Finishing the network protocol

Wed Feb 23 05:52:24 PST 2011

On 02/23/2011 07:26 AM, Andreas Hartmetz wrote:
> On Wednesday 23 February 2011 00:31:56 Marty Jack wrote:
>> On 02/22/2011 01:16 PM, Andreas Hartmetz wrote:
>>> Hello,
>>>
>>> I've started a Wayland implementation currently called "area", written in
>>> C++ and with the main goal to work on all hardware that currently works
>>> on Linux in some way (Framebuffer or X11).
>>> I've started with the network code and noticed a few things that still
>>> look like prototype code in Wayland, probably unchanged from early
>>> versions.
>>> Well, Wayland is becoming a serious project, so I think we should start
>>> finishing and fixing the protocol. Additionally I don't know if I will
>>> finish my own project so I'm looking to contribute my findings to the
>>> main project.
>>>
>>> First some hopefully correct primer of the Wayland protocol for
>>> interested onlookers.
>>> - The Wayland protocol is a remote procedure call protocol of sorts. All
>>>
>>>   messages are exchanged between objects; the protocol is asynchronous
>>>   and no methods have return values as such.
>>>   Methods can "return values" by triggering a message back.
>>>   Everything is asynchronous, but order of messages is preserved.
>>>
>>> - Each protocol-level object exists on both client and server
>>> - It's easiest to think of all objects being created by the server on
>>>
>>>   behalf of the clients.
>>>
>>> - Object constructors don't return anything, they have an object ID
>>>
>>>   argument that is pre-chosen and can later be used to refer to the
>>>   created object. If creation succeeded there is no message back.
>>>
>>> - The conversation between client and server is bootstrapped by creating
>>>
>>>   the "display" object on each client, which then starts talking to the
>>>   global display object on the server.
>>>
>>> I see the following issues:
>>> - There is no way to subscribe to events, or rather there is no way not
>>>
>>>   to subscribe to all events.
>>>
>>> - Range-based ID issuance for object IDs (obviously can't use pointers
>>>
>>>   between processes) is not bulletproof. It is possible for ranges to
>>>   become fragmented insofar that they can't be reclaimed because there's
>>>   one ID in every range. There is also currently no code that tries to
>>>   reclaim ranges.
>>>   The practical implication is that a Wayland server can, by design, not
>>>   run indefinitely without exhausting ID (range)s.
>>>   Another kind-of-problem is that a client can interfere with another
>>>   client's operation by, intentionally or not, using IDs belonging to
>>>   the other client.
>>>
>>> I've looked at the TODO and come up with a few ideas of my own for the
>>> following suggestions to modify the protocol:
>>>
>>> - Have one ID<->object map per client, except for global objects where
>>>
>>>   there is a global map in the server.
>>>   This is suggested in the TODO file; I've done it this way right away.
>>>   Obviously each client will have its own map in the client process
>>>   anyway.
>>>
>>> - Have three ID ranges:
>>>    a) for global objects
>>>    b) for client-specific objects created by the server (do they exist?)
>>>    c) for client-specific objects created by the client
>>>   
>>>   where "created by" really means "creation initiated (assigning an
>>>   object ID) by".
>>>
>>> - Handle subscription without an extra mechanism by creating or not
>>>
>>>   creating the object that will receive the desired events. Might need
>>>   some splitting of existing objects.
>>>   This would IMHO be an elegant and minimal way to handle the matter.
>>>
>>> - A scheme to recycle object IDs. When a new ID is needed, pick a free
>>>
>>>   one at random. This introduces a problem:
>>>   Suppose the client destroys object A with ID n, then by chance
>>>   immediately reuses ID n for object B.
>>>   The server will only receive this information later, the Wayland
>>>   protocol being asynchronous and the server not having to respond to an
>>>   object creation request, unless it goes wrong. In the meantime the
>>>   server could send an event intended for A which would end up at B,
>>>   causing Bad Things to happen - in my implementation most likely an
>>>   assert failure unless the objects are of the same class.
>>>   (This is the trickiest failure mode I could think of)
>>>   The suggested solution is a kind of "rendezvous" for objects where this
>>>   can happen, or for simplicity all objects:
>>>   On both client and server, have a function that needs to be called
>>>   twice to unregister an object ID.
