xen problem with save/restore

xen problem with save/restore

[Paul Larson]

I think I've seen something similar to this posted recently, but I can't
find the message to reply to.  I build xen-2.0-testing from source
yesterday.  I have debian sarge running on all my domains right now.  I
was trying the save/restore feature out, and got this after restoring
and consoling into a session:

************ REMOTE CONSOLE: CTRL-] TO QUIT ********
Unable to handle kernel NULL pointer dereference at virtual address
00000007
 printing eip:
c030e587
*pde = ma 00000000 pa 55555000
 [<c0109f80>] __do_suspend+0x19f/0x1e0

 [<c012ca65>] worker_thread+0x22b/0x32f

 [<c010a088>] __shutdown_handler+0x0/0x48

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c012c83a>] worker_thread+0x0/0x32f

 [<c0130b89>] kthread+0xa5/0xab

 [<c0130ae4>] kthread+0x0/0xab

 [<c010f411>] kernel_thread_helper+0x5/0xb

Oops: 0002 [#1]
PREEMPT DEBUG_PAGEALLOC
Modules linked in:
CPU:    0
EIP:    0061:[<c030e587>]    Not tainted VLI
EFLAGS: 00010203   (2.6.9-xenU)
EIP is at init_tsc+0x4b/0x9e
eax: c02c6dd4   ebx: 0000c000   ecx: fbffc000   edx: 00000001
esi: 00000010   edi: c0102000   ebp: 00000000   esp: c0039f04
ds: 007b   es: 007b   ss: 0069
Process events/0 (pid: 3, threadinfo=c0038000 task=c003cac0)
Stack: c010eadd 00000000 c0109f80 fbffc000 000001e1 00000063 8ed0dd7b
00000042
       c02c3200 c2dc5000 c02c5d80 00000000 c0038000 00000000 c012ca65
00000000
       c0039f74 00000000 c003bf10 c0038000 c0038000 c0038000 c010a088
c0038000
Call Trace:
 [<c010eadd>] time_resume+0x12/0x51

 [<c0109f80>] __do_suspend+0x19f/0x1e0

 [<c012ca65>] worker_thread+0x22b/0x32f

 [<c010a088>] __shutdown_handler+0x0/0x48

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c012c83a>] worker_thread+0x0/0x32f

 [<c0130b89>] kthread+0xa5/0xab

 [<c0130ae4>] kthread+0x0/0xab

 [<c010f411>] kernel_thread_helper+0x5/0xb

Code: 00 00 00 00 00 00 00 00 70 0d 5a f6 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 21 00 00 00 00 00 <00> 4e
f7 ac 41 c8 a8 af 41 61 21 4c 41 00 00 00 00 00 00 01 00

I'm just starting to play with xen, so I haven't figured out much of the
debugging stuff yet.  If someone can point me in the right direction
though, I'd be happy to try to collect some more data if there's
anything that would be useful.

Thanks,
Paul Larson



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Jérôme Petazzoni]

Paul Larson wrote:

>I think I've seen something similar to this posted recently, but I can't
>find the message to reply to.  I build xen-2.0-testing from source
>yesterday.  I have debian sarge running on all my domains right now.  I
>was trying the save/restore feature out, and got this after restoring
>and consoling into a session:
>
>************ REMOTE CONSOLE: CTRL-] TO QUIT ********
>Unable to handle kernel NULL pointer dereference at virtual address
>00000007
> printing eip:
>c030e587
>*pde = ma 00000000 pa 55555000
> [<c0109f80>] __do_suspend+0x19f/0x1e0
>  
>
[...]

I had exactly this problem with exactly the same setup (xen-2.0-testing 
on sarge).

Ian told that he would be interested with the vmlinux* files (not the 
vmlinuz ; those with the debug symbols, if I understood correctly).

I couldn't yet reproduce this bug (I recompiled the kernels with 
iptables support, and now the domain crashes on restoring without any 
message ; and I will get physical access to the machines this week-end 
only - and then apply the patch to show all kernel messages from all 
domains on the console, to guess what's happening).

