| Summary: | VirtualBox 5.1.10 | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | David Walser <luigiwalser> |
| Component: | Security | Assignee: | QA Team <qa-bugs> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Normal | CC: | andrewsfarm, brtians1, davidwhodgins, herman.viaene, jim, lewyssmith, lists.jjorge, lovaren, mageia, mageia, sysadmin-bugs, tarazed25, tmb, wilcal.int, youpburden |
| Version: | 5 | Keywords: | validated_update |
| Target Milestone: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| URL: | http://lwn.net/Vulnerabilities/704586/ | ||
| Whiteboard: | MGA5-64-OK advisory mga5-32-ok | ||
| Source RPM: | virtualbox-5.1.10-1.1.mga5 | CVE: | CVE-2016-5501, CVE-2016-5538, CVE-2016-5605, CVE-2016-5608, CVE-2016-5610, CVE-2016-5611, CVE-2016-5613, CVE-2016-6304 |
| Status comment: | |||
| Bug Depends on: | 19577 | ||
| Bug Blocks: | |||
|
Description
David Walser
2016-08-18 15:22:27 CEST
I don't have an i586 host to test anymore since I converted my workstation at home to x86_64 last weekend. x86_64 host with Mageia 5 guest works fine. I had to change the audio setting in the guest to Intel HD Audio and then in the guest run "alsactl init" and turn the volumes on the mixer channels all the way up to restore sound after the update. It stutters at the beginning of every new track (playing with Amarok). Playing with ogg123 sounds completely chopped up and horrible. Failure in my i586 install with a Mageia 5 guest. I had not booted into this install for several weeks, so there were over 100 updates waiting, including the update to the latest server kernel and vbox 5.1.2. I completed those updates first and then rebooted, but did not check operation of vbox 5.1.2, or get the extension pack. (I should have.) I then updated vbox to 5.1.4, and rebooted once more. The 5.1.4 extension pack downloaded and installed normally. But when I went to start my Mageia 5 guest to update that, it appeared to begin normally, but a progress bar box with the bar set at 0% opened, with a busy cursor inside the vbox guest. After around a minute (I did not time it) a second progress bar box opened with progress at 20%, with the busy cursor still there. I waited another 4-5 minutes with no change, then attempted to close the guest with the "X" in the corner, with no response. No response to an attempt to close the progress bar box, either. The vbox gui could be put on top, but was also completely unresponsive. I could use the host normally to reboot, which I did, and vbox showed the Mageia guest had been "aborted." I attempted to run the guest again with the same result as above. Hardware: ASRock M2+ motherboard, Athlon X2 7750 processor, nvidia Geforce 9800GT video card, 8GB RAM. Mageia 5 guest is configured to use the maximum allowed for a 32-bit guest, as a sort of a test of the 32-bit server kernel's ability to use more than 4GB. This hardware functions perfectly in 64-bit with a Mageia 5 guests (both 32-bit and 64-bit) in vbox 5.1.2. After this test I am hesitant to try 5.1.4 on it, because I do need it for production work from time to time. CC:
(none) =>
andrewsfarm Question: Is it possible that I need one or more of the updates that have been validated but not yet pushed? MGA5-32 on Acer D620 No installation issues. Tried to install Mag4-32 on a virtual machine (this version ran previously well on the real machine). Error at the end of the installation: mkinitrd failed (mkinitrd -v -f /boot/initrd-3.12.21-desktop-2.mga4.img 3.12.21-desktop-2.mga4 CC:
(none) =>
herman.viaene On real hardware, M5, KDE, 64-bit
Package(s) under test:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
default install of:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
[root@localhost wilcal]# uname -a
Linux localhost 4.4.16-desktop-1.mga5 #1 SMP Tue Jul 26 09:23:40 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.16-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.2-2.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.2-1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.2-2.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Created clients:
M5 i586 Gnome Live-DVD runs as a Vbox client. Screen sizes are correct.
M6 x86_64 Plasma Live-DVD installed, updates and runs as a Vbox client. Screen sizes are correct.
install from updates_testing:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
from updates_testing
[root@localhost wilcal]# uname -a
Linux localhost 4.4.16-desktop-1.mga5 #1 SMP Tue Jul 26 09:23:40 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.16-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.4-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.4-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.4-1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.4-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.4-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.4-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[wilcal@localhost ~]$ lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Clients created Pre-update:
M5 i586 Gnome Live-DVD still runs as a Vbox client. Screen sizes are correct. sound ok
M6 x86_64 Plasma Live-DVD still runs as a Vbox client. Screen sizes are correct.
