Xen devel - Feb 2011 - [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!

Same problem here,

what is the actual state of this bug?

Is there any workaround for "older" kernels like mine
(2.6.32-5-xen-amd64) on debian-squeeze?

BTW if I start the server with the same kernel (without xen) the problem
does not exist.

 

Since Giam gave us a lot of relevant informations on this bug, I don''t
know which details 

I could post here to help further on this.

 

regards,

Guido

 



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Mon, Feb 21, 2011 at 12:02:55PM +0100, Guido Hecken
wrote:
>    Same problem here,
> 
>    what is the actual state of this bug?
> 
>    Is there any workaround for "older" kernels like mine
(2.6.32-5-xen-amd64)
>    on debian-squeeze?
> 
>    BTW if I start the server with the same kernel (without xen) the problem
>    does not exist.
> 
> 
> 
>    Since Giam gave us a lot of relevant informations on this bug, I
don''t
>    know which details
> 
>    I could post here to help further on this.
> 
What do you do to reproduce this bug? 

Can you post the full stacktrace/BUG log ?

-- Pasi



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Pasi,
thanks for the quick response.

As mentioned in previous posts, creating and afterwards deleting snapshots
crashes the system.
I can do the following commands only twice, max. three times, before the server
crashes.

lvcreate -s -L 10G some_vg/some_lv -n snap_some_vg_some_lv
lvremove /dev/some_vg/some_lv -f

Also tried this:
sync; sleep 10; lvcreate -s -L 10G some_vg/some_lv -n snap_some_vg_some_lv;
sync; sleep 10; lvremove /dev/some_vg/some_lv -f

Result:
---------------------------------------------------------------------------------------------------
Message from syslogd@svd-xen at Feb 21 14:32:33 ...
 kernel:[22916.840309] ------------[ cut here ]------------

Message from syslogd@svd-xen at Feb 21 14:32:33 ...
 kernel:[22916.840323] invalid opcode: 0000 [#2] SMP

Message from syslogd@svd-xen at Feb 21 14:32:33 ...
 kernel:[22916.840329] last sysfs file:
/sys/devices/pci0000:00/0000:00:1f.2/host1/target1:0:0/1:0:0:0/block/sdb/sdb1/alignment_offset

Message from syslogd@svd-xen at Feb 21 14:32:33 ...
 kernel:[22916.840488] Stack:

Message from syslogd@svd-xen at Feb 21 14:32:33 ...
 kernel:[22916.840512] Call Trace:

Message from syslogd@svd-xen at Feb 21 14:32:33 ...
 kernel:[22916.840549] Code: ec 28 89 3c 24 48 89 f7 e8 a2 fd ff ff 48 89 e7 48
89 44 24 08 be 01 00 00 00 31 d2 41 ba f0 7f 00 00 e8 b0 cc ff     ff 85 c0 74
04 <0f> 0b eb fe 48 83 c4 28 c3 55 49 89 ca 48 89 d5 40 88 f1 48 89
---------------------------------------------------------------------------------------------------

/sys/devices/pci0000:00/0000:00:1f.2/host1/target1:0:0/1:0:0:0/block/sdb/sdb1/alignment_offset

The above line is amazing, since the disk sdb isn''t involved at all.
The LVM stuff resides on sda.