I don't think Debian is at the source of the problem :-) ... but the 
precise version of gcc could be. FYI, I was using gcc-3.3.5 (I compiled 
my kernels on my "fast" sid box, not on the xen box). Which gcc are you 
using ?



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Brian Wolfe]

I had attempted to enforce gcc 3.2 on compiling these packages. Not
certain that I suceeded though... Any tips would be usefull and
appreciated. :)

On Fri, 2004-12-03 at 18:48 +0100, Jérôme Petazzoni wrote:
> Paul Larson wrote:
>
> >I think I've seen something similar to this posted recently, but I can't
> >find the message to reply to.  I build xen-2.0-testing from source
> >yesterday.  I have debian sarge running on all my domains right now.  I
> >was trying the save/restore feature out, and got this after restoring
> >and consoling into a session:
> >
> >************ REMOTE CONSOLE: CTRL-] TO QUIT ********
> >Unable to handle kernel NULL pointer dereference at virtual address
> >00000007
> > printing eip:
> >c030e587
> >*pde = ma 00000000 pa 55555000
> > [<c0109f80>] __do_suspend+0x19f/0x1e0
> >
> >
> [...]
>
> I had exactly this problem with exactly the same setup (xen-2.0-testing
> on sarge).
>
> Ian told that he would be interested with the vmlinux* files (not the
> vmlinuz ; those with the debug symbols, if I understood correctly).
>
> I couldn't yet reproduce this bug (I recompiled the kernels with
> iptables support, and now the domain crashes on restoring without any
> message ; and I will get physical access to the machines this week-end
> only - and then apply the patch to show all kernel messages from all
> domains on the console, to guess what's happening).
>
> I don't think Debian is at the source of the problem :-) ... but the
> precise version of gcc could be. FYI, I was using gcc-3.3.5 (I compiled
> my kernels on my "fast" sid box, not on the xen box). Which gcc are you
> using ?
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel
--



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Paul Larson]

On Fri, 2004-12-03 at 11:48, Jérôme Petazzoni wrote:
> I don't think Debian is at the source of the problem :-) ... but the 
> precise version of gcc could be. FYI, I was using gcc-3.3.5 (I compiled 
> my kernels on my "fast" sid box, not on the xen box). Which gcc are you 
> using ?

gcc version 3.2.3 (Debian)



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Paul Larson]

Interesting, I'm using this on a different machine now and getting a
slightly different error when I restore:
Unable to handle kernel NULL pointer dereference at virtual address
00000007
 printing eip:
c0302407
*pde = ma 00000000 pa 55555000
 [<c012bbee>] worker_thread+0x1ea/0x2e0

The new machine is my laptop, an IBM T41p, but otherwise much the same
(debian sarge, same compiler, etc).