Clients created Post-update:
M6 x86_64 Gnome Live-DVD runs as a Vbox client. Screen sizes are correct.
M5 x86_64 KDE CI installed, updates and runs as a Vbox client. Screen sizes are correct. sound ok
did not need the /etc/rc.d/rc.local "ethtool -K enp0s3 gro off" thingy
For some reason Clients after update seemed quicker.
Test platform:
Intel Core i7-2600K Sandy Bridge 3.4GHz
GIGABYTE GA-Z68X-UD3-B3 LGA 1155 MoBo
GIGABYTE GV-N440D3-1GI Nvidia GeForce GT 440 (Fermi) 1GB
RTL8111/8168B PCI Express 1Gbit Ethernet
DRAM 16GB (4 x 4GB)
Mageia 5 64-bit, Nvidia driverCC:
(none) =>
wilcal.int Tried MGA4 installation again (see Comment 5), this time with almost all defaults. Installation went OK, but was seriously hampered by poor mouse handling: when approaching the right side of the virtual machine screen, the mouse pointer jumped out of this window to the host screen, making most of the time the "Next" button unreachable. I was able to get the normal behavior back by pressing the right "CTRL" twice, and then mouse operation was OK until the installer draws a new screen, and then the hoopla starts all over again. Once the installation was finished, this mouse problem did not appear again. But then, the Wifi adapter was not passed (choosen bridged adapter) to the virtual machine, so not able to test the Wifi printer configuration. All is quite sluggish on this old laptop, I wonder whether it is not too restricted to run virtualbox successfully. MGA5-64 on Lenovo B50 KDE Noinstallation issues. Tried same MGA4 installation as Comment 7. Got the same mouse problems and the numlock was not respected in the virtual machine. Also same problem with Wifi, the device is not visible in the virtual machine. This machine is much quicker and better equiped in RAM as the 32 bit I tested above, but the problems are the same. I'm gonna give this an "OK" for x86_64. IMO i586 installs of Vbox on real i586 platforms is unusable. Whiteboard:
(none) =>
MGA5-64-OK Removing the OK for now due to comment 2, etc. CC:
(none) =>
davidwhodgins M5 as a VB guest 64bit, fully updated, installed from updates_testing: x11-driver-video-vboxvideo-5.1.4-1.mga5 vboxaddition-kernel-4.4.16-desktop-1.mga5 Playing ogg audio works for me with: parole and vlc. pulseaudio works. (vbox 5.1.2: pulseaudio was broken, had to switch to also for audio to work) No further issues. Auto resize guest display: works. Shared folders: works. Shared clipboard: works. CC:
(none) =>
shybluenight I think this bug should be closed as wontfix, and a new bug created for vb 5.1.6. Specifically, for the items from the changelog ... Linux Additions: Linux 4.8 fix for the kernel display driver (bugs #15890 and #15896) Linux Additions: don't load the kernel modules provided by the Linux distribution but load the kernel modules from the official Guest Additions package instead (bug #15324) This is on hold for now, until we can get 5.1.6 fixed/built or maybe we'll have to wait for 5.1.8. Whiteboard:
(none) =>
feedback 5.1.6 has now been released to the public. This is a maintenance release. Can someone please add it to Cauldron? CC:
(none) =>
hamnisdude
Kristoffer Grundström
2016-09-19 08:28:49 CEST
Source RPM:
virtualbox-5.1.2-1.mga5.src.rpm =>
(none) Don't do that Kristoffer, if you read this bug report you'll see that it is intended for a Mageia 5 update. Version:
Cauldron =>
5 Update to 5.1.6 is built. If it works OK, it should go out with the kernel update. Advisory: This update provides the VirtualBox 5.1.6 maintenance release. Reference: https://www.virtualbox.org/wiki/Changelog SRPMS: virtualbox-5.1.6-1.mga5.src.rpm kmod-virtualbox-5.1.6-1.mga5.src.rpm kmod-vboxadditions-5.1.6-1.mga5.src.rpm i586: dkms-vboxadditions-5.1.6-1.mga5.noarch.rpm dkms-virtualbox-5.1.6-1.mga5.noarch.rpm python-virtualbox-5.1.6-1.mga5.i586.rpm vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.i586.rpm vboxadditions-kernel-4.4.22-desktop586-1.mga5-5.1.6-1.mga5.i586.rpm vboxadditions-kernel-4.4.22-server-1.mga5-5.1.6-1.mga5.i586.rpm vboxadditions-kernel-desktop586-latest-5.1.6-1.mga5.i586.rpm vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.i586.rpm vboxadditions-kernel-server-latest-5.1.6-1.mga5.i586.rpm virtualbox-5.1.6-1.mga5.i586.rpm virtualbox-devel-5.1.6-1.mga5.i586.rpm