Here the whole thing from syslog:
--------------------------------------------------------------------------------------------------
Feb 21 14:44:33 svd-xen kernel: [23636.279456] CPU 4:
Feb 21 14:44:33 svd-xen kernel: [23636.279458] Modules linked in: xt_tcpudp
nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack xt_physdev
ip6table_filter ip6_tables iptable_filter ip_tables ebtable_nat ebtables
x_tables bridge stp parport_pc ppdev lp parport cpufreq_powersave cpufreq_stats
cpufreq_userspace cpufreq_conservative xen_evtchn xenfs fuse bonding ext2 loop
evdev i2c_i801 psmouse i2c_core pcspkr serio_raw ioatdma button processor
acpi_processor ext3 jbd mbcache dm_snapshot dm_mod sg sr_mod cdrom sd_mod
crc_t10dif usbhid hid uhci_hcd ahci ehci_hcd libata usbcore nls_base aacraid
scsi_mod thermal igb dca thermal_sys [last unloaded: scsi_wait_scan]
Feb 21 14:44:33 svd-xen kernel: [23636.279508] Pid: 6881, comm: udisks-dm-expor
Tainted: G      D    2.6.32-5-xen-amd64 #1 X8DT3
Feb 21 14:44:33 svd-xen kernel: [23636.279510] RIP:
e030:[<ffffffff8130cb1b>]  [<ffffffff8130cb1b>] _spin_lock+0x18/0x1b
Feb 21 14:44:33 svd-xen kernel: [23636.279516] RSP: e02b:ffff88059ad21b10 
EFLAGS: 00000297
Feb 21 14:44:33 svd-xen kernel: [23636.279518] RAX: 0000000000000027 RBX:
ffff88059ad21b28 RCX: ffff88059ad21b68
Feb 21 14:44:33 svd-xen kernel: [23636.279520] RDX: 0000000000000026 RSI:
0000000000000200 RDI: ffff8805dc0db180
Feb 21 14:44:33 svd-xen kernel: [23636.279522] RBP: 0000000000000200 R08:
0000000000000008 R09: ffffffff814eb870
Feb 21 14:44:33 svd-xen kernel: [23636.279524] R10: 000000000000000b R11:
00000000000186a0 R12: ffff8805dc0db100
Feb 21 14:44:33 svd-xen kernel: [23636.279526] R13: 000000000000c580 R14:
ffff88059ad21b28 R15: ffffffff814eb830
Feb 21 14:44:33 svd-xen kernel: [23636.279531] FS:  00007feca22657a0(0000)
GS:ffff88001c6c2000(0000) knlGS:0000000000000000
Feb 21 14:44:33 svd-xen kernel: [23636.279533] CS:  e033 DS: 0000 ES: 0000 CR0:
000000008005003b
Feb 21 14:44:33 svd-xen kernel: [23636.279535] CR2: 00007feca1c043c0 CR3:
0000000001001000 CR4: 0000000000002660
Feb 21 14:44:33 svd-xen kernel: [23636.279537] DR0: 0000000000000000 DR1:
0000000000000000 DR2: 0000000000000000
Feb 21 14:44:33 svd-xen kernel: [23636.279539] DR3: 0000000000000000 DR6:
00000000ffff0ff0 DR7: 0000000000000400
Feb 21 14:44:33 svd-xen kernel: [23636.279541] Call Trace:
Feb 21 14:44:33 svd-xen kernel: [23636.279545]  [<ffffffff8100dd87>] ?
xen_exit_mmap+0xf8/0x136
Feb 21 14:44:33 svd-xen kernel: [23636.279548]  [<ffffffff810d1208>] ?
exit_mmap+0x5a/0x148
Feb 21 14:44:33 svd-xen kernel: [23636.279551]  [<ffffffff8104cb09>] ?
mmput+0x3c/0xdf
Feb 21 14:44:33 svd-xen kernel: [23636.279554]  [<ffffffff81050702>] ?
exit_mm+0x102/0x10d
Feb 21 14:44:33 svd-xen kernel: [23636.279557]  [<ffffffff8130ca72>] ?
_spin_lock_irq+0x7/0x22
Feb 21 14:44:33 svd-xen kernel: [23636.279560]  [<ffffffff81052127>] ?
do_exit+0x1f8/0x6c6
Feb 21 14:44:33 svd-xen kernel: [23636.279564]  [<ffffffff8100ecf2>] ?
check_events+0x12/0x20
Feb 21 14:44:33 svd-xen kernel: [23636.279567]  [<ffffffff8130d9dd>] ?
oops_end+0xaf/0xb4
Feb 21 14:44:33 svd-xen kernel: [23636.279570]  [<ffffffff810135f0>] ?
do_invalid_op+0x8b/0x95
Feb 21 14:44:33 svd-xen kernel: [23636.279573]  [<ffffffff8100c694>] ?
pin_pagetable_pfn+0x2d/0x36
Feb 21 14:44:33 svd-xen kernel: [23636.279576]  [<ffffffff810baf07>] ?
__alloc_pages_nodemask+0x11c/0x5f5
Feb 21 14:44:33 svd-xen kernel: [23636.279579]  [<ffffffff8101293b>] ?
invalid_op+0x1b/0x20
Feb 21 14:44:33 svd-xen kernel: [23636.279582]  [<ffffffff8100c694>] ?
pin_pagetable_pfn+0x2d/0x36
Feb 21 14:44:33 svd-xen kernel: [23636.279585]  [<ffffffff8100c690>] ?
pin_pagetable_pfn+0x29/0x36
Feb 21 14:44:33 svd-xen kernel: [23636.279588]  [<ffffffff810cd4e2>] ?
__pte_alloc+0x6b/0xc6
Feb 21 14:44:33 svd-xen kernel: [23636.279591]  [<ffffffff810cb394>] ?
pmd_alloc+0x28/0x5b
Feb 21 14:44:33 svd-xen kernel: [23636.279594]  [<ffffffff810cd60b>] ?
handle_mm_fault+0xce/0x80f
Feb 21 14:44:33 svd-xen kernel: [23636.279597]  [<ffffffff8102ddc0>] ?
pvclock_clocksource_read+0x3a/0x8b
Feb 21 14:44:33 svd-xen kernel: [23636.279600]  [<ffffffff8130f016>] ?
do_page_fault+0x2e0/0x2fc
Feb 21 14:44:33 svd-xen kernel: [23636.279603]  [<ffffffff8130ceb5>] ?
page_fault+0x25/0x30
--------------------------------------------------------------------------------------------------