Thanks,
Paul Larson
On Fri, 2004-12-03 at 10:58, Paul Larson wrote:
> I think I've seen something similar to this posted recently, but I can't
> find the message to reply to.  I build xen-2.0-testing from source
> yesterday.  I have debian sarge running on all my domains right now.  I
> was trying the save/restore feature out, and got this after restoring
> and consoling into a session:
> 
> ************ REMOTE CONSOLE: CTRL-] TO QUIT ********
> Unable to handle kernel NULL pointer dereference at virtual address
> 00000007
>  printing eip:
> c030e587
> *pde = ma 00000000 pa 55555000
>  [<c0109f80>] __do_suspend+0x19f/0x1e0
> 
>  [<c012ca65>] worker_thread+0x22b/0x32f
> 
>  [<c010a088>] __shutdown_handler+0x0/0x48
> 
>  [<c01196c0>] default_wake_function+0x0/0x12
> 
>  [<c01196c0>] default_wake_function+0x0/0x12
> 
>  [<c012c83a>] worker_thread+0x0/0x32f
> 
>  [<c0130b89>] kthread+0xa5/0xab
> 
>  [<c0130ae4>] kthread+0x0/0xab
> 
>  [<c010f411>] kernel_thread_helper+0x5/0xb
> 
> Oops: 0002 [#1]
> PREEMPT DEBUG_PAGEALLOC
> Modules linked in:
> CPU:    0
> EIP:    0061:[<c030e587>]    Not tainted VLI
> EFLAGS: 00010203   (2.6.9-xenU)
> EIP is at init_tsc+0x4b/0x9e
> eax: c02c6dd4   ebx: 0000c000   ecx: fbffc000   edx: 00000001
> esi: 00000010   edi: c0102000   ebp: 00000000   esp: c0039f04
> ds: 007b   es: 007b   ss: 0069
> Process events/0 (pid: 3, threadinfo=c0038000 task=c003cac0)
> Stack: c010eadd 00000000 c0109f80 fbffc000 000001e1 00000063 8ed0dd7b
> 00000042
>        c02c3200 c2dc5000 c02c5d80 00000000 c0038000 00000000 c012ca65
> 00000000
>        c0039f74 00000000 c003bf10 c0038000 c0038000 c0038000 c010a088
> c0038000
> Call Trace:
>  [<c010eadd>] time_resume+0x12/0x51
> 
>  [<c0109f80>] __do_suspend+0x19f/0x1e0
> 
>  [<c012ca65>] worker_thread+0x22b/0x32f
> 
>  [<c010a088>] __shutdown_handler+0x0/0x48
> 
>  [<c01196c0>] default_wake_function+0x0/0x12
> 
>  [<c01196c0>] default_wake_function+0x0/0x12
> 
>  [<c012c83a>] worker_thread+0x0/0x32f
> 
>  [<c0130b89>] kthread+0xa5/0xab
> 
>  [<c0130ae4>] kthread+0x0/0xab
> 
>  [<c010f411>] kernel_thread_helper+0x5/0xb
> 
> Code: 00 00 00 00 00 00 00 00 70 0d 5a f6 00 00 00 00 00 00 00 00 00 00
> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 21 00 00 00 00 00 <00> 4e
> f7 ac 41 c8 a8 af 41 61 21 4c 41 00 00 00 00 00 00 01 00
> 
> I'm just starting to play with xen, so I haven't figured out much of the
> debugging stuff yet.  If someone can point me in the right direction
> though, I'd be happy to try to collect some more data if there's
> anything that would be useful.
> 
> Thanks,
> Paul Larson
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now. 
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Paul Larson]

I just tried this on the same system/compiler with the xeno-unstable.bk
tree and save/restore worked for me.  Anyone know of a something in
unstable that might make this work?

