Bug 18727 - update request: virtualbox 5.0.24 maintenance release
Summary: update request: virtualbox 5.0.24 maintenance release
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 5
Hardware: All Linux
Priority: Normal normal
Target Milestone: ---
Assignee: QA Team
QA Contact:
Whiteboard: MGA5-32-OK MGA5-64-OK advisory
Keywords: validated_update
Depends on: 18726
  Show dependency treegraph
Reported: 2016-06-18 01:05 CEST by Thomas Backlund
Modified: 2016-09-15 19:33 CEST (History)
4 users (show)

See Also:
Source RPM: virtualbox
Status comment:


Description Thomas Backlund 2016-06-18 01:05:58 CEST
This update provides the virtualbox 5.0.22 maintenance release.

Fixes included in this update are:
 * VMM: fixes for certain Intel Atom hosts (bug #14915)
 * VMM: properly restore the complete FPU state for 32-bit guests on 64-bit
   hosts on Intel Sandy Bridge and Ivy Bridge CPUs
 * VMM: new I/O-APIC implementation fixing several bugs and improving the
   performance under certain conditions (bug #15295 and others)
 * VMM: fixed a potential Linux guest panic on AMD hosts
 * VMM: fixed a potential hang with 32-bit EFI guests on Intel CPUs (VT-x
   without unrestricted guest execution)
 * GUI: don't allow to start subsequent separate VM instances
 * GUI: raised upper limit for video capture screen resolution (bug #15432)
 * GUI: warn if the VM has less than 128MB VRAM configured and 3D enabled
 * Main: properly display an error message if the VRDE server cannot be
   enabled at runtime, for example because another service is using the
   same port
 * NAT: Initialize guest address guess for wildcard port-forwarding rules
   with default guest address (bug #15412)
 * VGA: fix for a problem which made certain legacy guests crash under
   certain conditions (bug #14811)
 * ACPI: properly notify the guest if the ACPI battery state changed
 * OVF: fixed import problems for some appliances using an AHCI controller
   created by 3rd party applications
 * SDK: reduced memory usage in the webservice Java bindings
 * Linux hosts: Linux 4.7 fix (bug #15459)
 * Linux Additions: Linux 4.7 fixes (bug #15444)
 * Linux Additions: fix for certain 32-bit guests (5.0.18 regression;
   bug #15320)
 * Linux Additions: fixed mouse pointer offset (5.0.18 regression;
   bug #15324)
 * Linux Additions: fixed X.Org crash after hard guest reset (5.0.18
   regression; bug #15354)
 * Linux Additions: don't stop the X11 setup if loading the shared
   folders module fails (5.0.18 regression)
 * Linux Additions: don't complain if the Drag and Drop service is not
   available on the host




Thomas Backlund 2016-06-18 01:06:19 CEST

Depends on: (none) => 18726

Thierry Vignaud 2016-06-19 07:26:26 CEST

CC: (none) => thierry.vignaud
Depends on: (none) => 18724

Comment 1 David Walser 2016-06-20 10:35:36 CEST
I guess we're going to make a note in the advisory or somewhere about the fact that some users may need to remove /etc/X11/xorg.conf, or at least remove the ModeLine lines from it for X to work.  Yet another upstream X-related regression that we can't do much about.

Otherwise, it's working just fine on Mageia 5 i586 guest and host.  I'll test x86_64 at work.

Whiteboard: (none) => MGA5-32-OK

Comment 2 David Walser 2016-06-20 17:07:19 CEST
Mageia 5 x86_64 host tested fine with Windows 7, Solaris 10, Solaris 11, PC-BSD, Fedora, Ubuntu, OpenSuSE, and Debian guests.
Comment 3 David Walser 2016-06-20 17:09:11 CEST
I'm having some issues with a Mageia 5 i586 guest on Mageia 5 x86_64 host.

The kernel modules for the guest are loaded, but /var/log/Xorg.0.log shows "vboxvideo: kernel driver found, not loading."