virtualbox-doc-5.1.6-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.6-1.mga5.i586.rpm virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.i586.rpm virtualbox-kernel-4.4.22-desktop586-1.mga5-5.1.6-1.mga5.i586.rpm virtualbox-kernel-4.4.22-server-1.mga5-5.1.6-1.mga5.i586.rpm virtualbox-kernel-desktop586-latest-5.1.6-1.mga5.i586.rpm virtualbox-kernel-desktop-latest-5.1.6-1.mga5.i586.rpm virtualbox-kernel-server-latest-5.1.6-1.mga5.i586.rpm x11-driver-video-vboxvideo-5.1.6-1.mga5.i586.rpm x86_64: dkms-vboxadditions-5.1.6-1.mga5.noarch.rpm dkms-virtualbox-5.1.6-1.mga5.noarch.rpm python-virtualbox-5.1.6-1.mga5.x86_64.rpm vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64.rpm vboxadditions-kernel-4.4.22-server-1.mga5-5.1.6-1.mga5.x86_64.rpm vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.x86_64.rpm vboxadditions-kernel-server-latest-5.1.6-1.mga5.x86_64.rpm virtualbox-5.1.6-1.mga5.x86_64.rpm virtualbox-devel-5.1.6-1.mga5.x86_64.rpm virtualbox-doc-5.1.6-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.6-1.mga5.x86_64.rpm virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64.rpm virtualbox-kernel-4.4.22-server-1.mga5-5.1.6-1.mga5.x86_64.rpm virtualbox-kernel-desktop-latest-5.1.6-1.mga5.x86_64.rpm virtualbox-kernel-server-latest-5.1.6-1.mga5.x86_64.rpm x11-driver-video-vboxvideo-5.1.6-1.mga5.x86_64.rpm Depends on:
(none) =>
19397 64-bit virtualbox 5.1.6 working OK on AMD/nvidia system. Guests checked are 32-bit and 64-bit Mageia 5, and 32-bit Windows XP. Sorry, updating here results from the loss of the nvidea driver and backing down to nouveau. Are we allowing the kernel update that's in the updates_testing repo at the same time as we update Vbox? I did that and maybe that's what's happening here. I'm gonna have to push away from the computer and try another day. FWIW: It takes me the better part of 2 -> 3 hours to test Vbox as I start with a blank drive, install M5, bring it to the present state with Vbox then install the new Vbox. That along with building four Vbox clients and updating them. I had already updated the kernel when I saw the notice about virtualbox, so my updates were done in two sessions. After the updates my machine is still using the nvidia 340 driver as it was before. I haven't been doing all the stuff Bill is doing. I've just been updating an existing system and guests, the way I would expect the usual user to do. (In reply to Thomas Andrews from comment #19) > I had already updated the kernel when I saw the notice about virtualbox, so > my updates were done in two sessions. After the updates my machine is still > using the nvidia 340 driver as it was before. > > I haven't been doing all the stuff Bill is doing. I've just been updating an > existing system and guests, the way I would expect the usual user to do. Ya, we'll go over this tomorrow at the QA meeting. I've been privileged to work on these Vbox/Kernel/Nvidia updates for a very long time with tmb. Last time around we had to determine what to do first and that was the kernel first then after that had settled the Vbox update. Doing them both at the same time as I did results in losing the Nvidia driver. Not the first time that's happened. So we gotta do one or the other first. We'll talk about it tomorrow. Also, IMO, we've been bitten a number of times on Vbox updates so it can be a long arduous process to test. IMO ya gotta start from ground zero to get it right. William, I don't know where you're getting this from. Typically we have done both together, as the kmod packages have to be built against a particular kernel and virtualbox version. It also greatly simplifies the update process for the users. There's no reason a virtualbox update would have an impact on the nvidia package. I fear that you're doing something strange with how you're installing the updates if you're having this issue. (In reply to David Walser from comment #21) > There's no reason a virtualbox update would > have an impact on the nvidia package. I fear that you're doing something > strange with how you're installing the updates if you're having this issue. Update failed doing both at the same time. Here are the packages that were updated: - dkms-virtualbox-5.1.6-1.mga5.noarch - kernel-desktop-4.4.22-1.mga5-1-1.mga5.x86_64 - kernel-desktop-latest-4.4.22-1.mga5.x86_64 - kernel-firmware-nonfree-20160914-1.mga5.nonfree.noarch - meta-task-5-28.1.mga5.noarch - radeon-firmware-20160920-1.mga5.nonfree.noarch - vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64 - vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.x86_64 - virtualbox-5.1.6-1.mga5.x86_64 - virtualbox-guest-additions-5.1.6-1.mga5.x86_64 - virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64 - virtualbox-kernel-desktop-latest-5.1.6-1.mga5.x86_64 - x11-driver-video-vboxvideo-5.1.6-1.mga5.x86_64 What's missing, what should not have been included? If it's your host machine, it should have only been these: - dkms-virtualbox-5.1.6-1.mga5.noarch - kernel-desktop-4.4.22-1.mga5-1-1.mga5.x86_64 - kernel-desktop-latest-4.4.22-1.mga5.x86_64 - kernel-firmware-nonfree-20160914-1.mga5.nonfree.noarch - meta-task-5-28.1.mga5.noarch - radeon-firmware-20160920-1.mga5.nonfree.noarch - virtualbox-5.1.6-1.mga5.x86_64 - virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64 - virtualbox-kernel-desktop-latest-5.1.6-1.mga5.x86_64 If it's your VM, it should have only been these: - kernel-desktop-4.4.22-1.mga5-1-1.mga5.x86_64 - kernel-desktop-latest-4.4.22-1.mga5.x86_64 - meta-task-5-28.1.mga5.noarch - vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64 - vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.x86_64 - virtualbox-guest-additions-5.1.6-1.mga5.x86_64 - x11-driver-video-vboxvideo-5.1.6-1.mga5.x86_64 FWIW, after updating both systems to vbox 5.1.6, I was able to successfully export Windows XP and Mageia 32-bit and 64-bit guests from my AMD/nvidia based 64-bit desktop system and import them into my Intel based 64-bit HP laptop system. There were a few missteps along the way, all my fault due to inexperience. Intel virtualization had to be enabled in the BIOS, one shared folder not present on the new system had to be created, and the 5.1.6 extension pack had to be installed before trying to start guests on the new system. Once those hurdles were cleared, the guests worked as expected. Everything looks good on both systems. Will do a test on a 32-bit system tomorrow. In all the times I've tested the Vbox/Kernel/nvidia combo the number of times I've had problems with Vbox you can count on one hand. It always seems to be the Nvidia driver that goes away and defaults back to nouveau. Which actually works just fine. Many times I've thought to just give it up and run Vbox with nouveau to heck with Nvidia.
Chris B
2016-09-29 09:44:27 CEST
CC:
shybluenight =>
(none) On real hardware, M5, KDE, 64-bit
Package(s) under test:
kernel-desktop-latest kernel-firmware-nonfree
nvidia-current-kernel-desktop-latest radeon-firmware
meta-task
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.16-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi kernel-firmware-nonfree
Package kernel-firmware-nonfree-20160712-1.mga5.nonfree.noarch is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# urpmi meta-task
Package meta-task-5-28.mga5.noarch is already installed
[root@localhost wilcal]# urpmi radeon-firmware
Package radeon-firmware-20160712-1.mga5.nonfree.noarch is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Boots to a working desktop, nvidia driver
install from updates_testing:
kernel-desktop-latest kernel-firmware-nonfree
nvidia-current-kernel-desktop-latest radeon-firmware
meta-task
The following 6 packages are going to be installed:
- kernel-desktop-4.4.22-1.mga5-1-1.mga5.x86_64
- kernel-desktop-latest-4.4.22-1.mga5.x86_64
- kernel-firmware-nonfree-20160914-1.mga5.nonfree.noarch
- meta-task-5-28.1.mga5.noarch
- meta-task-5-28.1.mga5.noarch
- radeon-firmware-20160920-1.mga5.nonfree.noarch
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.22-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi kernel-firmware-nonfree
Package kernel-firmware-nonfree-20160914-1.mga5.nonfree.noarch is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# urpmi meta-task
Package meta-task-5-28.1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi radeon-firmware
Package radeon-firmware-20160920-1.mga5.nonfree.noarch is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nouveau
Kernel modules: nvidiafb, nouveau
Boots to a working desktop but lost the proprietary nvidia driver.
And had nothing to do with Vbox as it's not installed on this system under test.
