State of GPLPV and Windows 8

State of GPLPV and Windows 8

[Andreas Kinzler]

What is the state of GPLPV and Windows 8? I am currently running tests 
with Windows 8 and GPLPV rev956 and operations (disk IO, net IO) look 
good. There seems to be a problem with shutdown/reboot where Windows 
hangs somewhat and displays a warning about improper shutdown on next 
reboot.

Regards Andreas

Re: State of GPLPV and Windows 8

[Andreas Kinzler]

Maybe the WinDbg log may be helpful. Sees to have some connection to 
hibernation?


1: kd> !analyze -v
*******************************************************************************
*                                                                             
*
*                        Bugcheck 
Analysis                                    *
*                                                                             
*
*******************************************************************************

KMODE_EXCEPTION_NOT_HANDLED (1e)
This is a very common bugcheck.  Usually the exception address pinpoints
the driver/function that caused the problem.  Always note this address
as well as the link date of the driver/image that contains this address.
Arguments:
Arg1: ffffffffc0000005, The exception code that was not handled
Arg2: fffff803c5e81400, The address that the exception occurred at
Arg3: 0000000000000000, Parameter 0 of the exception
Arg4: ffffffffffffffff, Parameter 1 of the exception

Debugging Details:
------------------


READ_ADDRESS:  ffffffffffffffff

EXCEPTION_CODE: (NTSTATUS) 0xc0000005 - The instruction at 0x%08lx 
referenced memory at 0x%08lx. The memory could not be %s.

FAULTING_IP:
nt!memcpy+240
fffff803`c5e81400 f30f6f440af0    movdqu  xmm0,xmmword ptr [rdx+rcx-10h]

EXCEPTION_PARAMETER2:  ffffffffffffffff

BUGCHECK_STR:  0x1E_c0000005_R

DEFAULT_BUCKET_ID:  WIN8_DRIVER_FAULT

PROCESS_NAME:  System

CURRENT_IRQL:  f

TAG_NOT_DEFINED_c000000f:  FFFFF88003622FB0

EXCEPTION_RECORD:  fffffa8003398e10 -- (.exr 0xfffffa8003398e10)
ExceptionAddress: fffffa8003398e18
    ExceptionCode: 000e0002
   ExceptionFlags: 00000001
NumberParameters: 93528968
    Parameter[0]: fffffa8005932388
    Parameter[1]: 0000000000000000
    Parameter[2]: 0000000000000100
    Parameter[3]: 0000000200000001
    Parameter[4]: 0000000000000001
    Parameter[5]: fffffa8003398e70
    Parameter[6]: 0000000fa1662b9d
    Parameter[7]: 0000000fa1662b9d
    Parameter[8]: fffffa8003398f00
    Parameter[9]: fffffa8003398f80
    Parameter[10]: 0000000000000000
    Parameter[11]: 0000000000000000
    Parameter[12]: 0000000000000000
    Parameter[13]: 0000000000000000
    Parameter[14]: 0000000000000000