Thanks,
Paul Larson

On Mon, 2004-12-06 at 22:44, Paul Larson wrote:
> Interesting, I'm using this on a different machine now and getting a
> slightly different error when I restore:
> Unable to handle kernel NULL pointer dereference at virtual address
> 00000007
>  printing eip:
> c0302407
> *pde = ma 00000000 pa 55555000
>  [<c012bbee>] worker_thread+0x1ea/0x2e0
> 
> The new machine is my laptop, an IBM T41p, but otherwise much the same
> (debian sarge, same compiler, etc).
> 
> Thanks,
> Paul Larson
> On Fri, 2004-12-03 at 10:58, Paul Larson wrote:
> > I think I've seen something similar to this posted recently, but I can't
> > find the message to reply to.  I build xen-2.0-testing from source
> > yesterday.  I have debian sarge running on all my domains right now.  I
> > was trying the save/restore feature out, and got this after restoring
> > and consoling into a session:
> > 
> > ************ REMOTE CONSOLE: CTRL-] TO QUIT ********
> > Unable to handle kernel NULL pointer dereference at virtual address
> > 00000007
> >  printing eip:
> > c030e587
> > *pde = ma 00000000 pa 55555000
> >  [<c0109f80>] __do_suspend+0x19f/0x1e0
> > 
> >  [<c012ca65>] worker_thread+0x22b/0x32f
> > 
> >  [<c010a088>] __shutdown_handler+0x0/0x48
> > 
> >  [<c01196c0>] default_wake_function+0x0/0x12
> > 
> >  [<c01196c0>] default_wake_function+0x0/0x12
> > 
> >  [<c012c83a>] worker_thread+0x0/0x32f
> > 
> >  [<c0130b89>] kthread+0xa5/0xab
> > 
> >  [<c0130ae4>] kthread+0x0/0xab
> > 
> >  [<c010f411>] kernel_thread_helper+0x5/0xb
> > 
> > Oops: 0002 [#1]
> > PREEMPT DEBUG_PAGEALLOC
> > Modules linked in:
> > CPU:    0
> > EIP:    0061:[<c030e587>]    Not tainted VLI
> > EFLAGS: 00010203   (2.6.9-xenU)
> > EIP is at init_tsc+0x4b/0x9e
> > eax: c02c6dd4   ebx: 0000c000   ecx: fbffc000   edx: 00000001
> > esi: 00000010   edi: c0102000   ebp: 00000000   esp: c0039f04
> > ds: 007b   es: 007b   ss: 0069
> > Process events/0 (pid: 3, threadinfo=c0038000 task=c003cac0)
> > Stack: c010eadd 00000000 c0109f80 fbffc000 000001e1 00000063 8ed0dd7b
> > 00000042
> >        c02c3200 c2dc5000 c02c5d80 00000000 c0038000 00000000 c012ca65
> > 00000000
> >        c0039f74 00000000 c003bf10 c0038000 c0038000 c0038000 c010a088
> > c0038000
> > Call Trace:
> >  [<c010eadd>] time_resume+0x12/0x51
> > 
> >  [<c0109f80>] __do_suspend+0x19f/0x1e0
> > 
> >  [<c012ca65>] worker_thread+0x22b/0x32f
> > 
> >  [<c010a088>] __shutdown_handler+0x0/0x48
> > 
> >  [<c01196c0>] default_wake_function+0x0/0x12
> > 
> >  [<c01196c0>] default_wake_function+0x0/0x12
> > 
> >  [<c012c83a>] worker_thread+0x0/0x32f
> > 
> >  [<c0130b89>] kthread+0xa5/0xab
> > 
> >  [<c0130ae4>] kthread+0x0/0xab
> > 
> >  [<c010f411>] kernel_thread_helper+0x5/0xb
> > 
> > Code: 00 00 00 00 00 00 00 00 70 0d 5a f6 00 00 00 00 00 00 00 00 00 00
> > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 21 00 00 00 00 00 <00> 4e
> > f7 ac 41 c8 a8 af 41 61 21 4c 41 00 00 00 00 00 00 01 00
> > 
> > I'm just starting to play with xen, so I haven't figured out much of the
> > debugging stuff yet.  If someone can point me in the right direction
> > though, I'd be happy to try to collect some more data if there's
> > anything that would be useful.
> > 
> > Thanks,
> > Paul Larson
> > 
> > 
> > 
> > -------------------------------------------------------
> > SF email is sponsored by - The IT Product Guide
> > Read honest & candid reviews on hundreds of IT Products from real users.
> > Discover which products truly live up to the hype. Start reading now. 
> > http://productguide.itmanagersjournal.com/
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/xen-devel
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now. 
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

RE: xen problem with save/restore

[Ian Pratt]

 > > The latest 2.0-testing and unstable trees only differ in that a new

> > 'blktap' driver was added to the unstable tree, which is not even 
> > built by default. It really looks to be impossible that you 
> could be 
> > seeing different behaviour between the two repositories. :-/
> I see, I was running a diff of the trees a while ago and I 
> see what you mean.  However, it has been very easy to 
> reproduce this under testing, but on unstable, it never 
> failed for me even once.  So, I installed gcc
> 3.2 and tried compiling testing with that, and it works now, 
> no crash! 
> So it looks like it is a compiler issue after all.