Please tell me, if you need more informations.
Sorry for posting with wrong subject line "(Xen-dev]" instead of
[Xen-devel].
Should I post it again to [Xen-devel]?

regards,
Guido



-----Ursprüngliche Nachricht-----
Von: Pasi Kärkkäinen [mailto:pasik@iki.fi] 
Gesendet: Montag, 21. Februar 2011 13:58
An: Guido Hecken
Cc: xen-devel@lists.xensource.com
Betreff: Re: [Xen-devel] [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!

On Mon, Feb 21, 2011 at 12:02:55PM +0100, Guido Hecken
wrote:
>    Same problem here,
> 
>    what is the actual state of this bug?
> 
>    Is there any workaround for "older" kernels like mine
(2.6.32-5-xen-amd64)
>    on debian-squeeze?
> 
>    BTW if I start the server with the same kernel (without xen) the problem
>    does not exist.
> 
> 
> 
>    Since Giam gave us a lot of relevant informations on this bug, I
don''t
>    know which details
> 
>    I could post here to help further on this.
> 
What do you do to reproduce this bug? 

Can you post the full stacktrace/BUG log ?

-- Pasi




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken
wrote:
> Pasi,
> thanks for the quick response.
> 
> As mentioned in previous posts, creating and afterwards deleting snapshots
crashes the system.
> I can do the following commands only twice, max. three times, before the
server crashes.
You need these two patches:

b2464c422fb44275deeb5770b668351860f68e0e
"x86: hold mm->page_table_lock while doing vmalloc_sync"
a8bf92faf72dba6e0dd2130256cb81c3e68f672b
"vmalloc: eagerly clear ptes on vunmap".

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Mon, Feb 21, 2011 at 11:15:39AM -0500, Konrad Rzeszutek Wilk
wrote:
> On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken wrote:
> > Pasi,
> > thanks for the quick response.
> > 
> > As mentioned in previous posts, creating and afterwards deleting
snapshots crashes the system.
> > I can do the following commands only twice, max. three times, before
the server crashes.
> 
> You need these two patches:
> 
> b2464c422fb44275deeb5770b668351860f68e0e
> "x86: hold mm->page_table_lock while doing vmalloc_sync"
> a8bf92faf72dba6e0dd2130256cb81c3e68f672b
> "vmalloc: eagerly clear ptes on vunmap".
>
Sounds like someone should open a bugreport to debian 
to add these patches to their kernel.

Those patches have been in xen/stable-2.6.32.x branch for around 5 months..

