boot failure in recent cset... dom0 compat issue?

boot failure in recent cset... dom0 compat issue?

[Dan Magenheimer]

Hmmm... somewhere between xen-unstable cset 19043 and 19109
my Intel Weybridge box stops booting dom0.  Is this a known
problem?  I am admittedly trying to boot a 4 months old
3.1-timeframe 2.6.18.8-based dom0.  Were hypervisor changes
made recently that render older dom0's useless? (I realize
this happens sometimes with xen-unstable, just wondering if
this is indeed the problem and, if so, if it was intentional.)

Thanks,
Dan


(XEN) PCI add device 00:03.2
(XEN) PCI remove device 00:03.2
(XEN) PCI add device 00:1f.5
(XEN) PCI remove device 00:1f.5
(XEN) PCI add device 00:1a.7
(XEN) PCI add device 00:1d.7
(XEN) PCI add device 00:1a.0
(XEN) PCI add device 00:1a.1
(XEN) PCI add device 00:1a.2
(XEN) PCI add device 00:1d.0
(XEN) PCI add device 00:1d.1
(XEN) PCI add device 00:1d.2
(XEN) PCI add device 00:1f.2
ata1.00: failed to set xfermode (err_mask=0x104)
ata1: port failed to respond (30 secs)
ata1: COMRESET failed (device not ready)
ata1.00: revalidation failed (errno=-5)
ata1: port failed to respond (30 secs)
ata1: COMRESET failed (device not ready)
ata1.00: revalidation failed (errno=-5)

Re: boot failure in recent cset... dom0 compat issue?

[Keir Fraser]

Shouldn't be a dom0 compat issue. More likely something to do with the Xen
heap removal. Some issues are fixed already but perhaps more are lurking.

 -- Keir

On 29/01/2009 22:12, "Dan Magenheimer" <dan.magenheimer@oracle.com> wrote:

> Hmmm... somewhere between xen-unstable cset 19043 and 19109
> my Intel Weybridge box stops booting dom0.  Is this a known
> problem?  I am admittedly trying to boot a 4 months old
> 3.1-timeframe 2.6.18.8-based dom0.  Were hypervisor changes
> made recently that render older dom0's useless? (I realize
> this happens sometimes with xen-unstable, just wondering if
> this is indeed the problem and, if so, if it was intentional.)
> 
> Thanks,
> Dan
> 
> 
> (XEN) PCI add device 00:03.2
> (XEN) PCI remove device 00:03.2
> (XEN) PCI add device 00:1f.5
> (XEN) PCI remove device 00:1f.5
> (XEN) PCI add device 00:1a.7
> (XEN) PCI add device 00:1d.7
> (XEN) PCI add device 00:1a.0
> (XEN) PCI add device 00:1a.1
> (XEN) PCI add device 00:1a.2
> (XEN) PCI add device 00:1d.0
> (XEN) PCI add device 00:1d.1
> (XEN) PCI add device 00:1d.2
> (XEN) PCI add device 00:1f.2
> ata1.00: failed to set xfermode (err_mask=0x104)
> ata1: port failed to respond (30 secs)
> ata1: COMRESET failed (device not ready)
> ata1.00: revalidation failed (errno=-5)
> ata1: port failed to respond (30 secs)
> ata1: COMRESET failed (device not ready)
> ata1.00: revalidation failed (errno=-5)
> 
> 
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

RE: boot failure in recent cset... dom0 compat issue?

[Dan Magenheimer]

I can confirm that c/s 19064 boots my dom0 fine and
c/s 19065 fails to boot.

> > 3.1-timeframe 2.6.18.8-based dom0.

Sorry, that was a typo, should have been 3.3(.0)-timeframe
2.6.18.8-based dom0, built 2008 Sept 18, probably from whatever
was auto-hg-pull'ed by a Xen build on that day.  (Is there
a way similar to xm info to check the c/s of a dom0?)