Which compiler from which distro were you using before? I'd like to get
to the bottom of this as other people have seen the same problem. We'd
be grateful if you could confirm that building with the other compiler
causes the bug.

Thanks,
Ian


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/

RE: xen problem with save/restore

[Paul Larson]

On Thu, 2004-12-09 at 09:47, Ian Pratt wrote:
> Which compiler from which distro were you using before? I'd like to get
> to the bottom of this as other people have seen the same problem. We'd
> be grateful if you could confirm that building with the other compiler
> causes the bug.
gcc version 3.3.4 (Debian 1:3.3.4-13) -> Bug with 2.0-testing
gcc version 3.2.3 (Debian) -> No bug with 2.0-testing

I can also try 3.4 pretty easily later today if you'd like

Thanks,
Paul Larson



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Nuutti Kotivuori]

Paul Larson wrote:
> On Thu, 2004-12-09 at 09:47, Ian Pratt wrote:
>> Which compiler from which distro were you using before? I'd like to
>> get to the bottom of this as other people have seen the same
>> problem. We'd be grateful if you could confirm that building with
>> the other compiler causes the bug.
>
> gcc version 3.3.4 (Debian 1:3.3.4-13) -> Bug with 2.0-testing
> gcc version 3.2.3 (Debian) -> No bug with 2.0-testing
>
> I can also try 3.4 pretty easily later today if you'd like

I don't know about the Xen guys, but I personally would like to see if
the bug appears with:

gcc version 3.3.5 (Debian 1:3.3.5-3)

This is the current latest Debian version. The 3.3.4-13 version is
from September, where as the 3.3.5-3 is really from gcc-3_3-branch CVS
20041204. In the meantime there have been fixes to bugs that cause
wrong code generation and crashes.

So, it would be really nice to know if there's a problem with the
default Debian gcc versions, or is it some specific version only.

-- Naked




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Paul Larson]

Paul Larson wrote:

>On Thu, 2004-12-09 at 09:47, Ian Pratt wrote:
>  
>
>>Which compiler from which distro were you using before? I'd like to get
>>to the bottom of this as other people have seen the same problem. We'd
>>be grateful if you could confirm that building with the other compiler
>>causes the bug.
>>    
>>
>gcc version 3.3.4 (Debian 1:3.3.4-13) -> Bug with 2.0-testing
>gcc version 3.2.3 (Debian) -> No bug with 2.0-testing
>
>I can also try 3.4 pretty easily later today if you'd like
>  
>
Seems that 2.0-testing and save/restore works with 3.4 as well.

Thanks,
Paul Larson


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

RE: xen problem with save/restore

[Ian Pratt]

What were you using before? I seem to recall it was 2.0-testing.bk?

There's currently very few changes between unstable and 2.0-testing
(though this will likely change from next week).  Does it now work
totally reliably? (it failed every time before, right?). You're using
the exact same machine and configuration?

