On Mon, Mar 31, 2008 at 09:11:38AM +1100, Russell Coker wrote:
> I think it's reasonable to expect a Lenny kernel to work in the same manner as
> an Etch kernel and to consider any failure to do so to be a bug.

Well, this is not possible in this case because they added code to
hijack other devices _and_ proposed the usage by default with the
knowledge that Linus won't ever accept this.

> The above is what I was using. I added the "console=hvc0" bit as you
> suggested (why is that needed?) and it displayed some console messages,

Because there is not code which overrides the default console. Fedora
have a patch, maybe I'll pull them in for now.

> mount: Mounting /dev/xvda on /root failed: No such device

Contents of /dev? /sys/block?

Bastian

--
"Get back to your stations!"
"We're beaming down to the planet, sir."
-- Kirk and Mr. Leslie, "This Side of Paradise",
stardate 3417.3



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Mon, Mar 31, 2008 at 09:26:32PM +0200, Bastian Blank wrote:
> Contents of /dev? /sys/block?

Oh, and xenstore-ls /local/domain/$domid.

Bastian

--
A little suffering is good for the soul.
-- Kirk, "The Corbomite Maneuver", stardate 1514.0



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Tuesday 01 April 2008 06:26, Bastian Blank wrote:
> On Mon, Mar 31, 2008 at 09:11:38AM +1100, Russell Coker wrote:
> > I think it's reasonable to expect a Lenny kernel to work in the same
> > manner as an Etch kernel and to consider any failure to do so to be a
> > bug.
>
> Well, this is not possible in this case because they added code to
> hijack other devices _and_ proposed the usage by default with the
> knowledge that Linus won't ever accept this.

Do we have people with enough spare time to fix this foolishness?

> > The above is what I was using. I added the "console=hvc0" bit as you
> > suggested (why is that needed?) and it displayed some console messages,
>
> Because there is not code which overrides the default console. Fedora
> have a patch, maybe I'll pull them in for now.

Fedora also messes things up. /dev/tty0 doesn't work for a getty with a
Fedora/RHEL/CentOS kernel in a DomU.

> > mount: Mounting /dev/xvda on /root failed: No such device
>
> Contents of /dev? /sys/block?

(initramfs) ls -l /sys/block/xvda
-r--r--r-- 1 0 0 4096 Mar 31 21:55 capability
-r--r--r-- 1 0 0 4096 Mar 31 21:55 dev
lrwxrwxrwx 1 0 0 0 Mar 31 21:55
device -> ../../devices/vbd-51712
drwxr-xr-x 2 0 0 0 Mar 31 21:55 holders
drwxr-xr-x 3 0 0 0 Mar 31 21:55 queue
-r--r--r-- 1 0 0 4096 Mar 31 21:55 range
-r--r--r-- 1 0 0 4096 Mar 31 21:55 removable
-r--r--r-- 1 0 0 4096 Mar 31 21:55 size
drwxr-xr-x 2 0 0 0 Mar 31 21:55 slaves
-r--r--r-- 1 0 0 4096 Mar 31 21:55 stat
lrwxrwxrwx 1 0 0 0 Mar 31 21:55
subsystem -> ../../block
--w------- 1 0 0 4096 Mar 31 21:55 uevent
(initramfs) ls -l /dev/xvda
brw-rw---- 1 0 0 202, 0 Mar 31 21:55 /dev/xvda

xenstore-ls /local/domain/$domid gives the following:
vm = "/vm/5787f4aa-a61e-c6b7-ce50-6a507cc9233c"
device = ""
vbd = ""
51712 = ""
virtual-device = "51712"
device-type = "disk"
protocol = "x86_32-abi"
backend-id = "0"
state = "4"
backend = "/local/domain/0/backend/vbd/23/51712"
ring-ref = "8"
event-channel = "6"
vif = ""
0 = ""
mac = "00:16:3e:66:66:69"
handle = "0"
protocol = "x86_32-abi"
backend-id = "0"
state = "4"
backend = "/local/domain/0/backend/vif/23/0"
tx-ring-ref = "768"
rx-ring-ref = "769"
event-channel = "7"
request-rx-copy = "1"
feature-rx-notify = "1"
feature-sg = "1"
feature-gso-tcpv4 = "1"
console = ""
0 = ""
protocol = "x86_32-abi"
state = "1"
backend-id = "0"
backend = "/local/domain/0/backend/console/23/0"
device-misc = ""
vif = ""
nextDeviceID = "1"
console = ""
nextDeviceID = "1"
console = ""
ring-ref = "77958"
port = "2"
limit = "1048576"
type = "xenconsoled"
tty = "/dev/pts/2"
image = ""
entry = "3224522752"
pae-mode = "yes"
loader = "generic"
virt-base = "3221225472"
features = ""
pae-pgdir-above-4gb = "1"
writable-page-tables = "0"
xen-version = "xen-3.0"
guest-os = "linux"
hypercall-page = "3224731648"
guest-version = "2.6"
store = ""
ring-ref = "77959"
port = "1"
cpu = ""
0 = ""
availability = "online"
control = ""
platform-feature-multiprocessor-suspend = "1"
name = "play-test"
domid = "23"
memory = ""
target = "102400"
serial = ""
0 = ""
tty = "/dev/pts/2"



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Tue, Apr 01, 2008 at 09:00:57AM +1100, Russell Coker wrote:
> Fedora also messes things up. /dev/tty0 doesn't work for a getty with a
> Fedora/RHEL/CentOS kernel in a DomU.

/dev/tty0 is a virtual console. A DomU have no virtual consoles except
with a framebuffer.

> > > mount: Mounting /dev/xvda on /root failed: No such device
> (initramfs) ls -l /dev/xvda
> brw-rw---- 1 0 0 202, 0 Mar 31 21:55 /dev/xvda

Okay, so this is the old race condition. Use rootdelay=X for now.

Bastian

--
A little suffering is good for the soul.
-- Kirk, "The Corbomite Maneuver", stardate 1514.0



--
To UNSUBSCRIBE, email to debian-bugs-dist-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org