LAST_CONTROL_TRANSFER:  from fffff803c5f7ddf2 to fffff803c5e86340

STACK_TEXT:
fffff880`0361a308 fffff803`c5f7ddf2 : 00000000`0000001e 
ffffffff`c0000005 fffff803`c5e81400 00000000`00000000 : nt!KeBugCheckEx
fffff880`0361a310 fffff803`c5ee68ad : fffff803`c658ac0e 
00000000`00000000 fffff880`0361a480 00000000`00000000 : 
nt!KiFatalExceptionHandler+0x22
fffff880`0361a350 fffff803`c5ee8633 : 00000000`00000000 
fffff880`03616000 00000000`00003990 fffff880`0361c000 : 
nt!RtlpExecuteHandlerForException+0xd
fffff880`0361a380 fffff803`c5f04abe : fffff880`0361b288 
fffff880`0361afc0 fffff880`0361b288 00000000`00000000 : 
nt!RtlDispatchException+0x44b
fffff880`0361aa90 fffff803`c5e85742 : fffffa80`03398e10 
fffffa80`04bd2101 fffffa80`04bd2190 00000000`00000103 : 
nt!KiDispatchException+0x455
fffff880`0361b150 fffff803`c5e83c4a : fffffa80`00000006 
56530701`00000000 00000000`00000003 00000000`00000000 : 
nt!KiExceptionDispatch+0xc2
fffff880`0361b330 fffff803`c5e81400 : fffff803`c6170a31 
00000000`00000003 00000000`00010000 fffff880`000001b9 : 
nt!KiGeneralProtectionFault+0x10a
fffff880`0361b4c8 fffff803`c6170a31 : 00000000`00000003 
00000000`00010000 fffff880`000001b9 fffff803`c617084f : nt!memcpy+0x240
fffff880`0361b4d0 fffff803`c61709a7 : fffffa80`05e00080 
fffff880`009c3101 fffffa80`047665a0 fffff880`0361b590 : 
nt!PopAddPagesToCompressedPageSet+0x5d
fffff880`0361b540 fffff803`c6173b8b : fffff880`0361b779 
fffff803`c5ebd2d6 fffffa80`05430680 fffffa80`047665a0 : 
nt!PopCompressHiberBlocks+0x87
fffff880`0361b6f0 fffff803`c658ac0e : fffffa80`03477340 
00000000`00000bb2 fffff880`009c3180 00000000`00000000 : 
nt!PopSaveHiberContext+0x7c
fffff880`0361b980 fffff803`c616b0a4 : fffffa80`04276510 
00000000`00000000 00000000`00000000 00000003`00000000 : 
hal!HaliAcpiSleep+0x54f
fffff880`0361ba10 fffff803`c616b26b : fffff803`c6093410 
fffff880`0361bbf0 fffff880`06a61b90 00000000`00f7fc98 : 
nt!PopHandleNextState+0x125
fffff880`0361ba60 fffff803`c5e7b6a8 : fffff880`009c5f00 
fffff880`009c3180 fffffa80`0410b118 fffff880`0361bca0 : 
nt!PopInvokeStateHandlerTargetProcessor+0x3c
fffff880`0361baf0 fffff803`c5ead1b0 : fffffa80`055f6980 
00000000`ffffffff fffffa80`055f6c58 00000000`00000000 : 
nt!KiExecuteAllDpcs+0x198
fffff880`0361bc30 fffff803`c5eb167a : fffff880`009c3180 
fffff880`009c3180 00000000`00000000 fffff880`009cedc0 : 
nt!KiRetireDpcList+0xd0
fffff880`0361bda0 00000000`00000000 : fffff880`0361c000 
fffff880`03616000 00000000`00000000 00000000`00000000 : nt!KiIdleLoop+0x5a


STACK_COMMAND:  kb

FOLLOWUP_IP:
nt!KiFatalExceptionHandler+22
fffff803`c5f7ddf2 cc              int     3

SYMBOL_STACK_INDEX:  1
SYMBOL_NAME:  nt!KiFatalExceptionHandler+22
FOLLOWUP_NAME:  MachineOwner
MODULE_NAME: nt
IMAGE_NAME:  ntkrnlmp.exe
DEBUG_FLR_IMAGE_TIMESTAMP:  50ab0e64
BUCKET_ID_FUNC_OFFSET:  22
FAILURE_BUCKET_ID:  0x1E_c0000005_R_nt!KiFatalExceptionHandler
BUCKET_ID:  0x1E_c0000005_R_nt!KiFatalExceptionHandler
Followup: MachineOwner
---------

Re: State of GPLPV and Windows 8

[Marc Dahlhaus]

[-- Attachment #1.1: Type: text/plain, Size: 623 bytes --]

Give this a try:

 1. Open Control Panel.
 2. Click Hardware and Sound.
 3. Click Power Options.
 4. Click the*Require a password on wakeup* option located on the left-pane.
 5. Click Change settings that are currently unavailable
 6. Under the *Windows Shutdown Settings* located at the end of the
    window,  uncheck the*Enable Hybrid Boot (recommended)* and click 
    the Save Changes button.

If it works, the hybrid boot of win8 isn't supported in the drivers...

Marc

Am 12.02.2013 16:01, schrieb Andreas Kinzler:
>
> Maybe the WinDbg log may be helpful. Sees to have some connection to 
> hibernation?
>
>
--8<--

[-- Attachment #1.2: Type: text/html, Size: 1220 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

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

Re: State of GPLPV and Windows 8

[James Harper]

> 
> Maybe the WinDbg log may be helpful. Sees to have some connection to
> hibernation?
> 

Were you actually hibernating at the time?

James

Re: State of GPLPV and Windows 8

[Andreas Kinzler]

I think Paul is right when he says

>Windows 8 doesn't actually shut down when you tell it to... It terminates all sessions
>apart from 0 (where services live) and then hibernates, so 'shutdown' is actually a lot
>more complex and error prone than it was before.

I clicked "shutdown" - Windows hibernates.



