Aron,
Seems we have two method, maybe we can use xen-unstable.hg to generate
xenlinux patch as first choice or talk to redhat side and persuade them pulling
CSet from xen-unstable.hg just at it syncing with xen-ia64-unstable.hg. If so,
it will reduce unnecessary extra work more :)
Just FYI.
Thanks & Best Regards
-Xiantao
OTC,Intel Corporation
> -----Original Message-----
> From: Aron Griffis [mailto:aron@xxxxxx]
> Sent: 2006年6月23日 21:34
> To: Zhang, Xiantao
> Cc: Yu, Fenghua; Yang, Fred; Tian, Kevin; You, Yongkang
> Subject: Re: How to make xen0 and xen with
> http://free.linux.hp.com/~agriffis/fedora-kernel-rpm
>
> Xiantao,
>
> Thank you for tracking this down! Unfortunately we can't simply
> rebase the patch because it is made by Juan from linux-2.6.tip-xen
> rather than directly from xen-unstable... i.e. we do not own the
> patch, and it is constantly being remade without our intervention.
>
> Because of the method Red Hat uses to generate this patch, this
> problem will continue to occur. I will bring up the issue today,
> because I think we need to find a long-term solution to this problem.
>
> Thanks again for your help.
>
> Regards,
> Aron
>
> -----Original Message-----
> From: Zhang, Xiantao [mailto:xiantao.zhang@xxxxxxxxx]
> Sent: Friday, June 23, 2006 09:28 PM
> Subject: RE: How to make xen0 and xen with
> http://free.linux.hp.com/~agriffis/fedora-kernel-rpm
>
>
> Hi Fenghua/Aron
> This issue should be caused by mismatch of xen and xenlinux's Cset.
> Xen's
> version should be xen-unstable.hg's Cset 10314, but xenlinux not based
> xen-unstalbe.hg's 10314 and seems that it based xen-ia64-unstable.hg's Cset
> 10314. If so, it will cause unpredictable aftermath. The obvious phenomena is
> that one page of xen0's kernel code was flushed and triggered breakfault at
> last. Seems we should rebase linux-2.6-xen.patch to xen-unstalble's Cset 10314
> first?
> Thanks & Best Regards
> -Xiantao
>
> OTC,Intel Corporation
>
> > -----Original Message-----
> > From: Yu, Fenghua > Sent: 2006年6月23日 10:50
> > To: Aron Griffis; Zhang, Xiantao
> > Cc: Yang, Fred; Tian, Kevin; You, Yongkang
> > Subject: RE: How to make xen0 and xen with
> > http://free.linux.hp.com/~agriffis/fedora-kernel-rpm
> >
> >
> > Hi, Xiantao,
> >
> > Could you try to reproduce the issue and fix it today? Please update us the
> > status then we can follow up. I think I saw this kind of issue (invalid
> > hypercalls) at the very beginning. The issue was gone after rebuilding the
> > xen/kernel rpms or updating xen/kernel rpms.
> >
> > Thanks.
> >
> > -Fenghua
> >
> > -----Original Message-----
> > From: Aron Griffis [mailto:aron@xxxxxx]
> > Sent: Thursday, June 22, 2006 7:09 PM
> > To: Zhang, Xiantao
> > Cc: Yang, Fred; Tian, Kevin; You, Yongkang; Yu, Fenghua
> > Subject: Re: How to make xen0 and xen with
> > http://free.linux.hp.com/~agriffis/fedora-kernel-rpm
> >
> > Well, this is mixed news, but it is all I have at the
> > moment...
> >
> > I updated fedora-kernel-ia64 today. It builds and you can
> > find the new rpms in my yum repository.
> >
> > However it doesn't boot. The kernel attempts to make
> > invalid hypercalls. I have not looked closely yet and don't
> > know the root cause of the problem.
> >
> > If you would be willing to test and debug those rpms, that
> > would be fantastic. Even though it doesn't boot at the
> > moment, I am going to create an RFC patchset for Red Hat to
> > get comments for the moment. I plan to send that out
> > tomorrow morning.
> >
> > Best regards,
> > Aron
> >
> > Zhang, Xiantao wrote: [Thu Jun 22 2006, 09:03:33PM EDT]
> > > Oh, seems I had some misunderstandings here. OK, Wish your good news. :)
> > > Thanks & Best Regards
> > > -Xiantao
> > >
> > > OTC,Intel Corporation
> > >
> > > > -----Original Message-----
> > > > From: Aron Griffis [mailto:aron@xxxxxx]
> > > > Sent: 2006年6月23日 1:26
> > > > To: Zhang, Xiantao
> > > > Cc: Yang, Fred; Tian, Kevin; You, Yongkang; Yu, Fenghua
> > > > Subject: Re: How to make xen0 and xen with
> > > > http://free.linux.hp.com/~agriffis/fedora-kernel-rpm
> > > >
> > > > Hi Xiantao,
> > > >
> > > > Zhang, Xiantao wrote: [Thu Jun 22 2006, 01:33:08AM EDT]
> > > > > We have noticed that you used this following tree to generate
|
