Re: [MirageOS-devel] NTP client integration with MirageOS / Mini-OS

[Thomas Leonard <talex5@xxxxxxxxx>]

On 19 August 2016 at 16:40, Hannes Mehnert <hannes@xxxxxxxxxxx> wrote:
> Hey,
>
> On 27/07/2016 17:07, Mindy wrote:
>> On 07/27/2016 02:05 PM, Matthew Gray wrote:
>>> Should MCLOCK also implement the DEVICE signature? I think so, for
>>> consistency.
>>
>> Making PCLOCK be a device looks completely reasonable to me, and I see
>> no reason not to make MCLOCK be one as well.  Thoughts?
>
> I can see value in making PCLOCK a device, but I fail to see any reason
> to make MCLOCK a device.  I don't expect MCLOCK to have internal state,
> neither to be able to error, nor that anyone would ever want to
> disconnect from it.
>
> For PCLOCK it makes sense since timezone and DST (plus adjustment to a
> real clock) may be needed, but MCLOCK is some arbitrary, always
> increasing, tick.  PCLOCK might then also need some adjust_t : t ->
> offset -> unit function, or how do people expect the timezone to be set?
>
> Or is there something I'm missing regarding MCLOCK?
>
> Sorry to not have responded earlier here (and I can see that some amount
> of implementation regarding this has already been done).

MCLOCK might have internal state for testing, although a global would
work there too if necessary.

While we're fixing all the clock APIs, how about moving sleep into
CLOCK/PCLOCK/MCLOCK?
It seems strange to have TIME and CLOCK be separate devices.

(and maybe have `CLOCK.at time` rather than or as well as `sleep`)


-- 
talex5 (GitHub/Twitter)        http://roscidus.com/blog/
GPG: 5DD5 8D70 899C 454A 966D  6A51 7513 3C8F 94F6 E0CC
GPG: DA98 25AE CAD0 8975 7CDA  BD8E 0713 3F96 CA74 D8BA

_______________________________________________
MirageOS-devel mailing list
MirageOS-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/cgi-bin/mailman/listinfo/mirageos-devel