> James Harper <mailto:james.harper@bendigoit.com.au>
> Tuesday, 12. February 2013 23:19
>
> Were you actually hibernating at the time?
>
> James
> Andreas Kinzler <mailto:ml-xen-devel@hfp.de>
> Tuesday, 12. February 2013 16:01
>
> Maybe the WinDbg log may be helpful. Sees to have some connection to 
> hibernation?
>
>
> 1: kd> !analyze -v
> ******************************************************************************* 
>
> *                                                                             
> *
> *                        Bugcheck 
> Analysis                                    *
> *                                                                             
> *
> ******************************************************************************* 
>
>

Re: State of GPLPV and Windows 8

[James Harper]

> 
> What is the state of GPLPV and Windows 8? I am currently running tests
> with Windows 8 and GPLPV rev956 and operations (disk IO, net IO) look
> good. There seems to be a problem with shutdown/reboot where Windows
> hangs somewhat and displays a warning about improper shutdown on next
> reboot.
> 

I have only tested with Windows 2012 but it shouldn't be any different to 8. My testing was pretty much limited to 'does it boot' testing, but I don't recall having any problems with shutdown or anything.

Are you interested in trying the very latest version? I've been doing a fairly major update in the way the child devices talk to the bus driver and it is now at a point where everything should be working but I haven't done a lot of testing yet.

James

Re: State of GPLPV and Windows 8

[Paul Durrant]

Windows 8 doesn't actually shut down when you tell it to... It terminates all sessions apart from 0 (where services live) and then hibernates, so 'shutdown' is actually a lot more complex and error prone than it was before.

  Paul 

> -----Original Message-----
> From: xen-devel-bounces@lists.xen.org [mailto:xen-devel-
> bounces@lists.xen.org] On Behalf Of James Harper
> Sent: 12 February 2013 22:19
> To: Andreas Kinzler; xen-devel@lists.xensource.com
> Subject: Re: [Xen-devel] State of GPLPV and Windows 8
> 
> >
> > What is the state of GPLPV and Windows 8? I am currently running tests
> > with Windows 8 and GPLPV rev956 and operations (disk IO, net IO) look
> > good. There seems to be a problem with shutdown/reboot where
> Windows
> > hangs somewhat and displays a warning about improper shutdown on next
> > reboot.
> >
> 
> I have only tested with Windows 2012 but it shouldn't be any different to 8.
> My testing was pretty much limited to 'does it boot' testing, but I don't recall
> having any problems with shutdown or anything.
> 
> Are you interested in trying the very latest version? I've been doing a fairly
> major update in the way the child devices talk to the bus driver and it is now
> at a point where everything should be working but I haven't done a lot of
> testing yet.
> 
> James
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel

Re: State of GPLPV and Windows 8

[James Harper]

> 
> Windows 8 doesn't actually shut down when you tell it to... It terminates all
> sessions apart from 0 (where services live) and then hibernates, so
> 'shutdown' is actually a lot more complex and error prone than it was before.
> 

I've gotten it all working now, but the restore from hibernate (which appears to use qemu) is slow. The hibernate on a 1GB machine takes seconds but the restore takes over a minute.

James

Re: State of GPLPV and Windows 8

[Paul Durrant]

> -----Original Message-----
> From: James Harper [mailto:james.harper@bendigoit.com.au]
> Sent: 19 February 2013 12:10
> To: Paul Durrant; Andreas Kinzler; xen-devel@lists.xensource.com
> Subject: RE: State of GPLPV and Windows 8
> 
> >
> > Windows 8 doesn't actually shut down when you tell it to... It terminates all
> > sessions apart from 0 (where services live) and then hibernates, so
> > 'shutdown' is actually a lot more complex and error prone than it was
> before.
> >
> 
> I've gotten it all working now, but the restore from hibernate (which appears
> to use qemu) is slow. The hibernate on a 1GB machine takes seconds but the
> restore takes over a minute.
> 

Yes, I've noticed Windows 8 seemingly taking a lot longer to load/restore than previous versions but never dug into it. I'm assuming it's going to be some suboptimal int13 handing, or maybe Windows 8 is bypassing the f/w and driving the IDE device itself?

