On Mon, Feb 08, 2021 at 11:04:12AM +0000, Jonathan Cameron wrote:
On Fri, 5 Feb 2021 15:53:55 +0000 Lorenzo Pieralisi lorenzo.pieralisi@arm.com wrote:
On Fri, Feb 05, 2021 at 11:53:08AM +0000, Jonathan Cameron wrote:
On Fri, 5 Feb 2021 09:11:59 +0800 Jammy Zhou jammy.zhou@linaro.org wrote:
On Fri, 5 Feb 2021 at 02:00, Lorenzo Pieralisi lorenzo.pieralisi@arm.com wrote:
On Thu, Feb 04, 2021 at 08:51:07AM +0800, Jammy Zhou wrote:
On Wed, 3 Feb 2021 at 21:55, Lorenzo Pieralisi <
lorenzo.pieralisi@arm.com>
wrote:
On Wed, Feb 03, 2021 at 01:30:33AM +0000, Jammy Zhou via Linaro-open-discussions wrote: > I have added these two topics to the agenda for meeting on Feb22nd.
We should probably move it to an afternoon slot GMT otherwise I don't think anyone in the US can attend (I know that an afternoon slot GMT is problematic for people in Asia, don't know how to solve this, please let me know).If we need to cover EU, US and Asia, I think the evening time for Asia
is the
only option, like 2pm(GMT)/10pm(GMT+8)/6am(GMT-8).
Or maybe we need to have two slots, one is Asia friendly and the other
is US
friendly.
Any preference?
I am UK based so the slot above is perfectly fine but others may disagree. We can have two separate calls, one for virt CPU hotplug and one for CXL, not sure though we have two completely disjoint sets of attendees.
Hmm, it is difficult. Why not list the required key attendees first for each topic? Then we can select a proper time according to their time zones.
Whilst others will of course be interested, from Huawei side I can cover CXL
- worth noting that I expect we are basically talking CXL 2.0 here not
earlier. The stuff currently going on is focused on spec verification and preenablement for 2.0 (not a whole lot of hardware available yet - if there were we wouldn't be actively scrubbing the spec whilst implementing it)
1.1 and earlier more or less just look like normal devices or memory + some horrible impdef stuff in firmware. Might be room for some collaboration in EDK2 - unclear though. There will be device specific stuff, but a lot of that is impdef (how to do bias switching etc).
I'm guessing I'll do the intro slides for this one unless someone else wants to do the summary of current status / open questions bit? - more than happy to collaborate on it as well!
So for CXL Me + Huawei UK team members(GMT) If we got firmware heavy - perhaps not on first call on this topic - Henson (GMT + 8) Others from Huawei in China probably interested, though if time isn't great probably not 'that interested' at this stage.
From other members we clearly have Lorenzo + ARM team (GMT + others?) Jon Masters (GMT - 5 I think)
Whilst I don't mind anyone from other interested companies (e.g. Intel) showing up, I suspect that would make limited sense at a first discussion of this topic?
CPU HP like to be more interesting to my China based colleagues alongside the UK team. So if we do go with a late slot for China rather than split meeting, let's do that topic first.
As to whether we want to split the meeting, that depends on whether we expect a detailed discussion on this occasion, or more of an update on direction followed by watch this space. Lorenzo any idea yet on which category we might be in?
I hope and work towards a virt CPU hotplug session that provides a finalized way forward (that is, what to do to upstream it rather than discussing how CPU hotplug should be implemented), which means I expect it to be fairly short - that's why I thought that CXL can be part of the same slot.
Perhaps we could do them as a unified call (late time slot for China) on the basis that we can have a follow up call on vCPU HP if it appears necessary (after people have had time to digest what you share etc).
Would that work for everyone?
It does for me, thanks a lot.
Lorenzo