Package: linux-image-2.6.26-1-xen-686
Version: 2.6.26-4
Severity: normal

Hi,

I know an Etch host with its stock Xen 3.0.3 hypervisor may be of no
interest, but I hope it still is. After all, a while ago I tested
linux-image-2.6.26-rc8-xen-686_2.6.26~rc8-1~experimental.1~snapshot.11752_i386.deb
and that booted OK with the same config on the same system.

Unfortunately, I can't find any more such images under
http://kernel-archive.buildserver.net/debian-kernel/, so I can't tell
which snapshot was the last workable version. From that pool,
2.6.25-7snapshot.11833 works, 2.6.27~rc5-1~experimental.1~snapshot.12183
crashes very much the same.

Here is the config:

name = "sandbox.aai"
kernel = "/home/wferi/vmlinuz-2.6.26-1-xen-686"
ramdisk = "/home/wferi/initrd.img-2.6.26-1-xen-686"
memory = 300
vif = [ 'mac=00:16:3e:01:00:31, bridge=br894' ]
disk = [ 'phy:/dev/mapper/sandbox.aai,xvda,w' ]
root = "/dev/mapper/sandbox-root ro"
extra = "2"
on_poweroff = 'destroy'
on_reboot = 'restart'
on_crash = 'restart'

And the full console output:

[ 0.000000] Reserving virtual address space above 0xf57fe000
[ 0.000000] Initializing cgroup subsys cpuset
[ 0.000000] Initializing cgroup subsys cpu
[ 0.000000] Linux version 2.6.26-1-xen-686 (Debian 2.6.26-4) (waldi@debian.org) (gcc version 4.1.3 20080623 (prerelease) (Debian 4.1.2-23)) #1 SMP Thu Aug 28 16:22:33 UTC 2008
[ 0.000000] BIOS-provided physical RAM map:
[ 0.000000] Xen: 0000000000000000 - 000000000009f000 (usable)
[ 0.000000] Xen: 0000000000100000 - 0000000012c00000 (usable)
[ 0.000000] 0MB HIGHMEM available.
[ 0.000000] 300MB LOWMEM available.
[ 0.000000] Zone PFN ranges:
[ 0.000000] DMA 0 -> 4096
[ 0.000000] Normal 4096 -> 76800
[ 0.000000] HighMem 76800 -> 76800
[ 0.000000] Movable zone start PFN for each node
[ 0.000000] early_node_map[1] active PFN ranges
[ 0.000000] 0: 0 -> 76800
[ 0.000000] DMI not present or invalid.
[ 0.000000] Allocating PCI resources starting at 20000000 (gap: 12c00000:ed400000)
[ 0.000000] SMP: Allowing 0 CPUs, 0 hotplug CPUs
[ 0.000000] PERCPU: Allocating 42056 bytes of per cpu data
[ 0.000000] Built 1 zonelists in Zone order, mobility grouping on. Total pages: 76200
[ 0.000000] Kernel command line: root=/dev/mapper/sandbox-root ro 2
[ 0.000000] Local APIC disabled by BIOS -- you can enable it with "lapic"
[ 0.000000] Enabling fast FPU save and restore... done.
[ 0.000000] Enabling unmasked SIMD FPU exception support... done.
[ 0.000000] Initializing CPU#0
[ 0.000000] PID hash table entries: 2048 (order: 11, 8192 bytes)
[ 0.000000] Detected 2392.232 MHz processor.
[ 0.004000] Console: colour dummy device 80x25
[ 0.004000] console [tty0] enabled
[ 0.004000] console [hvc0] enabled
[ 0.004000] Dentry cache hash table entries: 65536 (order: 6, 262144 bytes)
[ 0.004000] Inode-cache hash table entries: 32768 (order: 5, 131072 bytes)
[ 0.004000] Memory: 292528k/307200k available (1603k kernel code, 14192k reserved, 635k data, 208k init, 0k highmem)
[ 0.004000] virtual kernel memory layout:
[ 0.004000] fixmap : 0xf5617000 - 0xf57fd000 (1944 kB)
[ 0.004000] pkmap : 0xf5200000 - 0xf5400000 (2048 kB)
[ 0.004000] vmalloc : 0xd3800000 - 0xf51fe000 ( 537 MB)
[ 0.004000] lowmem : 0xc0000000 - 0xd2c00000 ( 300 MB)
[ 0.004000] .init : 0xc0337000 - 0xc036b000 ( 208 kB)
[ 0.004000] .data : 0xc0290d03 - 0xc032fa80 ( 635 kB)
[ 0.004000] .text : 0xc0100000 - 0xc0290d03 (1603 kB)
[ 0.004000] Checking if this processor honours the WP bit even in supervisor mode...Ok.
[ 0.004000] installing Xen timer for CPU 0
[ 0.080011] Calibrating delay using timer specific routine.. 3831.81 BogoMIPS (lpj=7663624)
[ 0.080090] Security Framework initialized
[ 0.080108] SELinux: Disabled at boot.
[ 0.080122] Capability LSM initialized
[ 0.080154] Mount-cache hash table entries: 512
[ 0.080354] Initializing cgroup subsys ns
[ 0.080375] Initializing cgroup subsys cpuacct
[ 0.080388] Initializing cgroup subsys devices
[ 0.080437] CPU: Trace cache: 12K uops, L1 D cache: 8K
[ 0.080457] CPU: L2 cache: 512K
[ 0.080468] CPU: Physical Processor ID: 3
[ 0.080483] Intel machine check architecture supported.
[ 0.080506] Intel machine check reporting enabled on CPU#0.
[ 0.080522] CPU0: Intel P4/Xeon Extended MCE MSRs (12) available
[ 0.080567] general protection fault: 0000 [#1] SMP
[ 0.080587] Modules linked in:
[ 0.080598]
[ 0.080607] Pid: 0, comm: swapper Not tainted (2.6.26-1-xen-686 #1)
[ 0.080625] EIP: e019:[<c010245b>] EFLAGS: 00010202 CPU: 0
[ 0.080644] EIP is at xen_write_cr4+0x2/0x6
[ 0.080657] EAX: 00000620 EBX: 00000020 ECX: c0331f3f EDX: 00000036
[ 0.080673] ESI: c0331fe4 EDI: 00000000 EBP: 00000000 ESP: c0331fc8
[ 0.080690] DS: e021 ES: e021 FS: 00d8 GS: 0000 SS: e021
[ 0.080706] Process swapper (pid: 0, ti=c0330000 task=c0307300 task.ti=c0330000)
[ 0.080725] Stack: c033f4e5 c033eaf7 c033784b c03527a0 003d0000 c033c882 00000000 9fab8171
[ 0.080762] 00004400 0702080b 00000f27 f5800000 00000000 c0b70000
[ 0.080792] Call Trace:
[ 0.080803] [<c033f4e5>] stop_mce+0x21/0x22
[ 0.080819] [<c033eaf7>] alternative_instructions+0xa/0x106
[ 0.080841] [<c033784b>] start_kernel+0x2bc/0x2c4
[ 0.080856] [<c033c882>] xen_start_kernel+0x276/0x27c
[ 0.080874] =======================
[ 0.080883] Code: 01 59 89 f0 5b 5e 5f 5d c3 c3 31 c0 c3 64 8b 15 08 20 36 c0 89 42 08 c3 64 a1 08 20 36 c0 8b 40 08 c3 64 a1 28 20 36 c0 c3 24 6f <0f> 22 e0 c3 64 a1 0c 20 36 c0 c3 64 a3 10 20 36 c0 c3 53 bb 02
[ 0.081059] EIP: [<c010245b>] xen_write_cr4+0x2/0x6 SS:ESP e021:c0331fc8
[ 0.081094] ---[ end trace 4eaa2a86a8e2da22 ]---
[ 0.081108] Kernel panic - not syncing: Attempted to kill the idle task!

