I am running Xen packaged with CentOS 5.4 and trying to install Ubuntu 10.04
as an HVM guest.
xen_major : 3
xen_minor : 1
xen_extra : .2-164.11.1.el5
After sprinkling the kernel with printks, I am seeing that it is stuck in
the function detect_extended_topology() in the following loop:
<snip>
sub_index = 1;
do {
cpuid_count(0xb, sub_index, &eax, &ebx, &ecx,
&edx);
printk("%s: after cpuid_count %d\n", __FUNCTION__,
sub_index);
/*
* Check for the Core type in the implemented sub leaves.
*/
if (LEAFB_SUBTYPE(ecx) == CORE_TYPE) {
core_level_siblings = LEVEL_MAX_SIBLINGS(ebx);
core_plus_mask_width = BITS_SHIFT_NEXT_LEVEL(eax);
break;
}
sub_index++;
} while (LEAFB_SUBTYPE(ecx) != INVALID_TYPE);
<snip>
The ECX leaf subtype never returns CORE_TYPE or INVALID_TYPE. So think I
might be running in to a bug / quirk in the CPUID handling code in Xen
packaged with CentOS 5.4. Is there a work around for it? Maybe specifying
something in the cpuid option in the config file.
Thanks!
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
David Markey
2010-Jun-09 18:03 UTC
Re: [Xen-users] Ubuntu 10.04 stuck in detect_extended_topology()
Is there a reason you cant use the PV installer? On 9 June 2010 19:01, AP Xen <apxeng@gmail.com> wrote:> I am running Xen packaged with CentOS 5.4 and trying to install Ubuntu > 10.04 as an HVM guest. > > > > xen_major : 3 > > xen_minor : 1 > > xen_extra : .2-164.11.1.el5 > > > > After sprinkling the kernel with printks, I am seeing that it is stuck in > the function detect_extended_topology() in the following loop: > > <snip> > > sub_index = 1; > > do { > > cpuid_count(0xb, sub_index, &eax, &ebx, &ecx, &edx); > > printk("%s: after cpuid_count %d\n", __FUNCTION__, > sub_index); > > > > /* > > * Check for the Core type in the implemented sub leaves. > > */ > > if (LEAFB_SUBTYPE(ecx) == CORE_TYPE) { > > core_level_siblings = LEVEL_MAX_SIBLINGS(ebx); > > core_plus_mask_width = BITS_SHIFT_NEXT_LEVEL(eax); > > break; > > } > > > > sub_index++; > > } while (LEAFB_SUBTYPE(ecx) != INVALID_TYPE); > > <snip> > > > > The ECX leaf subtype never returns CORE_TYPE or INVALID_TYPE. So think I > might be running in to a bug / quirk in the CPUID handling code in Xen > packaged with CentOS 5.4. Is there a work around for it? Maybe specifying > something in the cpuid option in the config file. > > > > Thanks! > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
AP Xen
2010-Jun-09 18:13 UTC
RE: [Xen-users] Ubuntu 10.04 stuck in detect_extended_topology()
I have standardized on HVM and really don''t want to go the PV route.
Alternatively, are you saying, do the install using PV and then switch to
HVM?
From: David Markey [mailto:admin@dmarkey.com]
Sent: Wednesday, June 09, 2010 11:03 AM
To: AP Xen
Cc: xen-users@lists.xensource.com
Subject: Re: [Xen-users] Ubuntu 10.04 stuck in detect_extended_topology()
Is there a reason you cant use the PV installer?
On 9 June 2010 19:01, AP Xen <apxeng@gmail.com> wrote:
I am running Xen packaged with CentOS 5.4 and trying to install Ubuntu 10.04
as an HVM guest.
xen_major : 3
xen_minor : 1
xen_extra : .2-164.11.1.el5
After sprinkling the kernel with printks, I am seeing that it is stuck in
the function detect_extended_topology() in the following loop:
<snip>
sub_index = 1;
do {
cpuid_count(0xb, sub_index, &eax, &ebx, &ecx,
&edx);
printk("%s: after cpuid_count %d\n", __FUNCTION__,
sub_index);
/*
* Check for the Core type in the implemented sub leaves.
*/
if (LEAFB_SUBTYPE(ecx) == CORE_TYPE) {
core_level_siblings = LEVEL_MAX_SIBLINGS(ebx);
core_plus_mask_width = BITS_SHIFT_NEXT_LEVEL(eax);
break;
}
sub_index++;
} while (LEAFB_SUBTYPE(ecx) != INVALID_TYPE);
<snip>
The ECX leaf subtype never returns CORE_TYPE or INVALID_TYPE. So think I
might be running in to a bug / quirk in the CPUID handling code in Xen
packaged with CentOS 5.4. Is there a work around for it? Maybe specifying
something in the cpuid option in the config file.
Thanks!
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users