Re: [Qemu-devel] [PATCH v2] vnc: added initial websockets support

[Tim Hardeck]

On 11/20/2012 10:47 AM, Daniel P. Berrange wrote:
> On Tue, Nov 20, 2012 at 10:21:58AM +0100, Tim Hardeck wrote:
>> This patch adds basic Websockets version 13 - RFC 6455 - support to QEMU
>> VNC. Binary encoding support on the client side is required.
>>
>> Because of the GnuTLS requirement the Websockets implementation is
>> optional (--enable-vnc-ws).
>>
>> To activate Websockets during runtime the VNC option "websocket" is
>> used, for example "-vnc :0,websocket" would activate Websockets.
>> The port for Websockets connections is (5700 + display) so if QEMU VNC
>> is started with :0 the websockets port would be 5700.
> 
> We need to be able to specify an explicit port number for the websockets
> listen address, separately from the main VNC port number. This automatic
> pick of websockets port might be nice for a user starting QEMU manually,
> but for management apps we need full direct control.
Ok, this should be no problem to add something like websocket=<port> but
I just thought that a correlation between the vnc and websocket port
would be quite useful especially when several QEMU instances are run on
one machine.
Would it be Ok to keep the correlation if no port is specified?

I am going to add the websocket port option after some more feedback.

>> Changes to v1
>> * removed automatic websocket recognition
>> * added new lwebsock socket on port 5700 + display when the vnc option
>>   "websocket" is passed on
>> * adapted vnc_connect vnc_listen_read to differ between websocket
>> * added separate event handler to read the Websocket handshake
>>
>> Would it be Ok to use a public domain SHA1 implementation like
>> tests/tcg/sha1.c and if so where should the sha1.c be stored?
>> Without the GnuTLS dependency would it be Ok to make Websockets not
>> optional because this would clean up the patch/code quite a bit?
> 
> IMHO using gcrypt/GNUTLS is a good thing. Creating our own copies of
> encryption algorithms in source tree causes significant complications
> getting QEMU security certified. GNUTLS is a common enough crypto
> library that I don't think it is unreasonably onerous to expect it to
> be used for WebSockets. It even works fine on Windows.
That's one of the reasons why I have used GnuTLS but it is quite a huge
dependency for just one algorithm and the many ifdefs don't really help
the code quality.

Regards
Tim

-- 
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix
Imendörffer, HRB 16746 (AG Nürnberg)
Maxfeldstr. 5, 90409 Nürnberg, Germany
T: +49 (0) 911 74053-0  F: +49 (0) 911 74053-483
http://www.suse.de/

signature.asc
Description: OpenPGP digital signature