lkml.org 
[lkml]   [2010]   [Feb]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [Xen-devel] [PATCH 6/6] xen/hybrid: Enable grant table and xenbus
Date
On Wednesday 03 February 2010 00:24:04 Ian Campbell wrote:
> On Tue, 2010-02-02 at 13:24 +0000, Sheng Yang wrote:
> > I am not sure if I understand you right, but I think the issue is,
> > there is no PVonHVM drivers in Linux upstream. The drivers are
> > currently maintained by OSVs, and the one in Xen upstream code only
> > support 2.6.18. So I didn't take them into consideration at the time.
>
> True, but this is something which should be taken care of by the core
> Xen-aware code not something which should be pushed down into each
> driver.
>
> Someone who wants to add PVonHVM functionality shouldn't have to go and
> remove a bunch of conditionals from each driver (or worse add
> alternative clauses to each check!).
>
> > I think the "xen_evtchn_enable()" looks much better. Would replace
> > these ugly lines in the next version.
>
> I think it would be cleaner to encapsulate this in the evtchn code
> rather than leaking platform knowledge into each driver. IOW the evtchn
> functions should return failure if event channels are not enabled and
> the driver should cope with this gracefully.

Agree. That what I suppose to do. What the drivers should only know is, if
event channel is enabled.

> Or perhaps at the xenbus driver level we should be deciding whether or
> not we have enough paravirtualisation to be worth probing the drivers at
> all?

I think current scheme is direct enough for now. We can improve it later.

--
regards
Yang, Sheng


\
 
 \ /
  Last update: 2010-02-02 17:51    [W:0.948 / U:0.076 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site