(In reply to William Kenney from comment #26) > > Boots to a working desktop but lost the proprietary nvidia driver. > And had nothing to do with Vbox as it's not installed on this system under > test. Thats because you havent updated your testing procedure since we stopped providing prebuilt proprietary modules. You did: <quote> install from updates_testing: kernel-desktop-latest kernel-firmware-nonfree nvidia-current-kernel-desktop-latest radeon-firmware meta-task </quote> What you should do: You must ensure matching kernel devel package gets installed... So installing packages from testing would be (at minimal): urpmi --search-media Testing kernel-desktop-latest kernel-desktop-devel-latest Trying to urpmi "nvidia-current-kernel-desktop-latest" wont change anything as it only ensures to pull in dkms-nvidia-current, and you already have both installed. But the dkms-nvidia* _needs_ the matching kernel devel package, or drakx will switch your system back to nouveau. CC:
(none) =>
tmb x86_64 host, i586 virtualbox Installed the kernel and vbox updates on the host then upgraded the Oracle extension pack. Rebooted to the new kernel OK, nvidia 352.79. Installed the updates in virtualbox and rebooted. Both systems look good at present. CC:
(none) =>
tarazed25 I'm never convinced these things will go through quietly. :-(( Keeping my fingers crossed. :p
Installed the virtualbox updates on 64bit vbox on x86_64 host. xorg.conf says
Driver "vboxvideo"
$ uname -r
4.4.22-desktop-1.mga5
I guess I should try installing the kernel-devel package.
Well, I jumped through all the hoops then re-read the comments and realized that wilcal is talking about losing the real hardware nvidia driver. Arrgh! (In reply to Len Lawrence from comment #31) > Well, I jumped through all the hoops then re-read the comments and realized > that wilcal is talking about losing the real hardware nvidia driver. Arrgh! Yup, all my testing is on real hardware. I've give it a shot again on Sat. Testing using Thomas's suggesting. Gotta get this right. (In reply to William Kenney from comment #22) > (In reply to David Walser from comment #21) > > > There's no reason a virtualbox update would > > have an impact on the nvidia package. I fear that you're doing something > > strange with how you're installing the updates if you're having this issue. > > Update failed doing both at the same time. Here are the packages > that were updated: > > - dkms-virtualbox-5.1.6-1.mga5.noarch > - kernel-desktop-4.4.22-1.mga5-1-1.mga5.x86_64 > - kernel-desktop-latest-4.4.22-1.mga5.x86_64 > - kernel-firmware-nonfree-20160914-1.mga5.nonfree.noarch > - meta-task-5-28.1.mga5.noarch > - radeon-firmware-20160920-1.mga5.nonfree.noarch > - vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64 > - vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.x86_64 > - virtualbox-5.1.6-1.mga5.x86_64 > - virtualbox-guest-additions-5.1.6-1.mga5.x86_64 > - virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64 > - virtualbox-kernel-desktop-latest-5.1.6-1.mga5.x86_64 > - x11-driver-video-vboxvideo-5.1.6-1.mga5.x86_64 > > What's missing, what should not have been included? Bill, weren't the kernel development packages flagged as dependencies of dkms-virtualbox? They should be, if you ask me, even if the kernel had already been updated somehow without updating those development packages. You shouldn't be able to update any dkms package without updating those, too. I'm going to take some serious time tomorrow going through the kernel update process and see what's needed and when. Now that I'm convinced that it's not a Vbox issue it should be easier to deal with. Audio regressions (stuttering in ogg123 and Amarok) not fixed in 5.1.6, but it works fine otherwise. Maybe we should just let this one go? (In reply to David Walser from comment #35) > Audio regressions (stuttering in ogg123 and Amarok) not fixed in 5.1.6, but > it works fine otherwise. Maybe we should just let this one go? IMO ogg is not a big player right now. (In reply to William Kenney from comment #36) > (In reply to David Walser from comment #35) > > > Audio regressions (stuttering in ogg123 and Amarok) not fixed in 5.1.6, but > > it works fine otherwise. Maybe we should just let this one go? > > IMO ogg is not a big player right now. That's incorrect and irrelevant. I noticed that the logon/logoff audio wasn't playing properly in my Mageia 5 32-bit guest. I just tried playing various videos (mp4, avi) with VLC and Dragon Player, and both stuttered badly, as well. No audio problems at all in my Windows XP guest. David, by "Let it go," did you mean let it go out as an update, or let it go from testing and wait for vbox 5.1.8? Curious. In the 32-bit Mageia guest, switching from the Pulseaudio driver to ALSA seems to eliminate the choppy audio. Audio isn't as choppy in my 64-bit Mageia 5 guest, even with the Pulseaudio driver, but videos are choppy and pixelated in VLC. Dragon Player is better. The host for these observations uses an Athlon X2 7750 processor, 