So then X fails to start.  If you remove xorg.conf and install x11-driver-video-modesetting, it will use that and work.
Comment 4 Thierry Vignaud 2016-06-20 17:32:40 CEST
Was there a kernel update in the guest?
Does it work better if you update the guest additions and/or the x11-driver-video-vboxvideo package?
David Walser 2016-06-20 17:42:05 CEST

Whiteboard: MGA5-32-OK => MGA5-32-OK MGA5-64-OK feedback

Comment 5 David Walser 2016-06-20 18:24:00 CEST
There's no harm in releasing this one though, since this VM has the same problem now with 5.0.20 guest additions too.  I would like to get to the bottom of this issue if you can help debug it.
Comment 6 David Walser 2016-06-20 21:19:01 CEST
(In reply to Thierry Vignaud from comment #4)
> Was there a kernel update in the guest?
> Does it work better if you update the guest additions and/or the
> x11-driver-video-vboxvideo package?

I have updated the guest to kernel 4.4.13 to match the prebuilt kmod for vboxadditions that's in there too.  There's no apparent reason it shouldn't work.
Comment 7 Thomas Andrews 2016-06-21 03:37:36 CEST
Updated in a MGA5 64-bit host, updated extension pack, started and added guest additions and X11 video driver to a 64-bit Mageia 5 guest, a 32-bit Mageia 5 guest, and a Windows XP guest. (Be aware that the Mageia guests had already been updated to kernel 4.4.13 for a previous test.)

No problems noted.

CC: (none) => andrewsfarm

David Walser 2016-06-21 20:10:05 CEST

Blocks: (none) => 15060, 18724
Depends on: 18724 => (none)

Comment 8 David Walser 2016-06-21 20:20:32 CEST
(In reply to David Walser from comment #3)
> I'm having some issues with a Mageia 5 i586 guest on Mageia 5 x86_64 host.
> The kernel modules for the guest are loaded, but /var/log/Xorg.0.log shows
> "vboxvideo: kernel driver found, not loading."

Now my VM at home is having the same problem :o(

CC: (none) => luigiwalser

Comment 9 David Walser 2016-06-22 00:44:07 CEST
I see this has been rebuilt as 1.1.mga5.  I upgraded my 64-bit host and 32-bit VM at work and there's no change, still no X.
Comment 10 David Walser 2016-06-29 13:48:17 CEST
Now 5.0.24 is out:

Unfortunately, they reverted the fix for the networking performance :o(
Comment 11 David Walser 2016-07-07 16:30:41 CEST
Tested 5.0.24, Comment 3 is unfortunately still valid.

I'll test it with other guests soon, and suspect it'll still work fine.

Summary: update request: virtualbox 5.0.22 maintenance release => update request: virtualbox 5.0.24 maintenance release

Comment 12 David Walser 2016-07-08 16:16:49 CEST
Indeed all other OSs work fine as guests.
Comment 13 Thomas Andrews 2016-07-09 01:50:13 CEST
Went to install this update, saw files for kernel 4.4.13-2 in the repo, so I updated all at the same time.

Everything seemed to work. The update seemed to take a very long time, but then at least one set of vbox kernel modules was being built, also modules built for the existing nvidia 340 driver. After downloading and installing the vbox extpack through the Vbox "check for updates," Mageia 32-bit and 64-bit guests were good to go, including after updating the vbox guest additions and kernel on each.

Windows XP guest also functioning, after the usual Windows XP folderol.
Comment 14 David Walser 2016-07-09 19:26:22 CEST
Working perfectly fine at home with my Mageia 5 i586 host and guest.
Comment 15 Thomas Backlund 2016-07-13 12:03:54 CEST
Should we push this one now, or go for the just released virtualbox 5.1 that has working apic, x2apic and io-apic implementations giving better performance...

if we push this one now, I will still push the new 5.1 series after for tests so it will match what we will have in mga6
Comment 16 David Walser 2016-07-13 12:35:57 CEST
Push it.  It works fine in most cases and is no worse that 5.0.20.  Then we can start testing 5.1.

Keywords: (none) => validated_update
Blocks: 15060, 18724 => (none)
Whiteboard: MGA5-32-OK MGA5-64-OK feedback => MGA5-32-OK MGA5-64-OK
CC: (none) => sysadmin-bugs

Comment 17 Thomas Backlund 2016-07-14 12:39:06 CEST
Updated advisory for virtualbox 5.0.24 added to svn

Whiteboard: MGA5-32-OK MGA5-64-OK => MGA5-32-OK MGA5-64-OK advisory

Comment 18 Mageia Robot 2016-07-14 12:48:04 CEST
An update for this issue has been pushed to the Mageia Updates repository.


Resolution: (none) => FIXED

Note You need to log in before you can comment on or make changes to this bug.