My machine is a standard Intel Weybridge box (though
its possible the firmware is old).

> -----Original Message-----
> From: Keir Fraser [mailto:keir.fraser@eu.citrix.com]
> Sent: Thursday, January 29, 2009 3:32 PM
> To: Dan Magenheimer; Xen-Devel (E-mail)
> Subject: Re: [Xen-devel] boot failure in recent cset... dom0 compat
> issue?
> 
> 
> Shouldn't be a dom0 compat issue. More likely something to do 
> with the Xen
> heap removal. Some issues are fixed already but perhaps more 
> are lurking.
> 
>  -- Keir
> 
> On 29/01/2009 22:12, "Dan Magenheimer" 
> <dan.magenheimer@oracle.com> wrote:
> 
> > Hmmm... somewhere between xen-unstable cset 19043 and 19109
> > my Intel Weybridge box stops booting dom0.  Is this a known
> > problem?  I am admittedly trying to boot a 4 months old
> > 3.1-timeframe 2.6.18.8-based dom0.  Were hypervisor changes
> > made recently that render older dom0's useless? (I realize
> > this happens sometimes with xen-unstable, just wondering if
> > this is indeed the problem and, if so, if it was intentional.)
> > 
> > Thanks,
> > Dan
> > 
> > 
> > (XEN) PCI add device 00:03.2
> > (XEN) PCI remove device 00:03.2
> > (XEN) PCI add device 00:1f.5
> > (XEN) PCI remove device 00:1f.5
> > (XEN) PCI add device 00:1a.7
> > (XEN) PCI add device 00:1d.7
> > (XEN) PCI add device 00:1a.0
> > (XEN) PCI add device 00:1a.1
> > (XEN) PCI add device 00:1a.2
> > (XEN) PCI add device 00:1d.0
> > (XEN) PCI add device 00:1d.1
> > (XEN) PCI add device 00:1d.2
> > (XEN) PCI add device 00:1f.2
> > ata1.00: failed to set xfermode (err_mask=0x104)
> > ata1: port failed to respond (30 secs)
> > ata1: COMRESET failed (device not ready)
> > ata1.00: revalidation failed (errno=-5)
> > ata1: port failed to respond (30 secs)
> > ata1: COMRESET failed (device not ready)
> > ata1.00: revalidation failed (errno=-5)
> > 
> > 
> > 
> > 
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xensource.com
> > http://lists.xensource.com/xen-devel
> 
> 
>

RE: boot failure in recent cset... dom0 compat issue?

[Dan Magenheimer]

> Sorry, that was a typo, should have been 3.3(.0)-timeframe
> 2.6.18.8-based dom0, built 2008 Sept 18, probably from whatever
> was auto-hg-pull'ed by a Xen build on that day.  (Is there
> a way similar to xm info to check the c/s of a dom0?)

In case it matters, it appears dom0 was c/s 664 from
http://xenbits.xensource.com/linux-2.6.18-xen.hg
(Found from code examination... is there a better way
to find the c/s of a dom0?)

