CentOS OS latest Update (fixed in 6.1.6 and 6.0.20)

[PaulL0]

While running the latest Centos version when I checked the software there was a software update message informing me an update for the OS is to be installed. The attached errors were received. As per Virtual GuruBox Meditation "A critical error has occurred while running the machine and the machine execution has been stopped. See attached log file VBox1 and the image file VBox.png.I have been running Virtual Box for a while and this is the first time I have come across this. I am not a software developer I am a newbie. Thanks for your help.

[aeichner]

The attached log doesn't show any guru meditation, please provide the correct VBox.log.

[PaulL0]

Guru-Log_File

[PaulL0]

Hi A Eichner,

I have attached what I think is the correct file. I ran a search for the word Guru see below|: Line 591: 00:00:12.653888 EMR3Init: fIemExecutesAll=false fGuruOnTripleFault=true

Line 1820: 01:07:58.226955 Changing the VM state from 'RUNNING' to 'GURU_MEDITATION' Line 1821: 01:07:58.227073 Console: Machine state changed to 'GuruMeditation' Line 1824: 01:07:58.227594 !! VCPU0: Guru Meditation -1607 (VERR_PGM_HANDLER_NOT_FOUND) Line 5866: 01:39:27.469293 GUI: User request to power VM off on Guru Meditation. Line 5871: 01:39:27.555791 Changing the VM state from 'GURU_MEDITATION' to 'POWERING_OFF'

I hope this is of help to you.

Thank you for your help.

Kind regards,

Paul

[PaulL0]

Hi A Eichner,

Can you provide an update to the information I submitted in answer to your request. Thanks, Paul

[Frank Batschulat (Oracle)]

So the log file shows:

https://www.215389.xyz/attachment/ticket/19361/Centos-2020-02-28-22-19-54.log

1388	01:03:28.072257 VMMDev: Guest Log: BIOS: VirtualBox 6.1.4
1389	01:03:28.072446 PCI: Setting up resources and interrupts
1390	01:03:28.072801 AssertLogRel F:\tinderbox\win-rel\src\VBox\Devices\Graphics\DevVGA.cpp(5648) int __cdecl vgaR3PciIORegionVRamMapUnmap(struct PDMDEVINSR3 *,struct PDMPCIDEV *,unsigned int,unsigned __int64,unsigned __int64,enum PCIADDRESSSPACE): RT_SUCCESS_NP(rc)
1391	01:03:28.097751 VERR_PGM_HANDLER_PHYSICAL_CONFLICT (-1603) - Attempt to register an access handler for a physical range of which a part was already handled.
1392	01:03:28.097995 ERROR [COM]: aRC=VBOX_E_VM_ERROR (0x80bb0003) aIID={4680b2de-8690-11e9-b83d-5719e53cf1de} aComponent={DisplayWrap} aText={Could not take a screenshot (VERR_NOT_SUPPORTED)}, preserve=false aResultDetail=-37 

followed by:

1451	01:03:37.861886 AssertLogRel F:\tinderbox\win-rel\src\VBox\Devices\Graphics\DevVGA.cpp(5648) int __cdecl vgaR3PciIORegionVRamMapUnmap(struct PDMDEVINSR3 *,struct PDMPCIDEV *,unsigned int,unsigned __int64,unsigned __int64,enum PCIADDRESSSPACE): RT_SUCCESS_NP(rc)
1452	01:03:37.862006 VERR_PGM_HANDLER_PHYSICAL_CONFLICT (-1603) - Attempt to register an access handler for a physical range of which a part was already handled.
1453	01:03:37.867476 AssertLogRel F:\tinderbox\win-rel\src\VBox\Devices\Graphics\DevVGA.cpp(5648) int __cdecl vgaR3PciIORegionVRamMapUnmap(struct PDMDEVINSR3 *,struct PDMPCIDEV *,unsigned int,unsigned __int64,unsigned __int64,enum PCIADDRESSSPACE): RT_SUCCESS_NP(rc)
1454	01:03:37.867632 VERR_PGM_HANDLER_PHYSICAL_CONFLICT (-1603) - Attempt to register an access handler for a physical range of which a part was already handled.
1455	01:03:37.872809 AssertLogRel F:\tinderbox\win-rel\src\VBox\Devices\Graphics\DevVGA.cpp(5648) int __cdecl vgaR3PciIORegionVRamMapUnmap(struct PDMDEVINSR3 *,struct PDMPCIDEV *,unsigned int,unsigned __int64,unsigned __int64,enum PCIADDRESSSPACE): RT_SUCCESS_NP(rc)
1456	01:03:37.872940 VERR_PGM_HANDLER_PHYSICAL_CONFLICT (-1603) - Attempt to register an access handler for a physical range of which a part was already handled. 