  Paul

Re: State of GPLPV and Windows 8

[James Harper]

> > >
> > > Windows 8 doesn't actually shut down when you tell it to... It terminates
> > > all
> > > sessions apart from 0 (where services live) and then hibernates, so
> > > 'shutdown' is actually a lot more complex and error prone than it was
> > > before.
> > >
> >
> > I've gotten it all working now, but the restore from hibernate (which
> > appears
> > to use qemu) is slow. The hibernate on a 1GB machine takes seconds but
> > the restore takes over a minute.
> >
> 
> Yes, I've noticed Windows 8 seemingly taking a lot longer to load/restore
> than previous versions but never dug into it. I'm assuming it's going to be
> some suboptimal int13 handing, or maybe Windows 8 is bypassing the f/w
> and driving the IDE device itself?
> 

Direct IDE seems a bit risky... it's all black box stuff though so anything goes :)

I've been doing a bit more reading and it seems that win8 is able to involve the storport drive in the restore process too, I think. From http://msdn.microsoft.com/en-us/library/windows/hardware/ff563901%28v=vs.85%29.aspx

DumpMode
A value indicating the use of the miniport during dump mode. This member is included starting with Windows 8. It can have one of the following values.
DUMP_MODE_CRASH - The miniport in dump mode is used for a crashdump.
DUMP_MODE_HIBER - The miniport in dump mode is used for a hibernation.
DUMP_MODE_MARK_MEMORY - The miniport in dump mode is used for marking required memory.
DUMP_MODE_RESUME - The miniport in dump mode is used for a resume from hibernation.

In order to play with that I'd need to download the win8 DDK (the visual studio based one) which I'm not real keen on doing just yet...

James

Re: State of GPLPV and Windows 8

[Paul Durrant]

> -----Original Message-----
> 
> Direct IDE seems a bit risky... it's all black box stuff though so anything goes :)
> 

True. IIRC Microsoft removed all their int10 code from Win8 though... was just wondering if they did the same thing with int13. 

> I've been doing a bit more reading and it seems that win8 is able to involve
> the storport drive in the restore process too, I think. From
> http://msdn.microsoft.com/en-
> us/library/windows/hardware/ff563901%28v=vs.85%29.aspx
> 
> DumpMode
> A value indicating the use of the miniport during dump mode. This member is
> included starting with Windows 8. It can have one of the following values.
> DUMP_MODE_CRASH - The miniport in dump mode is used for a crashdump.
> DUMP_MODE_HIBER - The miniport in dump mode is used for a hibernation.
> DUMP_MODE_MARK_MEMORY - The miniport in dump mode is used for
> marking required memory.
> DUMP_MODE_RESUME - The miniport in dump mode is used for a resume
> from hibernation.
> 

Interesting. I'll have a read too.

> In order to play with that I'd need to download the win8 DDK (the visual
> studio based one) which I'm not real keen on doing just yet...
> 

I sympathize. I've been through the pain. I wouldn't want to do it again.

Cheers,

    Paul

Re: State of GPLPV and Windows 8

[Andreas Kinzler]

In the meantime I disabled hibernation (powercfg.exe -h off) and it 
seems to work now. Which revision from your repo do you wish to be 
tested? I am interested in testing, but my time is somewhat limited - I 
cannot make any promise.

As I said in my original post, rev956 works quite well with its core 
operations: net IO and disk IO.

Regards Andreas
> James Harper <mailto:james.harper@bendigoit.com.au>
> Tuesday, 12. February 2013 23:19
>
> I have only tested with Windows 2012 but it shouldn't be any different 
> to 8. My testing was pretty much limited to 'does it boot' testing, 
> but I don't recall having any problems with shutdown or anything.
>
> Are you interested in trying the very latest version? I've been doing 
> a fairly major update in the way the child devices talk to the bus 
> driver and it is now at a point where everything should be working but 
> I haven't done a lot of testing yet.
>
> James
> Andreas Kinzler <mailto:ml-xen-devel@hfp.de>
> Tuesday, 12. February 2013 14:38
> What is the state of GPLPV and Windows 8? I am currently running tests 
> with Windows 8 and GPLPV rev956 and operations (disk IO, net IO) look 
> good. There seems to be a problem with shutdown/reboot where Windows 
> hangs somewhat and displays a warning about improper shutdown on next 
> reboot.
>
> Regards Andreas
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel

Re: State of GPLPV and Windows 8

[James Harper]

> 
> In the meantime I disabled hibernation (powercfg.exe -h off) and it
> seems to work now. Which revision from your repo do you wish to be
> tested? I am interested in testing, but my time is somewhat limited - I
> cannot make any promise.
> 
> As I said in my original post, rev956 works quite well with its core
> operations: net IO and disk IO.
> 

I've just fixed a hibernate problem under 4.2 and have just installed windows 8 to test the hibernate issues (previously I'd been testing Windows 2012 which doesn't have this feature I assume).

Once I've got those sorted I'll upload what I've got.

James