> -----Original Message-----
> From: Dan Magenheimer 
> Sent: Friday, January 30, 2009 8:40 AM
> To: Keir Fraser; Xen-Devel (E-mail)
> Subject: RE: [Xen-devel] boot failure in recent cset... dom0 compat
> issue?
> 
> 
> I can confirm that c/s 19064 boots my dom0 fine and
> c/s 19065 fails to boot.
> 
> > > 3.1-timeframe 2.6.18.8-based dom0.
> 
> Sorry, that was a typo, should have been 3.3(.0)-timeframe
> 2.6.18.8-based dom0, built 2008 Sept 18, probably from whatever
> was auto-hg-pull'ed by a Xen build on that day.  (Is there
> a way similar to xm info to check the c/s of a dom0?)
> 
> My machine is a standard Intel Weybridge box (though
> its possible the firmware is old).
> 
> > -----Original Message-----
> > From: Keir Fraser [mailto:keir.fraser@eu.citrix.com]
> > Sent: Thursday, January 29, 2009 3:32 PM
> > To: Dan Magenheimer; Xen-Devel (E-mail)
> > Subject: Re: [Xen-devel] boot failure in recent cset... dom0 compat
> > issue?
> > 
> > 
> > Shouldn't be a dom0 compat issue. More likely something to do 
> > with the Xen
> > heap removal. Some issues are fixed already but perhaps more 
> > are lurking.
> > 
> >  -- Keir
> > 
> > On 29/01/2009 22:12, "Dan Magenheimer" 
> > <dan.magenheimer@oracle.com> wrote:
> > 
> > > Hmmm... somewhere between xen-unstable cset 19043 and 19109
> > > my Intel Weybridge box stops booting dom0.  Is this a known
> > > problem?  I am admittedly trying to boot a 4 months old
> > > 3.1-timeframe 2.6.18.8-based dom0.  Were hypervisor changes
> > > made recently that render older dom0's useless? (I realize
> > > this happens sometimes with xen-unstable, just wondering if
> > > this is indeed the problem and, if so, if it was intentional.)
> > > 
> > > Thanks,
> > > Dan
> > > 
> > > 
> > > (XEN) PCI add device 00:03.2
> > > (XEN) PCI remove device 00:03.2
> > > (XEN) PCI add device 00:1f.5
> > > (XEN) PCI remove device 00:1f.5
> > > (XEN) PCI add device 00:1a.7
> > > (XEN) PCI add device 00:1d.7
> > > (XEN) PCI add device 00:1a.0
> > > (XEN) PCI add device 00:1a.1
> > > (XEN) PCI add device 00:1a.2
> > > (XEN) PCI add device 00:1d.0
> > > (XEN) PCI add device 00:1d.1
> > > (XEN) PCI add device 00:1d.2
> > > (XEN) PCI add device 00:1f.2
> > > ata1.00: failed to set xfermode (err_mask=0x104)
> > > ata1: port failed to respond (30 secs)
> > > ata1: COMRESET failed (device not ready)
> > > ata1.00: revalidation failed (errno=-5)
> > > ata1: port failed to respond (30 secs)
> > > ata1: COMRESET failed (device not ready)
> > > ata1.00: revalidation failed (errno=-5)
> > > 
> > > 
> > > 
> > > 
> > > _______________________________________________
> > > Xen-devel mailing list
> > > Xen-devel@lists.xensource.com
> > > http://lists.xensource.com/xen-devel
> > 
> > 
> >
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
>

Re: boot failure in recent cset... dom0 compat issue?

[Keir Fraser]

On 30/01/2009 15:39, "Dan Magenheimer" <dan.magenheimer@oracle.com> wrote:

> I can confirm that c/s 19064 boots my dom0 fine and
> c/s 19065 fails to boot.
> 
>>> 3.1-timeframe 2.6.18.8-based dom0.
> 
> Sorry, that was a typo, should have been 3.3(.0)-timeframe
> 2.6.18.8-based dom0, built 2008 Sept 18, probably from whatever
> was auto-hg-pull'ed by a Xen build on that day.  (Is there
> a way similar to xm info to check the c/s of a dom0?)
> 
> My machine is a standard Intel Weybridge box (though
> its possible the firmware is old).

Do you yet know whether an up-to-date dom0 kernel works for you?

 -- Keir

Re: boot failure in recent cset... dom0 compat issue?

[Keir Fraser]

On 30/01/2009 15:51, "Dan Magenheimer" <dan.magenheimer@oracle.com> wrote:

>> Sorry, that was a typo, should have been 3.3(.0)-timeframe
>> 2.6.18.8-based dom0, built 2008 Sept 18, probably from whatever
>> was auto-hg-pull'ed by a Xen build on that day.  (Is there
>> a way similar to xm info to check the c/s of a dom0?)
> 
> In case it matters, it appears dom0 was c/s 664 from
> http://xenbits.xensource.com/linux-2.6.18-xen.hg
> (Found from code examination... is there a better way
> to find the c/s of a dom0?)

