[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH 09/12] evtchn: move FIFO-private struct declarations



Hi Jan,

On 29/09/2020 13:49, Jan Beulich wrote:
On 29.09.2020 14:26, Julien Grall wrote:
On 28/09/2020 12:00, Jan Beulich wrote:
There's no need to expose them.

We are going to need them for LiveUpdate and Non-cooperative Live
Migration as the save/restore is happening outside of event_fifo.c.

This is because we tried to keep all the save/restore code in a separate
directory.

I'm afraid I don't view this as a reason for the change not to be made
right now. If, when, and in what shape LU will hit upstream is unknown
at the moment (unless I've missed some indication of a time line).

I was merely pointing out long term use case as I vaguely remember a discussion in the past to avoid short term change.

In
fact, if we didn't expose things like the ones here to far too wide an
"audience", I wonder whether ... >
Although, I could also see pros for scatter save/restore across the code
base.

... you wouldn't have chosen this route anyway, just to avoid exposing
items widely that are supposed to be (almost) private.

I would still chose that route because it helps to keep all the save/restore code in one place. FWIW, this follows what HVM context does today.

Anyway, as I said, this I can see pros/cons for each. So at least this discussion clarify the preferred approach.

For the patch itself:

Acked-by: Julien Grall <jgrall@xxxxxxxxxx>

Cheers,

--
Julien Grall



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.