VirtualBox

Ticket #18802 (new defect)

Opened 4 months ago

Last modified 3 months ago

cannot do "savestate" machine => fixed in SVN/next maintenance

Reported by: kjonca Owned by:
Component: VM control Version: VirtualBox 6.0.10
Keywords: Cc:
Guest type: other Host type: other

Description

When I try to do VBoxManage controlvm hel savestate machine ends in "Aborted" state and in console I have

0%...10%...20%...30%...40%...50%...60%...70%...80%...90%...
Progress state: NS_ERROR_ABORT
VBoxManage: error: Failed to save machine state
VBoxManage: error: The object is not ready
VBoxManage: error: Details: code E_ACCESSDENIED (0x80070005), component SessionMachine, interface IMachine
VBoxManage: error: Context: "RTEXITCODE handleControlVM(HandlerArg*)" at line 622 of file VBoxManageControlVM.cpp

dmesg shows:

[162555.293618] device vbox0 left promiscuous mode
[162561.558953] VDMA[29097]: segfault at 180b ip 00007fcaee190275 sp 00007fc97427eac0 error 4 in VBoxSharedCrOpenGL.so[7fcaee0df000+f5000]
[162561.558962] Code: ff 48 83 c4 20 5b 41 5c 41 5d 41 5e 5d c3 0f 1f 80 00 00 00 00 41 8b 7c 24 04 c7 45 dc 00 00 00 00 48 8d 55 dc be 88 8b 00 00 <41> ff 95 d0 17 00 00 8b 7d dc 85 ff 7f 7d 48 89 de 31 ff e8 93 fe
[162561.814962] vboxnetflt: 41 out of 104 packets were not sent (directed to host)

If I do it with GUI effect is the same (message shows in popup)

Attachments

hel-2019-08-01-19-42-48.log Download (222.4 KB) - added by kjonca 4 months ago.
vbox_config.diff Download (22.8 KB) - added by fbatschu 3 months ago.

Change History

Changed 4 months ago by kjonca

comment:1 Changed 3 months ago by socratis

From your log:

00:03:27.742026 OpenGL Warning: Assertion failed: rc == VINF_SUCCESS=0, file /build/virtualbox-kgswdk/virtualbox-6.0.10-dfsg/src/VBox/GuestHost/OpenGL/state_tracker/state_snapshot.c, line 622

Great! Somebody finally filed a bug report about this, I have about a gazillion posts in the forums that need to be tied to a ticket.

comment:3 Changed 3 months ago by fbatschu

storing some technical information from the OpenSuse ticket work here so it does not get lost:

<snip> Is there a fix for the problem reported in ticket 18802 (https://www.virtualbox.org/ticket/18802)? I have gotten a bug report on the openSUSE Bugzilla ( https://bugzilla.novell.com/show_bug.cgi?id=1146308)reporting this problem with 6.0.10 and a Windows 7 guest. I have not been able to duplicate the problem. I compared the .vbox files for the two systems. The only things that I see that might cause the problem are as follows (The - lines are from the one that fails.):

         <PAE enabled="false"/>
-        <LongMode enabled="true"/>
-        <HardwareVirtExLargePages enabled="true"/>
+        <LongMode enabled="false"/>
+        <HardwareVirtExLargePages enabled="false"/>
       </CPU>
 
and
 
     <StorageControllers>
-      <StorageController name="IDE" type="PIIX4" PortCount="2" useHostIOCache="true" Bootable="true">
-        <AttachedDevice passthrough="false" type="DVD" hotpluggable="false" port="1" device="0">
-          <Image uuid="{a8f03b82-28a7-4748-856c-873fa3c0866f}"/>
+      <StorageController name="SATA" type="AHCI" PortCount="2" useHostIOCache="false" Bootable="true" IDE0MasterEmulationPort="0" IDE0SlaveEmulationPort="1" IDE1MasterEmulationPort="2" IDE1SlaveEmulationPort="3">
+        <AttachedDevice nonrotational="true" type="HardDisk" hotpluggable="false" port="0" device="0">
+          <Image uuid="{924ac17c-9bc7-4e88-ad57-60dc81a42e90}"/>
         </AttachedDevice>
-      </StorageController>
-      <StorageController name="SCSI" type="LsiLogic" PortCount="16" useHostIOCache="false" Bootable="true">
-        <AttachedDevice type="HardDisk" hotpluggable="false" port="0" device="0">
-          <Image uuid="{662215da-93df-4092-aae5-d8a86ed9bcaf}"/>
+        <AttachedDevice passthrough="false" type="DVD" hotpluggable="false" port="1" device="0">
+          <Image uuid="{6607ded7-e5a1-423e-833a-995baf38e84f}"/>
         </AttachedDevice>
       </StorageController>
     </StorageControllers>

In case I missed something important, I have attached the complete diffs for the two .vbox files. <snip end>

I have attached the file from the Poster as : vbox_config.diff

Changed 3 months ago by fbatschu

comment:4 Changed 3 months ago by fbatschu

  • Component changed from other to VM control

comment:5 Changed 3 months ago by socratis

Actually the offending line is none of these that you (or actually Larry) quoted, the offending line is:

      <Display VRAMSize="256" accelerate3D="true" accelerate2DVideo="true"/>

and more specifically the "accelerate3D="true"" part. If no 3D acceleration is involved, the SavedState does not crash. This was due to a security fix in the 3D code in 6.0.8 that didn't quite go as planned.

I already replied to Larry in the mailing list...

comment:6 Changed 3 months ago by aeichner

  • Summary changed from cannot do "savestate" machine to cannot do "savestate" machine => fixed in SVN/next maintenance

Thanks for the report, this should be fixed in the next maintenance release.

comment:7 Changed 3 months ago by socratis

@aeichner,

Do you have a changeset and a revision that this got fixed? Is it available in the Testbuilds?

comment:8 Changed 3 months ago by tenzap

Problem is also present in 5.2.32 and it looks like it is also fixed in Version 5.2.33 r133021

Note: See TracTickets for help on using tickets.

www.oracle.com
ContactPrivacy policyTerms of Use