* dom0 pvops crash when used as domU kernel
@ 2010-01-26 13:50 Ian Jackson
2010-01-26 20:14 ` Jeremy Fitzhardinge
0 siblings, 1 reply; 3+ messages in thread
From: Ian Jackson @ 2010-01-26 13:50 UTC (permalink / raw)
To: xen-devel, Jeremy Fitzhardinge
Is the dom0 pvops kernel supposed to be useable for a PV guest ?
If so, it doesn't work on amd64. See attached console output from the
guest.
The host and guest are both xen-unstable x86_64 with the dom0 pvops
kernel. The host and guest operating system are Debian lenny amd64.
The guest is installed with Debian's xen-create-image, which uses
debootstrap on the host to construct the guest's filesystem in an LV
and set it up ready to be booted PV.
The problem occurs during or shortly after save/restore. My test
system did this:
LOG executing ssh ... root@10.80.249.102 xm list
LOG executing ssh ... root@10.80.249.102 xm save one.guest.osstest image
LOG ping 10.80.249.241 down
LOG executing ssh ... root@10.80.249.102 xm restore image
LOG ping 10.80.249.241 up
LOG await tcp one.guest.osstest 22: waiting 5s...
LOG await tcp one.guest.osstest 22: ok. (0s)
LOG executing ssh ... root@10.80.249.241 echo guest one.guest.osstest: restored
ssh_exchange_identification: Connection closed by remote host
status 65280 at Osstest.pm line 190.
Ian.
[ 19.662648] Grant tables using version 2 layout.
[ 23.870414] ------------[ cut here ]------------
[ 23.870439] kernel BUG at /home/osstest/build.1142.build-amd64/xen-unstable/linux-2.6-pvops.git/include/linux/dcache.h:336!
[ 23.870458] invalid opcode: 0000 [#1] SMP
[ 23.870474] last sysfs file: /sys/kernel/uevent_seqnum
[ 23.870485] CPU 0
[ 23.870497] Modules linked in: [last unloaded: scsi_wait_scan]
[ 23.870518] Pid: 1438, comm: sshd Not tainted 2.6.31.6 #1
[ 23.870529] RIP: e030:[<ffffffff81106c70>] [<ffffffff81106c70>] __follow_mount+0x4f/0x76
[ 23.870555] RSP: e02b:ffff88001ed53c18 EFLAGS: 00010246
[ 23.870566] RAX: 0000000000000000 RBX: ffff88001e84ca00 RCX: 0000000000000002
[ 23.870580] RDX: ffff88001f4826c0 RSI: 0000000000000001 RDI: ffff88001f5ad840
[ 23.870594] RBP: ffff88001ed53c38 R08: 0000000000000000 R09: 0000000000000000
[ 23.870607] R10: 0000000000000007 R11: 0000000000100000 R12: 0000000000000000
[ 23.870620] R13: ffff88001ed53ca8 R14: ffff88001ed53d68 R15: 0000000000000001
[ 23.870638] FS: 00007f0ee1912790(0000) GS:ffffc90000000000(0000) knlGS:0000000000000000
[ 23.870654] CS: e033 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 23.870666] CR2: 00007f0edf52fca0 CR3: 000000001f146000 CR4: 0000000000002660
[ 23.870680] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 23.870695] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 23.870709] Process sshd (pid: 1438, threadinfo ffff88001ed52000, task ffff88001edc8f20)
[ 23.870723] Stack:
[ 23.870731] ffff88001ed53d68 ffff88001f5ad840 ffff88001ed53d68 ffff88001ed53ca8
[ 23.870752] <0> ffff88001ed53c88 ffffffff81106de4 ffff88001ed53ca8 ffff88001ed53c98
[ 23.870778] <0> ffff88001e955200 ffff88001ed541ed ffff88001ed53d68 ffff88001ed53ca8
[ 23.870807] Call Trace:
[ 23.870821] [<ffffffff81106de4>] do_lookup+0x61/0x15d
[ 23.870849] [<ffffffff811079fd>] __link_path_walk+0x61a/0x770
[ 23.870866] [<ffffffff81107d9a>] path_walk+0x69/0xd4
[ 23.870881] [<ffffffff811081e5>] do_path_lookup+0x2a/0x86