>>>   One call from the destructor of the local object when it destroys
>>>   itself, one call from the remote counterpart object when it destroys
>>>   itself. No matter in which order the method is called, the first call
>>>   removes the ID<->object mapping and puts the object ID on a waiting
>>>   list to avoid reuse. The second call removes the ID from the waiting
>>>   list, making it free to reuse.
>>>
>>> - Specify how parent-child relationships work, e.g. (bad example, the
>>>
>>>   answer is probably no here) is a surface automatically destroyed when
>>>   its screen goes away? By whom?
>>>
>>> - Specify who gets do delete objects and how that looks in the protocol
>>>
>>>   - this is probably more a matter of documentation; I didn't read all
>>>   of Wayland's code carefully and implementers ideally shouldn't have to.
>>>
>>> - Add information in the protocol description XML file about things like
>>>
>>>   an object being global or not, and basically everything mentioned above
>>>   that can benefit from help from the code generator.
>>>
>>> I'm not publishing a repository URL right now because I haven't chosen a
>>> license yet and because I've copied over the wayland.xml protocol
>>> description file that bears no license header. Kristian, what about the
>>> license of that file?
>>> If there is interest I can polish my code a bit and publish an URL.
>>>
>>> Cheers,
>>> Andreas
>>> _______________________________________________
>>> wayland-devel mailing list
>>> wayland-devel at lists.freedesktop.org
>>> http://lists.freedesktop.org/mailman/listinfo/wayland-devel
>>
>> I haven't analyzed this in any detail but maybe there is a simple unique-ID
>> encoding along the lines of partitioning the encoding space into "created
>> by client", "created by server" and then a wraparound counter sufficiently
>> big, say 28 bits of counter and 4 bits of who allocated the value, that a
>> collision would require the connection to be up for an insanely long time
>> and the risk of collision is low (but would need to be checked for).
>>
>> Usually the only time you want a random value is when there is some
>> security risk involved in being able to predict one.
>>
> Using sequential IDs is fine for most aspects, but it doesn't prevent 
> collisions. Without some list of "dangerous" IDs it can always happen that you 
> get one that was just freed, however. Saying that wraparound is not allowed 
> makes the protocol "mostly reliable", which should be avoided IMHO. Even if it 
> never causes problems it's always something you have to think of when 
> debugging;  you can't completetely forget about lower-level details.
> Picking an ID that is not in use on both sides with complete (not 99.9999995%) 
> reliability unfortunately takes a little more work.
> Maybe somebody has a great idea how to do it with less effort?
> Sequential 64-bit integers maybe? Adds quite a few nines to the likeliness 
> that no collision occurs, taking a little more space and time...
> _______________________________________________
> wayland-devel mailing list
> wayland-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/wayland-devel
> 

No, I disagree about the possibility of getting one that was just freed, assuming you buy into the namespace partitioning.

If you have a variable in the client that holds the last one allocated, and you increment it, you are guaranteed not to get the same value for client-allocated IDs until the counter wraps around.  If you partition the ID space, you are guaranteed not to get the same value as any other ID allocator can allocate.  Assuming you have a data structure that maps the ID back to your structure corresponding to the ID, which I think you would need anyway, and you are about to put your pointer into that structure, it is easy to check if the ID is used and retry the allocation.

I would be tempted to use a multilevel page table for the aforementioned data structure the same way page frame data bases work.  That way you don't have a lot of wasted pointer space; you can clean up the second (and third, if you wanted) level page table when it becomes free after those objects are deleted and it won't be needed again until the ID wraps.

Also, if you put the "who allocated this" in the lowest order bits, and increment just the upper bits, this is a very easy way to do it, most computers these days have add instructions that will handle the overflow for you.

	static uint32_t id;	/* WhoAmI in 0:3; sequential counter in 4:31 */
	Initialize:  id = WhoAmI;
	Allocate:  id += 1 << 4;

(This idea comes long ago from a professor at CMU who wanted to put a protection ring ID in those low order bits)