The cset doesn't get embedded in the kernel anywhere, unfortunately.

 -- Keir

RE: boot failure in recent cset... dom0 compat issue?

[Dan Magenheimer]

> > I can confirm that c/s 19064 boots my dom0 fine and
> > c/s 19065 fails to boot.
> > 
> >>> 3.1-timeframe 2.6.18.8-based dom0.
> > 
> > Sorry, that was a typo, should have been 3.3(.0)-timeframe
> > 2.6.18.8-based dom0, built 2008 Sept 18, probably from whatever
> > was auto-hg-pull'ed by a Xen build on that day.  (Is there
> > a way similar to xm info to check the c/s of a dom0?)
> > 
> > My machine is a standard Intel Weybridge box (though
> > its possible the firmware is old).
> 
> Do you yet know whether an up-to-date dom0 kernel works for you?

linux-2.6.18-xen.hg tip is c/s 786

xen-19043 boots dom0linux-644 fine
xen-19109 fails to boot dom0linux-644
xen-19064 boots dom0linux-644 fine
xen-19065 fails to boot dom0linux-644
xen-19043 boots dom0linux-786 but I get no networking
xen-19109 boots dom0linux-786 but I get no networking

Has something changed in tools that might need to
be matched against latest xen and/or latest dom0, which
might explain my networking disappearance?  I'd
upgrade tools too, but that's harder to back out if
it doesn't work, and I'm not keen to reinstall this
machine from scratch right now.

Re: boot failure in recent cset... dom0 compat issue?

[Keir Fraser]

On 30/01/2009 17:56, "Dan Magenheimer" <dan.magenheimer@oracle.com> wrote:

> linux-2.6.18-xen.hg tip is c/s 786
> 
> xen-19043 boots dom0linux-644 fine
> xen-19109 fails to boot dom0linux-644
> xen-19064 boots dom0linux-644 fine
> xen-19065 fails to boot dom0linux-644
> xen-19043 boots dom0linux-786 but I get no networking
> xen-19109 boots dom0linux-786 but I get no networking
> 
> Has something changed in tools that might need to
> be matched against latest xen and/or latest dom0, which
> might explain my networking disappearance?  I'd
> upgrade tools too, but that's harder to back out if
> it doesn't work, and I'm not keen to reinstall this
> machine from scratch right now.

I don't think there have been any intentional interface breakages there just
now. So I don't really have an easy solution.

 -- Keir

RE: boot failure in recent cset... dom0 compat issue?

[Dan Magenheimer]

> <dan.magenheimer@oracle.com> wrote:
> 
> > linux-2.6.18-xen.hg tip is c/s 786
> > 
> > xen-19043 boots dom0linux-644 fine
> > xen-19109 fails to boot dom0linux-644
> > xen-19064 boots dom0linux-644 fine
> > xen-19065 fails to boot dom0linux-644
> > xen-19043 boots dom0linux-786 but I get no networking
> > xen-19109 boots dom0linux-786 but I get no networking
> > 
> > Has something changed in tools that might need to
> > be matched against latest xen and/or latest dom0, which
> > might explain my networking disappearance?  I'd
> > upgrade tools too, but that's harder to back out if
> > it doesn't work, and I'm not keen to reinstall this
> > machine from scratch right now.
> 
> I don't think there have been any intentional interface 
> breakages there just
> now. So I don't really have an easy solution.
> 
>  -- Keir

Apologies, the networking problem was a red herring...
a NIC driver issue that I had forgotten needs to be
fixed every time I build a new dom0 kernel.

So, yes, a matched hypervisor/dom0 tip set works
for me now.

So it appears, at least on some machines, that c/s
19065 renders dom0 no longer backwards-compatible.
No big deal I guess but probably worth noting in
the release notes for the 3.4 release.

Dan