Thanks,
Feri.

-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (600, 'stable'), (50, 'unstable')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.26-1-686
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)



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

On Wed, 2008-09-10 at 11:01 +0200, Ferenc Wagner wrote:
> [ 0.080644] EIP is at xen_write_cr4+0x2/0x6

Unfortunately the 3.0.3 hypervisor is stricter in what it will allow CR4
to be set to than necessary and won't allow some of the fields which
newer kernels use. I think we are unlikely to fix this in the kernel in
Debian. You could ask upstream but I'm not sure they would want/be able
to change this behaviour.

Ian.
--
Ian Campbell

13. ... r-q1




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

Ian Campbell <ijc@hellion.org.uk> writes:

> On Wed, 2008-09-10 at 11:01 +0200, Ferenc Wagner wrote:
>> [ 0.080644] EIP is at xen_write_cr4+0x2/0x6
>
> Unfortunately the 3.0.3 hypervisor is stricter in what it will allow CR4
> to be set to than necessary and won't allow some of the fields which
> newer kernels use. I think we are unlikely to fix this in the kernel in
> Debian. You could ask upstream but I'm not sure they would want/be able
> to change this behaviour.

Hi Ian,

thanks for the quick and very informative response! Is it enough if I
upgrade to 3.2.0-3~bpo4+2 or should I go for 3.2.1-2 in Lenny?

Interesting, though, that a CR flag changed going from rc8 to final...
--
Thanks,
Feri.



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

On Wed, 2008-09-10 at 12:57 +0200, Ferenc Wagner wrote:
> Ian Campbell <ijc@hellion.org.uk> writes:
>
> > On Wed, 2008-09-10 at 11:01 +0200, Ferenc Wagner wrote:
> >> [ 0.080644] EIP is at xen_write_cr4+0x2/0x6
> >
> > Unfortunately the 3.0.3 hypervisor is stricter in what it will allow CR4
> > to be set to than necessary and won't allow some of the fields which
> > newer kernels use. I think we are unlikely to fix this in the kernel in
> > Debian. You could ask upstream but I'm not sure they would want/be able
> > to change this behaviour.
>
> Hi Ian,
>
> thanks for the quick and very informative response! Is it enough if I
> upgrade to 3.2.0-3~bpo4+2 or should I go for 3.2.1-2 in Lenny?

I don't recall when the relaxation of cr4 changes happened. Newer is
normally better ( ;-) ) but 3.2.0 is probably worth a go if you have an
aversion to 3.2.1.

> Interesting, though, that a CR flag changed going from rc8 to final...

it was probably side effect of an otherwise innocent looking change...

--
Ian Campbell

BOFH excuse #87:

Password is too complex to decrypt

Hi Ian,

I upgraded to the 3.2 hypervisor from Backports.org, and that indeed
works perfectly. You can close this bug.

Special thanks (to the team) for the suspend and the D-I support!

I seem to recall that I read somewhere a description of Xen support in
Lenny. It stated the lack of dom0 support, and suggested to stay with
Etch in the dom0s. If there'll be a Xen section in Lenny release
notes, we'd rather make sure to point out that this is *not* possible.
--
Regards,
Feri.



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