1463	01:03:37.897171 AssertLogRel F:\tinderbox\win-rel\src\VBox\Devices\Graphics\DevVGA.cpp(5648) int __cdecl vgaR3PciIORegionVRamMapUnmap(struct PDMDEVINSR3 *,struct PDMPCIDEV *,unsigned int,unsigned __int64,unsigned __int64,enum PCIADDRESSSPACE): RT_SUCCESS_NP(rc)
1464	01:03:37.897311 VERR_PGM_HANDLER_PHYSICAL_CONFLICT (-1603) - Attempt to register an access handler for a physical range of which a part was already handled. 

and finally:

1820	01:07:58.226955 Changing the VM state from 'RUNNING' to 'GURU_MEDITATION'
1821	01:07:58.227073 Console: Machine state changed to 'GuruMeditation'
1822	01:07:58.227591 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
1823	01:07:58.227593 !!
1824	01:07:58.227594 !!         VCPU0: Guru Meditation -1607 (VERR_PGM_HANDLER_NOT_FOUND)
1825	01:07:58.227608 !!
1826	01:07:58.227616 !!
1827	01:07:58.227616 !! {mappings, <NULL>}
1828	01:07:58.227618 !!
1829	01:07:58.227629 !!
1830	01:07:58.227630 !! {hma, <NULL>}
1831	01:07:58.227631 !! 

[Frank Batschulat (Oracle)]

This sounds a lot like an internal bug we've just fixed today:

Guru Meditation -1607 (VERR_PGM_HANDLER_NOT_FOUND) trying to force a crash dump on Linux

Any chance that your Linux guest VM was trying to a crash dump after a kernel panic perhaps?

[Klaus Espenlaub]

No, it's the "start VM from saved state, reset VM, guru" variant. Which is not exactly the same, but ultimately runs into very similar problems. We should be able to provide builds with a fix rather soon... likely next week.

[Frank Batschulat (Oracle)]

So this is apparently not what I suspected to be fixed by revision r136435 in trunk but instead it is what Klaus said, fixed with revision r136428 in trunk.

I will provide public test builds for the 6.1 (backport) and Trunk train asap next week containing both fixes and will update this bug once they become available.

[Frank Batschulat (Oracle)]

So more fix details:

in trunk we have the following 2 related changesets:

r136435 VMSVGA: Do not needlessly register VGA LFB handler when updating framebuffer mapping
r136428 VMSVGA: Unregister VRAM handler after loading state if necessary

Those have been backported to the 6.1 branch with revisions r136455 and r136456 respectively.

I'll produce test builds for trunk and the 6.1 branch asap.

[Frank Batschulat (Oracle)]

WORKAROUND:

trunk test builds containing above mentioned revisions with the fix are available on:

https://www.215389.xyz/wiki/Testbuilds

Section: Development snapshots

I will update once the 6.1 branch testbuilds are available.

[Frank Batschulat (Oracle)]

We have also released Linux test builds for the 6.0.X branch that contain mentioned fix for:

VMSVGA: Do not needlessly register VGA LFB handler when updating framebuffer mapping

which fixes the Linux crash dump reboot problem in revision r136470 on:

https://www.215389.xyz/wiki/Testbuilds

Section: Latest 6.0.x test builds

Linux 64-bit	6.0.x revision 136475
Linux EL6 64-bit	6.0.x revision 136475
Linux EL7 64-bit	6.0.x revision 136475
Linux EL8 64-bit	6.0.x revision 136475
Extension Pack	6.0.x revision 136470

[Frank Batschulat (Oracle)]

this has been fixed in the public releases 6.1.6 and 6.0.20.