Ian

> I just tried this on the same system/compiler with the 
> xeno-unstable.bk tree and save/restore worked for me.  Anyone 
> know of a something in unstable that might make this work?
> 
> Thanks,
> Paul Larson
> 
> On Mon, 2004-12-06 at 22:44, Paul Larson wrote:
> > Interesting, I'm using this on a different machine now and 
> getting a 
> > slightly different error when I restore:
> > Unable to handle kernel NULL pointer dereference at virtual address
> > 00000007
> >  printing eip:
> > c0302407
> > *pde = ma 00000000 pa 55555000
> >  [<c012bbee>] worker_thread+0x1ea/0x2e0
> > 
> > The new machine is my laptop, an IBM T41p, but otherwise 
> much the same 
> > (debian sarge, same compiler, etc).
> > 
> > Thanks,
> > Paul Larson
> > On Fri, 2004-12-03 at 10:58, Paul Larson wrote:
> > > I think I've seen something similar to this posted 
> recently, but I 
> > > can't find the message to reply to.  I build xen-2.0-testing from 
> > > source yesterday.  I have debian sarge running on all my domains 
> > > right now.  I was trying the save/restore feature out, 
> and got this 
> > > after restoring and consoling into a session:
> > > 
> > > ************ REMOTE CONSOLE: CTRL-] TO QUIT ******** Unable to 
> > > handle kernel NULL pointer dereference at virtual address
> > > 00000007
> > >  printing eip:
> > > c030e587
> > > *pde = ma 00000000 pa 55555000
> > >  [<c0109f80>] __do_suspend+0x19f/0x1e0
> > > 
> > >  [<c012ca65>] worker_thread+0x22b/0x32f
> > > 
> > >  [<c010a088>] __shutdown_handler+0x0/0x48
> > > 
> > >  [<c01196c0>] default_wake_function+0x0/0x12
> > > 
> > >  [<c01196c0>] default_wake_function+0x0/0x12
> > > 
> > >  [<c012c83a>] worker_thread+0x0/0x32f
> > > 
> > >  [<c0130b89>] kthread+0xa5/0xab
> > > 
> > >  [<c0130ae4>] kthread+0x0/0xab
> > > 
> > >  [<c010f411>] kernel_thread_helper+0x5/0xb
> > > 
> > > Oops: 0002 [#1]
> > > PREEMPT DEBUG_PAGEALLOC
> > > Modules linked in:
> > > CPU:    0
> > > EIP:    0061:[<c030e587>]    Not tainted VLI
> > > EFLAGS: 00010203   (2.6.9-xenU)
> > > EIP is at init_tsc+0x4b/0x9e
> > > eax: c02c6dd4   ebx: 0000c000   ecx: fbffc000   edx: 00000001
> > > esi: 00000010   edi: c0102000   ebp: 00000000   esp: c0039f04
> > > ds: 007b   es: 007b   ss: 0069
> > > Process events/0 (pid: 3, threadinfo=c0038000 task=c003cac0)
> > > Stack: c010eadd 00000000 c0109f80 fbffc000 000001e1 00000063 
> > > 8ed0dd7b
> > > 00000042
> > >        c02c3200 c2dc5000 c02c5d80 00000000 c0038000 00000000 
> > > c012ca65 00000000
> > >        c0039f74 00000000 c003bf10 c0038000 c0038000 c0038000 
> > > c010a088 c0038000 Call Trace:
> > >  [<c010eadd>] time_resume+0x12/0x51
> > > 
> > >  [<c0109f80>] __do_suspend+0x19f/0x1e0
> > > 
> > >  [<c012ca65>] worker_thread+0x22b/0x32f
> > > 
> > >  [<c010a088>] __shutdown_handler+0x0/0x48
> > > 
> > >  [<c01196c0>] default_wake_function+0x0/0x12
> > > 
> > >  [<c01196c0>] default_wake_function+0x0/0x12
> > > 
> > >  [<c012c83a>] worker_thread+0x0/0x32f
> > > 
> > >  [<c0130b89>] kthread+0xa5/0xab
> > > 
> > >  [<c0130ae4>] kthread+0x0/0xab
> > > 
> > >  [<c010f411>] kernel_thread_helper+0x5/0xb
> > > 
> > > Code: 00 00 00 00 00 00 00 00 70 0d 5a f6 00 00 00 00 00 
> 00 00 00 00 
> > > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 21 00 00 00 00 00 
> > > <00> 4e
> > > f7 ac 41 c8 a8 af 41 61 21 4c 41 00 00 00 00 00 00 01 00
> > > 
> > > I'm just starting to play with xen, so I haven't figured 
> out much of 
> > > the debugging stuff yet.  If someone can point me in the right 
> > > direction though, I'd be happy to try to collect some 
> more data if 
> > > there's anything that would be useful.
> > > 
> > > Thanks,
> > > Paul Larson
> > > 
> > > 
> > > 
> > > -------------------------------------------------------
> > > SF email is sponsored by - The IT Product Guide Read 
> honest & candid 
> > > reviews on hundreds of IT Products from real users.
> > > Discover which products truly live up to the hype. Start 
> reading now. 
> > > http://productguide.itmanagersjournal.com/
> > > _______________________________________________
> > > Xen-devel mailing list
> > > Xen-devel@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/xen-devel
> > 
> > 
> > 
> > -------------------------------------------------------
> > SF email is sponsored by - The IT Product Guide Read honest 
> & candid 
> > reviews on hundreds of IT Products from real users.
> > Discover which products truly live up to the hype. Start 
> reading now. 
> > http://productguide.itmanagersjournal.com/
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/xen-devel
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide Read honest & 
> candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now. 
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel
> 
> 