-- Pasi


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Konrad,

thanks for the fast response and your hint.
I suppose, I ''ll have to recompile the debian standard kernel to
integrate these patches or did I miss something?
If so, this drives me back to a non-standard system which I can''t
update by normal system-tools without taking special care.
Is there any other way to integrate these patches?

regards,
Guido

-----Ursprüngliche Nachricht-----
Von: Konrad Rzeszutek Wilk [mailto:konrad.wilk@oracle.com] 
Gesendet: Montag, 21. Februar 2011 17:16
An: Guido Hecken
Cc: Pasi Kärkkäinen; xen-devel@lists.xensource.com
Betreff: Re: [Xen-devel] [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!

On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken
wrote:
> Pasi,
> thanks for the quick response.
> 
> As mentioned in previous posts, creating and afterwards deleting snapshots
crashes the system.
> I can do the following commands only twice, max. three times, before the
server crashes.
You need these two patches:

b2464c422fb44275deeb5770b668351860f68e0e
"x86: hold mm->page_table_lock while doing vmalloc_sync"
a8bf92faf72dba6e0dd2130256cb81c3e68f672b
"vmalloc: eagerly clear ptes on vunmap".


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Mon, 2011-02-21 at 16:28 +0000, Pasi Kärkkäinen
wrote:
> On Mon, Feb 21, 2011 at 11:15:39AM -0500, Konrad Rzeszutek Wilk wrote:
> > On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken wrote:
> > > Pasi,
> > > thanks for the quick response.
> > > 
> > > As mentioned in previous posts, creating and afterwards deleting
snapshots crashes the system.
> > > I can do the following commands only twice, max. three times,
before the server crashes.
> > 
> > You need these two patches:
> > 
> > b2464c422fb44275deeb5770b668351860f68e0e
> > "x86: hold mm->page_table_lock while doing vmalloc_sync"
> > a8bf92faf72dba6e0dd2130256cb81c3e68f672b
> > "vmalloc: eagerly clear ptes on vunmap".
> >
> 
> Sounds like someone should open a bugreport to debian 
> to add these patches to their kernel.
Yes please. Even if I remember I''ll need a bug report to justify the
change for a stable update.
> Those patches have been in xen/stable-2.6.32.x branch for around 5 months..
5 months ago is post the xen.git snapshot used in Squeeze unfortunately
so it missed out.

Ian.
> -- Pasi
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
-- 
Ian Campbell
Current Noise: Ludicra - Stagnant Pond

I never said all Democrats were saloonkeepers; what I said was all
saloonkeepers were Democrats.


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Mon, 2011-02-21 at 16:55 +0000, Guido Hecken wrote:
> Konrad,
> 
> thanks for the fast response and your hint.
> I suppose, I ''ll have to recompile the debian standard kernel to
integrate these patches or did I miss something?
> If so, this drives me back to a non-standard system which I can''t
update by normal system-tools without taking special care.
> Is there any other way to integrate these patches?
Only by reporting the issue to Debian and waiting for them to release an
update. There is nothing this list can do to update an existing kernel
binary for you.

Ian.
> 
> regards,
> Guido
> 
> -----Ursprüngliche Nachricht-----
> Von: Konrad Rzeszutek Wilk [mailto:konrad.wilk@oracle.com] 
> Gesendet: Montag, 21. Februar 2011 17:16
> An: Guido Hecken
> Cc: Pasi Kärkkäinen; xen-devel@lists.xensource.com
> Betreff: Re: [Xen-devel] [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!
> 
> On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken wrote:
> > Pasi,
> > thanks for the quick response.
> > 
> > As mentioned in previous posts, creating and afterwards deleting
snapshots crashes the system.
> > I can do the following commands only twice, max. three times, before
the server crashes.
> 
> You need these two patches:
> 
> b2464c422fb44275deeb5770b668351860f68e0e
> "x86: hold mm->page_table_lock while doing vmalloc_sync"
> a8bf92faf72dba6e0dd2130256cb81c3e68f672b
> "vmalloc: eagerly clear ptes on vunmap".
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
-- 
Ian Campbell
Current Noise: Ludicra - Stagnant Pond

What''s page one, a preemptive strike?
		-- Professor Freund, Communication, Ramapo State College


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Mon, Feb 21, 2011 at 05:55:41PM +0100, Guido Hecken
wrote:
> Konrad,
> 
> thanks for the fast response and your hint.
> I suppose, I ''ll have to recompile the debian standard kernel to
integrate these patches or did I miss something?
> If so, this drives me back to a non-standard system which I can''t
update by normal system-tools without taking special care.
> Is there any other way to integrate these patches?
> 
That''s why Debian needs to integrate those patches to their kernel.
They''ve been in xen/stable-2.6.32.x branch for around 5 months already.