8GB DDR2 RAM, nvidia 340 video driver. Will try different hardware tomorrow. I don't use PulseAudio on my host so the driver in the VM settings has always been set to ALSA for me. The problem appears to be hardware dependent for me. Using the same guest (exported/imported) on my Probook 6550b laptop (Intel Core i3 M350 processor, 4GB DDR3 RAM, Intel graphics), audio and video are much better in all cases. But it is also dependent on the encoding of the video. A video taken with my Android tablet two years ago shows video problems it doesn't show on real hardware, while a downloaded video shows no problems, other than a lower resolution than you would see on the host. The problems are not as bad as with my desktop host, but still there. Dragon Player looks better than the VLC from the tainted repositories. (In reply to Thomas Andrews from comment #41) > The problem appears to be hardware dependent for me. Thomas could you attach, or provide a link, to a file(s) that exhibit this issue. Thanks. I have created a test page that contains videos encoded with the most commonly encountered video formats, including ogg. This to use in the field as I encounter different and varied platforms. This to find out if and/or how well any platform handles these formats. The results are quite surprising. Feel free to use this page at: http://23.117.228.191:8080/~wilcal/video_tests/html5/html5.html Package(s) under test:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
default install of:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
[root@localhost wilcal]# uname -a
Linux localhost 4.4.16-desktop-1.mga5 #1 SMP Tue Jul 26 09:23:40 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.16-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.2-2.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.2-1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.2-2.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Created clients:
M5 i586 Gnome Live-CD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
install kernel under test from updates_testing using:
urpmi --search-media Testing kernel-desktop-latest kernel-desktop-devel-latest
[root@localhost wilcal]# uname -a
Linux localhost 4.4.22-desktop-1.mga5 #1 SMP Sat Sep 24 20:22:09 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.22-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi kernel-desktop-devel-latest
Package kernel-desktop-devel-latest-4.4.22-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-nvidia-current
Package x11-driver-video-nvidia-current-352.79-3.mga5.nonfree.x86_64 is already installe
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
System boots to a working desktop. Common apps work. Screen dimensions are correct.
Virtualbox still runs just fine
Install Virtualbox update:
The following 9 packages are going to be installed:
- dkms-virtualbox-5.1.6-1.mga5.noarch
- meta-task-5-28.1.mga5.noarch
- vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64
- vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.x86_64
- virtualbox-5.1.6-1.mga5.x86_64
- virtualbox-guest-additions-5.1.6-1.mga5.x86_64
- virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64
- virtualbox-kernel-desktop-latest-5.1.6-1.mga5.x86_64
- x11-driver-video-vboxvideo-5.1.6-1.mga5.x86_64
[root@localhost wilcal]# uname -a
Linux localhost 4.4.22-desktop-1.mga5 #1 SMP Sat Sep 24 20:22:09 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.22-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.6-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.6-1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.6-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.6-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.6-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Previously created client:
M5 i586 Gnome Live-CD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
Created clients:
M5 x86_64 Gnome Live-DVD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
M5 x86_64 KDE Live-DVD installed runs and updates as a Vbox client
Common apps work. Screen dimensions are correct.
I still have to use the "nomodeset" option on the kernel
I no longer need the: "ethtool -K enp0s3 gro off in" /etc/rc.d/rc.local
(In reply to William Kenney from comment #42) > (In reply to Thomas Andrews from comment #41) > > > The problem appears to be hardware dependent for me. > > Thomas could you attach, or provide a link, to a file(s) that > exhibit this issue. Thanks. I have two brief video files that show the behavior, but they are each around 5MB in length, so are too large to attach here. I will attach them to an email sent to you. To clarify what I am seeing and hearing: I hear an audio "stutter" in all Mageia 5 guests, on any hardware. It seems worse when I am using Pulseaudio as my Vbox audio driver on my AMD box than if I use ALSA. It is easy to pick out in the two videos, as the only sound in them is the constant drone of a gasoline-powered fan running. When played on real hardware, there are no skips in the