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/

RE: xen problem with save/restore

[Paul Larson]

On Tue, 2004-12-07 at 19:20, Ian Pratt wrote:
> What were you using before? I seem to recall it was 2.0-testing.bk?
Yes, 2.0-testing.bk fails, but xeno-unstable.bk passes.

> There's currently very few changes between unstable and 2.0-testing
Great! Maybe it'll be easier to see what fixed it in xeno-unstable.

> (though this will likely change from next week).  Does it now work
> totally reliably? (it failed every time before, right?). You're using
> the exact same machine and configuration?
Yes, I repulled and recompiled just to be certain.  Under unstable, the
save/restore/reconnect works without anything crashing.  Under
xen-2.0-testing though, I'm getting this when I reconnect:
Unable to handle kernel NULL pointer dereference at virtual address
00000007
 printing eip:
c0302407
*pde = ma 00000000 pa 55555000
 [<c012bbee>] worker_thread+0x1ea/0x2e0

I am seeing this in xm log when I do the destroy:
[2004-12-08 18:38:51 xend] DEBUG (XendDomainInfo:634) Closing console,
domain 1
[2004-12-08 18:38:51 xend] ERROR (XendDomainInfo:627) Domain destroy
failed: VM1Traceback (most recent call last):
  File "/lib/python/xen/xend/XendDomainInfo.py", line 625, in
destroy_domain
    return xc.domain_destroy(dom=self.dom)
error: (3, 'No such process')

Thanks,
Paul Larson



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Keir Fraser]

> On Tue, 2004-12-07 at 19:20, Ian Pratt wrote:
> > What were you using before? I seem to recall it was 2.0-testing.bk?
> Yes, 2.0-testing.bk fails, but xeno-unstable.bk passes.
> 
> > There's currently very few changes between unstable and 2.0-testing
> Great! Maybe it'll be easier to see what fixed it in xeno-unstable.

The latest 2.0-testing and unstable trees only differ in that a new
'blktap' driver was added to the unstable tree, which is not even
built by default. It really looks to be impossible that you could be
seeing different behaviour between the two repositories. :-/

 -- Keir

> > (though this will likely change from next week).  Does it now work
> > totally reliably? (it failed every time before, right?). You're using
> > the exact same machine and configuration?
> Yes, I repulled and recompiled just to be certain.  Under unstable, the
> save/restore/reconnect works without anything crashing.  Under
> xen-2.0-testing though, I'm getting this when I reconnect:
> Unable to handle kernel NULL pointer dereference at virtual address
> 00000007
>  printing eip:
> c0302407
> *pde = ma 00000000 pa 55555000
>  [<c012bbee>] worker_thread+0x1ea/0x2e0
> 
> I am seeing this in xm log when I do the destroy:
> [2004-12-08 18:38:51 xend] DEBUG (XendDomainInfo:634) Closing console,
> domain 1
> [2004-12-08 18:38:51 xend] ERROR (XendDomainInfo:627) Domain destroy
> failed: VM1Traceback (most recent call last):
>   File "/lib/python/xen/xend/XendDomainInfo.py", line 625, in
> destroy_domain
>     return xc.domain_destroy(dom=self.dom)
> error: (3, 'No such process')
> 
> Thanks,
> Paul Larson
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now. 
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

Re: xen problem with save/restore

[Paul Larson]

