VirtualBox

Ticket #3525 (closed defect: fixed)

Opened 5 years ago

Last modified 5 years ago

PGMPOOLKIND_FREE Assertion

Reported by: costing Owned by:
Priority: critical Component: other
Version: VirtualBox 2.1.4 Keywords: Assertion Failed pgmPoolGetPage
Cc: Guest type: other
Host type: other

Description

After a couple of weeks of running I got this exception:

!!R0-Assertion Failed!!
Expression: pPage && pPage->enmKind != PGMPOOLKIND_FREE
Location  : /home/vbox/vbox-2.1.4/src/VBox/VMM/PGMInternal.h(4569) PGMPOOLPAGE* pgmPoolGetPage(PGMPOOL*, RTHCPHYS)
HCPhys=00000003e2589000 pPage=00007f00b46d56c8 idx=248
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!
!!                 Guru Meditation -2701 (VERR_VMM_RING0_ASSERTION)

Attachments

VBox.log Download (151.5 KB) - added by costing 5 years ago.
VBox_Mac.log Download (111.0 KB) - added by jeyrich 5 years ago.
another occurence on a a Mac (added by jeyrich)
VBox-vm1.log Download (218.7 KB) - added by shjorth 5 years ago.
Solaris Guest 1 Guru Meditation
VBox-vm2.log Download (180.3 KB) - added by shjorth 5 years ago.
Solaris Guest 2 Guru Meditation

Change History

Changed 5 years ago by costing

comment:1 Changed 5 years ago by frank

  • Summary changed from R0-Assertion Failed in pgmPoolGetPage to PGMPOOLKIND_FREE Assertion

The interesting thing is that your host processor is quite fast but the log file contains a lot of complaints that ATA commands took a long time to complain. Though this shouldn't be related to the PGMPOOLKIND_FREE assertion...

comment:2 Changed 5 years ago by costing

Yes, we had quite poor hard drives (5400 RPM) that even in RAID 5 of 4 of them were performing bad. The guests were also crashing from time to time complaining about IO errors, FS was remounted r/o ... all kind of crap. We have assumed they were caused by the poor IO so we've replaced the disks recently. If we see the errors again we'll come back with bug reports.

Changed 5 years ago by jeyrich

another occurence on a a Mac (added by jeyrich)

comment:3 Changed 5 years ago by jeyrich

I have added a log file for a similar occurence on a Mac Pro. It happened on the first day of installing VBox, though. Could disabling Nested Paging and/or Hardware Virtualization help to circumvent the problem for now?

comment:4 Changed 5 years ago by jeyrich

Oh, and I found this in system.log after the crash

Mar 28 00:06:38 hostname kernel[0]: Resetting IOCatalogue.
Mar 28 00:06:38 hostname kextd[757]: 0 cached, 444 uncached personalities to catalog
Mar 28 00:15:00 hostname kernel[0]: ^M
Mar 28 00:15:00 hostname kernel[0]: !!Assertion Failed!!^M
Mar 28 00:15:00 hostname kernel[0]: Expression: pPage && pPage->enmKind != PGMPOOLKIND_FREE^M
Mar 28 00:15:00 hostname kernel[0]: Location  : /Volumes/ServerHD2/Users2/vbox/tinderbox/2.1-mac-rel/src/VBox/VMM/PGMInternal.h(4569) PGMPOOLPAGE* pgmPoolGetPage(PGMPOOL*, RTHCPHYS)^M
Mar 28 00:15:00 hostname kernel[0]: HCPhys=0000000000000000 pPage=00000000 idx=0
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x5c1
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x24f
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x405
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x6dc
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x86
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x62a
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x3ad
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x710
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x50b
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x40a
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x332
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x412
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0xda
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x24c
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0x4e
Mar 28 00:15:37 hostname kernel[0]: PGMR0DynMapTermVM: 1 dangling refs to 0xed

comment:5 Changed 5 years ago by neilcawse

Notebook:

This guru meditation failure has just occured on my Ubuntu XPS M1530 running Ubuntu Jaunty Beta and VirtualBox 2.

Drive:

I have a Vertex SSD Drive from OCZ running EXT 4.

Suspected Problem:

I suspect it may be a ssd drive firmware issue as some of my colleges have been reporting a few problems. Im logging this in case it is not or maybe an EXT 4 problem or something else.

Log:

!!Assertion Failed![[BR]]

Expression: pPage && pPage->enmKind != PGMPOOLKIND_FREE
Location : /home/vbox/vbox-2.2.0/src/VBox/VMM/PGMInternal.h(4274) PGMPOOLPAGE* pgmPoolGetPage(PGMPOOL*, RTHCPHYS)
HCPhys=000000002c9e8000 pPage=00000000 idx=0

06:33:26.058 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
06:33:26.058 !!
06:33:26.058 !! Guru Meditation 1103 (VINF_EM_DBG_HYPER_ASSERTION)
06:33:26.059 !!
06:33:26.060 !!Assertion Failed!!
06:33:26.060 Expression: pPage && pPage->enmKind != PGMPOOLKIND_FREE
06:33:26.060 Location : /home/vbox/vbox-2.2.0/src/VBox/VMM/PGMInternal.h(4274) PGMPOOLPAGE* pgmPoolGetPage(PGMPOOL*, RTHCPHYS)
06:33:26.060 HCPhys=000000002c9e8000 pPage=00000000 idx=0
06:33:26.060 !!
06:33:26.060 !! EIP=91c01580 NOTRAP
06:33:26.060 !! EIP in VMMGC.gc (91bd8000) at rva 29580 near symbols:
06:33:26.060 !! 91c014e0 rva 000294e0 off 000000a0 pgmPoolGetPageByHCPhys
06:33:26.060 !! 91c01590 rva 00029590 off -00000010 pgmPoolTrackFlushGCPhysPTsSlow
06:33:26.072 !! fff8:91c01580 90 nop
06:33:26.073 !!
06:33:26.073 !!
06:33:26.073 !!

Changed 5 years ago by shjorth

Solaris Guest 1 Guru Meditation

Changed 5 years ago by shjorth

Solaris Guest 2 Guru Meditation

comment:6 Changed 5 years ago by shjorth

Experienced PGMPOOLKIND_FREE Assertions for two Solaris 10u6 guests running concurrently (VBox-vm1.log & VBox-vm2.log attached) running on the same Ubuntu Linux host with Virtualbox 2.2. One of the Solaris 10u6 guests was performing a code build at the time (VBox-vm2.log) so there would have been a fair amount of I/O and CPU load.

This is the second occurrence of PGMPOOLKIND_FREE causing the two VM's to crash in two days.

EXT3 used as the host file-system.

The two remaining guest VM's (one Solaris 10u6 and the other a Redhat 4 EL guest) are still running fine.

comment:7 Changed 5 years ago by frank

  • Status changed from new to closed
  • Resolution set to fixed

This problem should be actually fixed with VirtualBox 2.2.2. Please reopen if this problem still occurs with VirtualBox 2.2.2.

Note: See TracTickets for help on using tickets.

www.oracle.com
ContactPrivacy policyTerms of Use