-- Pasi
> regards,
> Guido
> 
> -----Ursprüngliche Nachricht-----
> Von: Konrad Rzeszutek Wilk [mailto:konrad.wilk@oracle.com] 
> Gesendet: Montag, 21. Februar 2011 17:16
> An: Guido Hecken
> Cc: Pasi Kärkkäinen; xen-devel@lists.xensource.com
> Betreff: Re: [Xen-devel] [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!
> 
> On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken wrote:
> > Pasi,
> > thanks for the quick response.
> > 
> > As mentioned in previous posts, creating and afterwards deleting
snapshots crashes the system.
> > I can do the following commands only twice, max. three times, before
the server crashes.
> 
> You need these two patches:
> 
> b2464c422fb44275deeb5770b668351860f68e0e
> "x86: hold mm->page_table_lock while doing vmalloc_sync"
> a8bf92faf72dba6e0dd2130256cb81c3e68f672b
> "vmalloc: eagerly clear ptes on vunmap".
> 
> 
> _______________________________________________
> 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

I have opened a bug report at debian.org on this problem.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=614400

Thanks for your time

regards,
Guido

-----Ursprüngliche Nachricht-----
Von: Pasi Kärkkäinen [mailto:pasik@iki.fi] 
Gesendet: Montag, 21. Februar 2011 18:01
An: Guido Hecken
Cc: Konrad Rzeszutek Wilk; xen-devel@lists.xensource.com
Betreff: Re: [Xen-devel] [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!

On Mon, Feb 21, 2011 at 05:55:41PM +0100, Guido Hecken
wrote:
> Konrad,
> 
> thanks for the fast response and your hint.
> I suppose, I ''ll have to recompile the debian standard kernel to
integrate these patches or did I miss something?
> If so, this drives me back to a non-standard system which I can''t
update by normal system-tools without taking special care.
> Is there any other way to integrate these patches?
> 
That''s why Debian needs to integrate those patches to their kernel.
They''ve been in xen/stable-2.6.32.x branch for around 5 months already.

-- Pasi
> regards,
> Guido
> 
> -----Ursprüngliche Nachricht-----
> Von: Konrad Rzeszutek Wilk [mailto:konrad.wilk@oracle.com] 
> Gesendet: Montag, 21. Februar 2011 17:16
> An: Guido Hecken
> Cc: Pasi Kärkkäinen; xen-devel@lists.xensource.com
> Betreff: Re: [Xen-devel] [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!
> 
> On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken wrote:
> > Pasi,
> > thanks for the quick response.
> > 
> > As mentioned in previous posts, creating and afterwards deleting
snapshots crashes the system.
> > I can do the following commands only twice, max. three times, before
the server crashes.
> 
> You need these two patches:
> 
> b2464c422fb44275deeb5770b668351860f68e0e
> "x86: hold mm->page_table_lock while doing vmalloc_sync"
> a8bf92faf72dba6e0dd2130256cb81c3e68f672b
> "vmalloc: eagerly clear ptes on vunmap".
> 
> 
> _______________________________________________
> 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

On Mon, Feb 21, 2011 at 11:28 PM, Pasi Kärkkäinen <pasik@iki.fi>
wrote:
> On Mon, Feb 21, 2011 at 11:15:39AM -0500, Konrad Rzeszutek Wilk wrote:
>> On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken wrote:
>> > Pasi,
>> > thanks for the quick response.
>> >
>> > As mentioned in previous posts, creating and afterwards deleting
snapshots crashes the system.
>> > I can do the following commands only twice, max. three times,
before the server crashes.
>>
>> You need these two patches:
>>
>> b2464c422fb44275deeb5770b668351860f68e0e
>> "x86: hold mm->page_table_lock while doing vmalloc_sync"
>> a8bf92faf72dba6e0dd2130256cb81c3e68f672b
>> "vmalloc: eagerly clear ptes on vunmap".
>>
>
> Sounds like someone should open a bugreport to debian
> to add these patches to their kernel.
>
> Those patches have been in xen/stable-2.6.32.x branch for around 5 months..
It might not be those patches that fixed the bug.