On Wed, 2004-12-08 at 13:59, Keir Fraser wrote:
> > On Tue, 2004-12-07 at 19:20, Ian Pratt wrote:
> > > What were you using before? I seem to recall it was 2.0-testing.bk?
> > Yes, 2.0-testing.bk fails, but xeno-unstable.bk passes.
> > 
> > > There's currently very few changes between unstable and 2.0-testing
> > Great! Maybe it'll be easier to see what fixed it in xeno-unstable.
> 
> The latest 2.0-testing and unstable trees only differ in that a new
> 'blktap' driver was added to the unstable tree, which is not even
> built by default. It really looks to be impossible that you could be
> seeing different behaviour between the two repositories. :-/
I see, I was running a diff of the trees a while ago and I see what you
mean.  However, it has been very easy to reproduce this under testing,
but on unstable, it never failed for me even once.  So, I installed gcc
3.2 and tried compiling testing with that, and it works now, no crash! 
So it looks like it is a compiler issue after all.

Thanks,
Paul Larson



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/

xen problem with save/restore

[Paul Larson]

I think I've seen something similar to this posted recently, but I can't
find the message to reply to.  I build xen-2.0-testing from source
yesterday.  I have debian sarge running on all my domains right now.  I
was trying the save/restore feature out, and got this after restoring
and consoling into a session:

************ REMOTE CONSOLE: CTRL-] TO QUIT ********
Unable to handle kernel NULL pointer dereference at virtual address
00000007
 printing eip:
c030e587
*pde = ma 00000000 pa 55555000
 [<c0109f80>] __do_suspend+0x19f/0x1e0

 [<c012ca65>] worker_thread+0x22b/0x32f

 [<c010a088>] __shutdown_handler+0x0/0x48

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c012c83a>] worker_thread+0x0/0x32f

 [<c0130b89>] kthread+0xa5/0xab

 [<c0130ae4>] kthread+0x0/0xab

 [<c010f411>] kernel_thread_helper+0x5/0xb

Oops: 0002 [#1]
PREEMPT DEBUG_PAGEALLOC
Modules linked in:
CPU:    0
EIP:    0061:[<c030e587>]    Not tainted VLI
EFLAGS: 00010203   (2.6.9-xenU)
EIP is at init_tsc+0x4b/0x9e
eax: c02c6dd4   ebx: 0000c000   ecx: fbffc000   edx: 00000001
esi: 00000010   edi: c0102000   ebp: 00000000   esp: c0039f04
ds: 007b   es: 007b   ss: 0069
Process events/0 (pid: 3, threadinfo=c0038000 task=c003cac0)
Stack: c010eadd 00000000 c0109f80 fbffc000 000001e1 00000063 8ed0dd7b
00000042
       c02c3200 c2dc5000 c02c5d80 00000000 c0038000 00000000 c012ca65
00000000
       c0039f74 00000000 c003bf10 c0038000 c0038000 c0038000 c010a088
c0038000
Call Trace:
 [<c010eadd>] time_resume+0x12/0x51

 [<c0109f80>] __do_suspend+0x19f/0x1e0

 [<c012ca65>] worker_thread+0x22b/0x32f

 [<c010a088>] __shutdown_handler+0x0/0x48

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c01196c0>] default_wake_function+0x0/0x12

 [<c012c83a>] worker_thread+0x0/0x32f

 [<c0130b89>] kthread+0xa5/0xab

 [<c0130ae4>] kthread+0x0/0xab

 [<c010f411>] kernel_thread_helper+0x5/0xb

Code: 00 00 00 00 00 00 00 00 70 0d 5a f6 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 21 00 00 00 00 00 <00> 4e
f7 ac 41 c8 a8 af 41 61 21 4c 41 00 00 00 00 00 00 01 00

I'm just starting to play with xen, so I haven't figured out much of the
debugging stuff yet.  If someone can point me in the right direction
though, I'd be happy to try to collect some more data if there's
anything that would be useful.

Thanks,
Paul Larson



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/