[Spice-devel] [RFC PATCH v2 00/20] Monitor ID rework

[Lukáš Hrázký]

On Mon, 2018-08-27 at 14:27 +0200, Gerd Hoffmann wrote:
>   Hi,
> 
> > 1. The IDs sent from the client in VDAgentMonitorsConfig and
> > MousePosition messages are equal to either `channel_id + monitor_id` or
> > `channel_id ? channel_id : monitor_id`. This is under the assumption
> > that there is either only one display_channel or more display channels
> > each with only a single monitor. Under this assumption the
> > aforementioned expressions are equivalent. It is not a reliable way to
> > identify monitors though.
> 
> Can we define this to be "channel_id + monitor_id", then use sparse
> channel ids?  qxl has room for up to 16 monitors in the protocol
> messages.  channel_id is 8 bit.  So we could support up to 16 display
> devices with up to 16 outputs each if we enumerate the display channels
> 0, 16, 32 instead of 0, 1, 2.

While I suppose technically possible, a couple issues:

1. It is AFAICS not backwards compatible anyway (you don't change the
protocol itself, but you change the content):

channel_id | monitor_id | old display_id | your display_id
----------------------------------------------------------
         0 |          0 |              0 |               0
         1 |          0 |              1 |              16
         2 |          0 |              2 |              32
...

2. I don't want to do this, while it's better than simple channel_id +
monitor_id, it's still unnecessary to encode the numbers this way, two
separate fields are much cleaner. (and with it not saving us
compatibility trouble...)

(on a side note, I'm still assimilating the information you shared in
the other emails and thinking about ways to move forward, that's why
I'm slow on the replies...)

Thanks,
Lukas

> cheers,
>   Gerd
>