[Fedora-xen] Simple testing results for Xen and VMX on FC5-test2.

[Stephen C. Tweedie]

Hi,

On Fri, 2006-01-27 at 11:03 +0800, You, Yongkang wrote:

> I have done a simple checking for Xen and VMX on FC5-test2. Basically, I can create VMX on FC5-test2 IA32. But FC5-test2 IA32e (x86-64) doesn't provide Xen and Xen0 kernel, so didn't do Xen and VMX testing for it.
> 
> Beside of testing Xen and VMX, I also found some issues about installation, which I have updated the bug #179078 as a NB. The following is the testing results only for Xen and VMX.
> =========================
> 1. *** IA32E hasn't RPMS about Xen hypervisor and Xen0/XenU kernels. __FAIL__

It's coming.  I hope.  The upstream x86_64 stuff from linux-2.6-merge
doesn't actually compile at the moment (though that's easy to fix), and
once compiled it fails in different, compiler-dependent ways on some of
my hardware.  I've got a minimally stable boot now on a different box,
but the upstream base in this case is proving hard to work with.  

> 2. *** In order to boot up IA32 Xen0, I have to add pci=noacpi as the boot 
> parameters of Xen0 kernel, or system will hang when initial PCI. I don't know if it is only related to my Machine. 

It's certainly hardware-dependent, but you're not the only person
reporting the need for such options.  Please open a bugzilla so that we
can track this one properly, as we'd really like to automate such
options as much as possible for the main release.

One thing that would be worth knowing is whether the main, non-xen
rawhide SMP kernel boots on this box.  The UP kernel disables a bunch of
advanced hardware enablement for compatibility purposes, but the SMP one
should match Xen as closely as possible.

> 8. Couldn't get Xen Changeset by "xm info" or "xm dmesg".

Not sure how easy that would be to add, since we don't build it out of
Mercurial directly but rather out of a tarballed copy from the CVS build
system.  I'm being careful to add the changeset info to the changelog in
the rpm, though.

Thanks,
 Stephen