lkml.org 
[lkml]   [2010]   [Mar]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [Xen-devel] [PATCH 7/7] xen: Enable event channel of PV extension of HVM
On Fri, 12 Mar 2010, Sheng Yang wrote:
> On Thursday 11 March 2010 19:43:00 Stefano Stabellini wrote:
> > On Thu, 11 Mar 2010, Sheng Yang wrote:
> > > On Wednesday 10 March 2010 23:29:30 Stefano Stabellini wrote:
> > > > On Wed, 10 Mar 2010, Sheng Yang wrote:
> > > > > I think we can leave the controversial thing later. At least, we want
> > > > > a framework for PV extension of HVM. We can work together to
> > > > > determine what is the better way for evtchn, as well as porting
> > > > > pirqs. (And the later MSI work may also depends on it)
> > > >
> > > > Most of my patch series can be upstreamed right now, the only thing
> > > > that needs some extra work is the pirq remapping.
> > > > So yes, we can upstream the rest and that would also make PV on HVM
> > > > work on linux upstream ASAP.
> > > > I think that is a worthy goal by itself, therefore I am going to send
> > > > another reduced patch series, ready to be upstreamed, without pirq
> > > > remappings, with support for PV on HVM.
> > >
> > > I think my first 6 patches can be used without modification. And I've
> > > already spent lots of efforts on them with Jeremy.
> > >
> > > We can work on evtchn and platform pci later.
> >
> > Agreed on evtchn but platform pci should be present even in the basic
> > version.
> > Please read the series "basic PV on HVM support" I posted yesterday
> > because I believe has already everything needed to be accepted upstream
> > right now; at the same time it introduces the pv clocksource and the
> > vector based callback mechanism that is going to be needed for the
> > interrupt remapping we are going to work on later on.
> >
> IIRC, I think your platform PCI driver also depends on evtchn to inject
> interrupt?
>

The original platform PCI driver used to register an interrupt based
callback mechanism but in the last patch series I sent I removed it in
favor of the new vector based callback you are familiar with.
However I am currently working on providing the former GSI based method
as a fallback, otherwise PV on HVM won't work on xen < 4.



\
 
 \ /
  Last update: 2010-03-12 11:45    [W:0.070 / U:0.044 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site