[ 23.870895] [<ffffffff8110a72f>] user_path_at+0x52/0x8c
[ 23.870910] [<ffffffff810cf02e>] ? unlock_page+0x22/0x26
[ 23.870925] [<ffffffff811001e7>] ? file_free+0x31/0x35
[ 23.870939] [<ffffffff810fe9e4>] sys_statfs+0x29/0x95
[ 23.870954] [<ffffffff8102f0f1>] ? xen_force_evtchn_callback+0xd/0xf
[ 23.870974] [<ffffffff8102f822>] ? check_events+0x12/0x20
[ 23.870989] [<ffffffff8102f80f>] ? xen_restore_fl_direct_end+0x0/0x1
[ 23.871007] [<ffffffff81576dd9>] ? _spin_unlock_irqrestore+0x34/0x36
[ 23.871013] [<ffffffff811ffc5a>] ? __up_read+0x92/0x9c
[ 23.871013] [<ffffffff81084c5b>] ? up_read+0x9/0xb
[ 23.871013] [<ffffffff8157959b>] ? do_page_fault+0x2d5/0x307
[ 23.871013] [<ffffffff81577245>] ? page_fault+0x25/0x30
[ 23.871013] [<ffffffff81033cc2>] system_call_fastpath+0x16/0x1b
[ 23.871013] Code: 45 49 8b 7d 08 e8 78 8d 00 00 45 85 e4 74 09 49 8b 7d 00 e8 97 fe ff ff 49 89 5d 00 48 8b 53 20 48 85 d2 74 0d 8b 02 85 c0 75 04 <0f> 0b eb fe 3e ff 02 49 89 55 08 41 bc 01 00 00 00 49 8b 45 08
[ 23.871013] RIP [<ffffffff81106c70>] __follow_mount+0x4f/0x76
[ 23.871013] RSP <ffff88001ed53c18>
[ 23.871013] ---[ end trace d3d7603d6373c307 ]---
[ 23.893521] ------------[ cut here ]------------
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: dom0 pvops crash when used as domU kernel
2010-01-26 13:50 dom0 pvops crash when used as domU kernel Ian Jackson
@ 2010-01-26 20:14 ` Jeremy Fitzhardinge
2010-01-29 14:10 ` Ian Jackson
0 siblings, 1 reply; 3+ messages in thread
From: Jeremy Fitzhardinge @ 2010-01-26 20:14 UTC (permalink / raw)
To: Ian Jackson; +Cc: Ian Campbell, xen-devel
On 01/26/2010 05:50 AM, Ian Jackson wrote:
> Is the dom0 pvops kernel supposed to be useable for a PV guest ?
>
Yep.
> If so, it doesn't work on amd64. See attached console output from the
> guest.
>
> The host and guest are both xen-unstable x86_64 with the dom0 pvops
> kernel. The host and guest operating system are Debian lenny amd64.
> The guest is installed with Debian's xen-create-image, which uses
> debootstrap on the host to construct the guest's filesystem in an LV
> and set it up ready to be booted PV.
>
The backtrace is strange; sshd doing statfs shouldn't be hitting any
Xen-specific paths. That suggests something else has gone wrong earlier.
> The problem occurs during or shortly after save/restore. My test
> system did this:
>
Hm, save/restore got pretty bitrotted for a while. IanC has done a mass
of work on it lately, but its possible there are still some issues. Are
you actually using an AMD processor? They tend to be undertested.
J
> LOG executing ssh ... root@10.80.249.102 xm list
> LOG executing ssh ... root@10.80.249.102 xm save one.guest.osstest image
> LOG ping 10.80.249.241 down
> LOG executing ssh ... root@10.80.249.102 xm restore image
> LOG ping 10.80.249.241 up
> LOG await tcp one.guest.osstest 22: waiting 5s...
> LOG await tcp one.guest.osstest 22: ok. (0s)
> LOG executing ssh ... root@10.80.249.241 echo guest one.guest.osstest: restored
> ssh_exchange_identification: Connection closed by remote host
> status 65280 at Osstest.pm line 190.
>
> Ian.
>
> [ 19.662648] Grant tables using version 2 layout.
> [ 23.870414] ------------[ cut here ]------------
> [ 23.870439] kernel BUG at /home/osstest/build.1142.build-amd64/xen-unstable/linux-2.6-pvops.git/include/linux/dcache.h:336!