sound. Video looks OK in VLC and Dragon Player in 32-bit Mageia 5 guests on all hardware, and in the 64-bit Mageia guest on my Intel-based host laptop. On my AMD/nvidia (340 driver)-based host desktop, video in the 64-bit Mageia 5 guest freezes at several points and pixelates at a couple of others, in both VLC and Dragon Player. Video is much better in my 64-bit guest if I tell it to use both cores of my CPU. KSysGuard was reporting 100% usage of one core when the guest only had one to use, and 70-90% of both in full screen mode with both active, but closer to 50% if not in full screen mode. So it looks like this was all a case of inadequate hardware for the job at hand, not a bug in virtualbox. The audio dropouts still exist in all Mageia guests, and both hosts, no matter how I change the settings. They can be reduced in number, but never go away completely. This is clearly an upstream bug, one we are not likely to fix. The question then becomes, is this bug serious enough for us to ignore any security fixes that are in this version? For my part, I say no. We've skipped one maintenance release already, it would not be a good idea to skip another. If this is the only problem we should push the update and hope the audio gets fixed upstream in the next one. (In reply to Thomas Andrews from comment #46) > closer to 50% if not in full screen mode. So it looks like this was all a > case of inadequate hardware for the job at hand, not a bug in virtualbox. Just curious. In the vb guest, try adding the kernel option divider=10, as per bug 44. Does that improve the situation? (In reply to Dave Hodgins from comment #47) > (In reply to Thomas Andrews from comment #46) > > closer to 50% if not in full screen mode. So it looks like this was all a > > case of inadequate hardware for the job at hand, not a bug in virtualbox. > > Just curious. In the vb guest, try adding the kernel option divider=10, > as per bug 44. Does that improve the situation? Quite the opposite. The first time I tried, as soon as I ran VLC, the guest crashed. The second time it ran, but with no difference. (In reply to Thomas Andrews from comment #3) > Failure in my i586 install with a Mageia 5 guest. > > I had not booted into this install for several weeks, so there were over 100 > updates waiting, including the update to the latest server kernel and vbox > 5.1.2. I completed those updates first and then rebooted, but did not check > operation of vbox 5.1.2, or get the extension pack. (I should have.) I then > updated vbox to 5.1.4, and rebooted once more. The 5.1.4 extension pack > downloaded and installed normally. But when I went to start my Mageia 5 > guest to update that, it appeared to begin normally, but a progress bar box > with the bar set at 0% opened, with a busy cursor inside the vbox guest. > After around a minute (I did not time it) a second progress bar box opened > with progress at 20%, with the busy cursor still there. I waited another 4-5 > minutes with no change, then attempted to close the guest with the "X" in > the corner, with no response. No response to an attempt to close the > progress bar box, either. The vbox gui could be put on top, but was also > completely unresponsive. I could use the host normally to reboot, which I > did, and vbox showed the Mageia guest had been "aborted." I attempted to run > the guest again with the same result as above. > > Hardware: ASRock M2+ motherboard, Athlon X2 7750 processor, nvidia Geforce > 9800GT video card, 8GB RAM. Mageia 5 guest is configured to use the maximum > allowed for a 32-bit guest, as a sort of a test of the 32-bit server > kernel's ability to use more than 4GB. > Finally tried updating to vbox 5.1.6 on this i586 install. I had previously restored vbox 5.1.2, which worked well. After updating the real hardware i586 Mageia install, I updated vbox. I then rebooted, and downloaded and installed the extension pack. All had gone smoothly. But, when I attempted to start the Mageia 5 32-bit guest, I saw the same behavior as I saw with 5.1.4, the same behavior described above. This makes it unusable as it is. I respectfully disagree with Comment #9, as I have found vbox to be quite usable in i586 installs, using version 5.1.2 and earlier. Perhaps the problem is the amount of memory I gave to the guest when I created it, the maximum allowed for a 32-bit Mageia guest. But it works with 5.1.2, it should work with 5.1.6.
Nicolas Lécureuil
2016-10-12 11:07:45 CEST
CC:
(none) =>