I was testing latest xen/stable-2.6.32.x (currently at 2.6.32.27,
definitely has those patches) with zvol when I got the bug, which
manifests itself when starting 4 domUs at boot using xendomains
service. So I tried xen/next-2.6.32 (which was just updated to
2.6.32.28), works fine so far.

So to Guido, if you have time, please try using both
xen/stable-2.6.32.x and xen/next-2.6.32 to verify which version can
solve your problem. Regarding your concern about non-standard systems,
"make-kpkg" might help
(http://newbiedoc.sourceforge.net/system/kernel-pkg.html).

-- 
Fajar

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Fajar,

thanks for your comments on this.

Here some explanatory notes concerning the case:
I thought it would be a good idea to have the base for production xen-servers on
a stable distribution with presumable stable packages installed, instead of
compiling my own.
In the past I used to compile the needed xen dom0 kernel on top of Centos 5.x or
Fedora but stumbled on not to be compliant to distribution specific update
methods anymore.
Especially compiling new xen kernels on top of the old 2.6.18 kernel in Centos
5.x was a pain. So I switched to Debian Squeeze which has the reputation to be
the distribution with the most actual xen packages integrated.
Another reason for choosing a presumable stable distribution was the long time
support (LTS) one would get...

I will have a closer look on your provided link and hopefully, it will lead to a
stable system with newest and stable xen-kernel, which leaves other installed
packages still uptadeable through distribution specific methods.

regards,
Guido


-----Ursprüngliche Nachricht-----
Von: Fajar A. Nugraha [mailto:list@fajar.net] 
Gesendet: Mittwoch, 23. Februar 2011 09:31
An: Guido Hecken
Cc: Konrad Rzeszutek Wilk; xen-devel@lists.xensource.com; Pasi Kärkkäinen
Betreff: Re: [Xen-devel] [Xen-dev] kernel BUG at arch/x86/xen/mmu.c:1860!

On Mon, Feb 21, 2011 at 11:28 PM, Pasi Kärkkäinen <pasik@iki.fi>
wrote:
> On Mon, Feb 21, 2011 at 11:15:39AM -0500, Konrad Rzeszutek Wilk wrote:
>> On Mon, Feb 21, 2011 at 02:58:53PM +0100, Guido Hecken wrote:
>> > Pasi,
>> > thanks for the quick response.
>> >
>> > As mentioned in previous posts, creating and afterwards deleting
snapshots crashes the system.
>> > I can do the following commands only twice, max. three times,
before the server crashes.
>>
>> You need these two patches:
>>
>> b2464c422fb44275deeb5770b668351860f68e0e
>> "x86: hold mm->page_table_lock while doing vmalloc_sync"
>> a8bf92faf72dba6e0dd2130256cb81c3e68f672b
>> "vmalloc: eagerly clear ptes on vunmap".
>>
>
> Sounds like someone should open a bugreport to debian
> to add these patches to their kernel.
>
> Those patches have been in xen/stable-2.6.32.x branch for around 5 months..
It might not be those patches that fixed the bug.

I was testing latest xen/stable-2.6.32.x (currently at 2.6.32.27,
definitely has those patches) with zvol when I got the bug, which
manifests itself when starting 4 domUs at boot using xendomains
service. So I tried xen/next-2.6.32 (which was just updated to
2.6.32.28), works fine so far.

So to Guido, if you have time, please try using both
xen/stable-2.6.32.x and xen/next-2.6.32 to verify which version can
solve your problem. Regarding your concern about non-standard systems,
"make-kpkg" might help
(http://newbiedoc.sourceforge.net/system/kernel-pkg.html).

-- 
Fajar


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel