresponds to the
procedure number in the packet header. The worker is thus kept busy
until the API call completes. The implementation of the API call
is responsible for doing semantic validation of parameters and any
MAC security checks on the objects affected.
</p>
<p>
Once the API call has completed, the worker thread will take the
return value and output parameters, or error object and encode
them into a reply packet. Again it does not attempt to do any
semantic validation of output data, aside from variable length
field limit checks. The worker thread puts the reply packet onto
the transmission queue for the client. The worker is now finished
and goes back to wait for another incoming method call.
</p>
<p>
The main event loop is back in charge and when the client socket
becomes writable, it will start sending the method reply packet
back to the client.
</p>
<p>
At any time the libvirt connection object can emit asynchronous
events. These are handled by callbacks in the main event thread.
The callback will simply encode the event parameters into a new
data packet and place the packet on the client transmission
queue.
</p>
<p>
Incoming and outgoing stream packets are also directly handled
by the main event thread. When an incoming stream packet is
received, instead of placing it in the global dispatch queue
for the worker threads, it is sidetracked into a per-stream
processing queue. When the stream becomes writable, queued
incoming stream packets will be processed, passing their data
payload onto the stream. Conversely when the stream becomes
readable, chunks of data will be read from it, encoded into
new outgoing packets, and placed on the client's transmit
queue
</p>
<h4><a name="apiserverdispatchex1">Example with overlapping methods</a></h4>
<p>
This example illustrates processing of two incoming methods with
overlapping execution
</p>
<pre>
Event thread Worker 1 Worker 2
| | |
V V V
Wait I/O Wait Job Wait Job
| | |
V | |
Recv method1 | |
| | |
V | |
Queue method1 V |
| Serve method1 |
V | |
Wait I/O V |
| Call API1() |
V | |
Recv method2 | |
| | |
V | |
Queue method2 | V
| | Serve method2
V V |
Wait I/O Return API1() V
| | Call API2()
| V |
V Queue reply1 |
Send reply1 | |
| V V
V Wait Job Return API2()
Wait I/O | |
| ... V
V Queue reply2
Send reply2 |
| V
V Wait Job
Wait I/O |
| ...
...
</pre>
<h4><a name="apiserverdispatchex2">Example with stream data</a></h4>
<p>
This example illustrates processing of stream data
</p>
<pre>
Event thread
|
V
Wait I/O
|
V
Recv stream1
|
V
Queue stream1
|
V
Wait I/O
|
V
Recv stream2
|
V
Queue stream2
|
V
Wait I/O
|
V
Write stream1
|
V
Write stream2
|
V
Wait I/O
|
...
</pre>
</body>
</html>