[Xen-devel] RE: win-pv assertion failed while migrating

["James Harper" <james.harper@xxxxxxxxxxxxxxxx>]

I have just pushed a fix. It was a stupid mistake on my part and I
didn't notice it because Windows 2008 doesn't seem to crash on ASSERT
failures.

Please try it now.

Thanks

James

> -----Original Message-----
> From: alice wan [mailto:wanjia19870902@xxxxxxxxx]
> Sent: Tuesday, 15 February 2011 19:20
> To: James Harper
> Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
> Subject: Re: win-pv assertion failed while migrating
> 
> new problem appeared in the fix and only after migrate once, it came
out.
> 
> XenNet --> XenNet_SuspendResume
> XenNet     New state SR_STATE_RESUMING
> XenNet <-- XenNet_SuspendResume
> XenNet --> XenNet_ResumeWorkItem
> XenPCI --> EvtChn_PdoEventChannelDpc
> 
> *** Assertion failed: !xi->resume_work_item
> ***   Source File: e:\win-pvdrivers\win-pvdrivers.hg\xennet\xennet.c,
line 248
> 
> XenPCI <-- EvtChn_PdoEventChannelDpc
> XenVbd --> DriverEntry
> XenVbd     IRQL = 31
> XenVbd     DriverObject = 00000000, RegistryPath = 00000000
> XenVbd --> XenVbd_HwScsiFindAdapter
> XenVbd     IRQL = 31
> XenVbd     xvdd = 8196B300
> XenVbd     BusInterruptLevel = 28
> XenVbd     BusInterruptVector = 01c
> XenVbd     NumberOfAccessRanges = 1
> XenVbd     RangeStart = 1f9eb000, RangeLength = 00001000
> XenVbd     XEN_INIT_TYPE_VECTORS
> XenVbd     XEN_INIT_TYPE_DEVICE_STATE - 820E93CC
> XenVbd     XEN_INIT_TYPE_RING - ring-ref = 815FC000
> XenVbd     XEN_INIT_TYPE_EVENT_CHANNEL - event-channel = 7
> XenVbd     XEN_INIT_TYPE_READ_STRING - device-type = disk
> XenVbd     device-type = Disk
> XenVbd     XEN_INIT_TYPE_READ_STRING - mode = w
> XenVbd     mode = w
> XenVbd     XEN_INIT_TYPE_READ_STRING - sectors = 104857600
> XenVbd     XEN_INIT_TYPE_READ_STRING - sector-size = 512
> XenVbd     XEN_INIT_TYPE_GRANT_ENTRIES - entries = 11
> XenVbd     qemu_hide_flags_value = 3
> XenVbd     aligned_buffer_data = 8196D274
> XenVbd     aligned_buffer = 8196E000
> XenVbd     ConfigInfo->MaximumTransferLength = 4096
> XenVbd     ConfigInfo->NumberOfPhysicalBreaks = 0
> XenVbd     ConfigInfo->CachesData was initialised to 0
> XenVbd     Dma64BitAddresses not supported
> XenVbd <-- XenVbd_HwScsiFindAdapter
> XenVbd --> XenVbd_HwScsiInitialize
> XenVbd     IRQL = 31
> XenVbd     dump_mode = 1
> XenVbd <-- XenVbd_HwScsiInitialize
> XenVbd     Command = INQUIRY
> XenVbd --> XenVbd_HwScsiResetBus
> XenVbd     IRQL = 31
> XenVbd <-- XenVbd_HwScsiResetBus
> XenVbd     Command = SCSIOP_START_STOP_UNIT
> XenVbd     Command = READ_CAPACITY
> XenVbd <-- DriverEntry
> XenVbd     max_dump_mode_blocks = 8
> XenVbd     max_dump_mode_length = 4096
> 
> *** Assertion failed: srb != NULL
> ***   Source File:
e:\win-pvdrivers\win-pvdrivers.hg\xenvbd\xenvbd_scsiport.c,
> line 988
> 
> XenNet --> XenNet_Shutdown
> XenNet <-- XenNet_Shutdown
> XenPCI     Bug check 0x0000007E (0x80000003, 0x80878113, 0xF88A566C,
> 0xF88A5368)
> 
> 
> 2011/2/15 alice wan <wanjia19870902@xxxxxxxxx>
> 
> 
>       I build from source. today, i'll use latest version to test, if
there's
> any results, i'll let u know
> 
> 
>       2011/2/14 James Harper <james.harper@xxxxxxxxxxxxxxxx>
> 
> 
>               I've found a race in xennet that could cause the
problems you are
> seeing
>               and have pushed a fix, but it's not well tested. From
what you've
> said I
>               think you can build from source. If not, I should be
able to
> upload some
>               binaries tomorrow. Please let me know how your testing
goes.
> 
> 
>               James
> 
>               > -----Original Message-----
>               > From: alice wan [mailto:wanjia19870902@xxxxxxxxx]
> 
>               > Sent: Monday, 14 February 2011 18:01
> 
>               > To: James Harper
>               > Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
> 
>               > Subject: Re: win-pv assertion failed while migrating
>               >
>               > serveral hundred times, not regular and vm blue screen
of death
>               >
>               >
>               >
>               >
>               > 2011/2/14 James Harper <james.harper@xxxxxxxxxxxxxxxx>
>               >
>               >
>               >       >
>               >       > yes, windows 2003-32.
>               >       >
>               >       > my test is migrate vm between host A and host
B
>               >       > continually(xen4.0.0.0+2.6.31.13)
>               >       >
>               >       > Almost after automatic migrate script ran for
one day,
> this
>               problem
>               >       appeared
>               >       >
>               >       > maybe i can also do some test on
xen4.0.1+2.6.32.x to
> check
>               whether it
>               >       matters
>               >       > kernel or pv driver.
>               >
>               >
>               >       So you are saying that it doesn't happen
immediately,
> only after
>               some
>               >       time? Do you know how many times?
>               >
>               >       Maybe I'm leaking a resource somewhere every
time migrate
>               happens and
>               >       eventually it runs out...
>               >
>               >       James
>               >
>               >
> 
> 
> 
> 


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