> [ 23.870458] invalid opcode: 0000 [#1] SMP
> [ 23.870474] last sysfs file: /sys/kernel/uevent_seqnum
> [ 23.870485] CPU 0
> [ 23.870497] Modules linked in: [last unloaded: scsi_wait_scan]
> [ 23.870518] Pid: 1438, comm: sshd Not tainted 2.6.31.6 #1
> [ 23.870529] RIP: e030:[<ffffffff81106c70>] [<ffffffff81106c70>] __follow_mount+0x4f/0x76
> [ 23.870555] RSP: e02b:ffff88001ed53c18 EFLAGS: 00010246
> [ 23.870566] RAX: 0000000000000000 RBX: ffff88001e84ca00 RCX: 0000000000000002
> [ 23.870580] RDX: ffff88001f4826c0 RSI: 0000000000000001 RDI: ffff88001f5ad840
> [ 23.870594] RBP: ffff88001ed53c38 R08: 0000000000000000 R09: 0000000000000000
> [ 23.870607] R10: 0000000000000007 R11: 0000000000100000 R12: 0000000000000000
> [ 23.870620] R13: ffff88001ed53ca8 R14: ffff88001ed53d68 R15: 0000000000000001
> [ 23.870638] FS: 00007f0ee1912790(0000) GS:ffffc90000000000(0000) knlGS:0000000000000000
> [ 23.870654] CS: e033 DS: 0000 ES: 0000 CR0: 000000008005003b
> [ 23.870666] CR2: 00007f0edf52fca0 CR3: 000000001f146000 CR4: 0000000000002660
> [ 23.870680] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> [ 23.870695] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> [ 23.870709] Process sshd (pid: 1438, threadinfo ffff88001ed52000, task ffff88001edc8f20)
> [ 23.870723] Stack:
> [ 23.870731] ffff88001ed53d68 ffff88001f5ad840 ffff88001ed53d68 ffff88001ed53ca8
> [ 23.870752]<0> ffff88001ed53c88 ffffffff81106de4 ffff88001ed53ca8 ffff88001ed53c98
> [ 23.870778]<0> ffff88001e955200 ffff88001ed541ed ffff88001ed53d68 ffff88001ed53ca8
> [ 23.870807] Call Trace:
> [ 23.870821] [<ffffffff81106de4>] do_lookup+0x61/0x15d
> [ 23.870849] [<ffffffff811079fd>] __link_path_walk+0x61a/0x770
> [ 23.870866] [<ffffffff81107d9a>] path_walk+0x69/0xd4
> [ 23.870881] [<ffffffff811081e5>] do_path_lookup+0x2a/0x86
> [ 23.870895] [<ffffffff8110a72f>] user_path_at+0x52/0x8c
> [ 23.870910] [<ffffffff810cf02e>] ? unlock_page+0x22/0x26
> [ 23.870925] [<ffffffff811001e7>] ? file_free+0x31/0x35
> [ 23.870939] [<ffffffff810fe9e4>] sys_statfs+0x29/0x95
> [ 23.870954] [<ffffffff8102f0f1>] ? xen_force_evtchn_callback+0xd/0xf
> [ 23.870974] [<ffffffff8102f822>] ? check_events+0x12/0x20
> [ 23.870989] [<ffffffff8102f80f>] ? xen_restore_fl_direct_end+0x0/0x1
> [ 23.871007] [<ffffffff81576dd9>] ? _spin_unlock_irqrestore+0x34/0x36
> [ 23.871013] [<ffffffff811ffc5a>] ? __up_read+0x92/0x9c
> [ 23.871013] [<ffffffff81084c5b>] ? up_read+0x9/0xb
> [ 23.871013] [<ffffffff8157959b>] ? do_page_fault+0x2d5/0x307
> [ 23.871013] [<ffffffff81577245>] ? page_fault+0x25/0x30
> [ 23.871013] [<ffffffff81033cc2>] system_call_fastpath+0x16/0x1b
> [ 23.871013] Code: 45 49 8b 7d 08 e8 78 8d 00 00 45 85 e4 74 09 49 8b 7d 00 e8 97 fe ff ff 49 89 5d 00 48 8b 53 20 48 85 d2 74 0d 8b 02 85 c0 75 04<0f> 0b eb fe 3e ff 02 49 89 55 08 41 bc 01 00 00 00 49 8b 45 08
> [ 23.871013] RIP [<ffffffff81106c70>] __follow_mount+0x4f/0x76
> [ 23.871013] RSP<ffff88001ed53c18>
> [ 23.871013] ---[ end trace d3d7603d6373c307 ]---
> [ 23.893521] ------------[ cut here ]------------
>
>
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: dom0 pvops crash when used as domU kernel
2010-01-26 20:14 ` Jeremy Fitzhardinge
@ 2010-01-29 14:10 ` Ian Jackson
0 siblings, 0 replies; 3+ messages in thread
From: Ian Jackson @ 2010-01-29 14:10 UTC (permalink / raw)
To: Jeremy Fitzhardinge; +Cc: Ian, Campbell, xen-devel@lists.xensource.com
Jeremy Fitzhardinge writes ("Re: dom0 pvops crash when used as domU kernel"):
> On 01/26/2010 05:50 AM, Ian Jackson wrote:
> > If so, it doesn't work on amd64. See attached console output from the
> > guest.
> >
> > The host and guest are both xen-unstable x86_64 with the dom0 pvops
> > kernel. The host and guest operating system are Debian lenny amd64.
> > The guest is installed with Debian's xen-create-image, which uses
> > debootstrap on the host to construct the guest's filesystem in an LV
> > and set it up ready to be booted PV.
>
> The backtrace is strange; sshd doing statfs shouldn't be hitting any
> Xen-specific paths. That suggests something else has gone wrong earlier.
Quite possibly. The problem only occurs with the 64-bit kernel.
32bit domU with 32bit dom0 on 64bit Xen works fine, as does
all-32-bit.
> > The problem occurs during or shortly after save/restore. My test
> > system did this:
>
> Hm, save/restore got pretty bitrotted for a while. IanC has done a mass
> of work on it lately, but its possible there are still some issues. Are
> you actually using an AMD processor? They tend to be undertested.
No,
model name : Intel(R) Pentium(R) D CPU 2.80GHz
Ian.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2010-01-29 14:10 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-01-26 13:50 dom0 pvops crash when used as domU kernel Ian Jackson
2010-01-26 20:14 ` Jeremy Fitzhardinge
2010-01-29 14:10 ` Ian Jackson
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.