mageia I am wondering if the problem I'm seeing with the 32-bit version has something to do with the size of my 32-bit Mageia guest. Could it be the 32-bit server kernel is not addressing the extra RAM as it should? I will try tonight to either import a smaller 32-bit guest from my 64-bit system, or maybe my 32-bit Windows XP guest, or perhaps create a new Mageia guest with 5.1.6. Or maybe all three. (In reply to Thomas Andrews from comment #50) > I am wondering if the problem I'm seeing with the 32-bit version has > something to do with the size of my 32-bit Mageia guest. Could it be the > 32-bit server kernel is not addressing the extra RAM as it should? > No, wait... It can't be that, as the guest works fine with the 4.4.22 kernel and vbox 5.1.2. Will still try other things tonight, though. It is my opinion that the kernel upgrade should not be held back because of the problem I'm seeing with 32-bit vbox. We need to move that kernel along if we are to get Mageia 5.1 out the door. Update supporting three virtual machines on the main test machine. VirtualBox updated about 13 days ago - guests running fine. Guests: x86_64 mga5.1 kernel 4.4.22-desktop-1 i586 mga5.1 kernel 4.4.22-desktop-1 i586 Cauldron kernel 4.7.0-desktop-2 (In reply to Thomas Andrews from comment #50) > I will try tonight to either import a smaller 32-bit guest from my 64-bit > system, or maybe my 32-bit Windows XP guest, or perhaps create a new Mageia > guest with 5.1.6. Or maybe all three. No joy trying to create a new Mageia 5 32-bit guest on my 32-bit host. Everything seemed to be fine until it got to the point where it was to boot the iso, and the behavior described in Comment 3 was repeated. Server: x86_64 kernel 4.4.22-desktop-1 fully updated Mageia 5 Guests: x86_64 kernel 4.4.16-desktop-1 Mageia 5.1 Live GNOME DVD (UEFI boot) i586 kernel 4.4.16-desktop-1 Mageia 5.1 Live GNOME DVD (legacy boot) No obvious issues. No issues found for x86_64 host and various guests. Host: # uname -a Linux marte 4.4.22-desktop-1.mga5 #1 SMP Sat Sep 24 20:22:09 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux # rpm -qa | grep -Ei 'kernel|virtualbox|vbox' | sort dkms-virtualbox-5.1.6-1.mga5 kernel-desktop-4.4.22-1.mga5-1-1.mga5 kernel-desktop-devel-4.4.22-1.mga5-1-1.mga5 kernel-desktop-devel-latest-4.4.22-1.mga5 kernel-desktop-latest-4.4.22-1.mga5 kernel-firmware-20160409-1.mga5 kernel-firmware-nonfree-20160914-1.mga5.nonfree kernel-userspace-headers-4.4.22-1.mga5 nvidia340-kernel-desktop-latest-340.96-6.mga5.nonfree virtualbox-5.1.6-1.mga5 virtualbox-doc-5.1.6-1.mga5 Guests: Mageia 5 (up-to-date) Mageia 6 (up-to-date) Ubuntu 16.04 (up-to-date) Fedora 24 (up-to-date) Windows XP Windows 7 Windows 10 CC:
(none) =>
mageia what about this bug on 32bit ? I need it validated to push the kernel update ( for mga5.1 ) Agreed at the IRC meeting that the best way forward for is to revert to 5.1.2 'feedback' noted as suggested by tmb. CC:
(none) =>
lewyssmith (In reply to Thomas Andrews from comment #49) > Perhaps the problem is the amount of memory I gave to the guest when I > created it, the maximum allowed for a 32-bit Mageia guest. But it works with > 5.1.2, it should work with 5.1.6. Could you give more details of your setup? I am testing 5.1.16 on real 32 bit hardware ( RAM 1GB - cpu without virt instructions Pentium M740). I cannot reproduce any bug for now with a 256MB VM of Mandriva One 2007... CC:
(none) =>
lists.jjorge (In reply to José Jorge from comment #58) > > Could you give more details of your setup? I am testing 5.1.16 on real 32 > bit hardware ( RAM 1GB - cpu without virt instructions Pentium M740). > I have a 32-bit Mageia 5 install on 64-bit hardware, which I created a little less than a year ago mostly for 32-bit QA testing purposes. I no longer have any strictly 32-bit hardware that is operational. The hardware under discussion uses an ASRock A790GXH/128M motherboard, with an Athlon X2 7750 processor and 8GB of RAM, with an nVidia GeForce 9800GT video card. When the 32-bit Mageia 5 install was created, the installer used the server kernel. When I created my original Mageia 5 32-bit guest, which BTW uses the desktop kernel, again for testing purposes, I told it to use up to the maximum amount of RAM, slightly less than 4GB. At the time I thought this would test the ability of the server kernel to use more than 4GB of RAM, but after monitoring memory usage while running VirtualBox I see that it doesn't actually use any more RAM than it needs. Everything seemed to work OK through vbox 5.1.2. The update to 5.1.4 wouldn't boot the guest, showing a progress bar that gets "stuck" on 20%, and necessitating a reboot of the whole system to abort it. So, I then uninstalled vbox 5.1.4 and installed 5.1.2, which again worked OK. When I attempted to test 5.1.6, which I did some time AFTER I had already updated to the 4.4.22 kernel, vbox did the same thing it did with 5.1.4. Thinking it might be the guest, I tried creating another Mageia 5 32-bit guest, this one with only 2GB of RAM allocated. It would not even boot the iso, showing the same 20% progress bar. Each time I update vbox I use the same procedure. After updating the packages, I do a reboot to make sure that the new kernel module is installed. I then run vbox and get the new extension pack using the "Check for Updates" option of the File menu. Then I attempt to run a guest, and if successful I get the guest additions etc. updates for it. This has always worked with the 64-bit vbox, and with the 32-bit vbox through 5.1.2. That is as far as I've gone. Vbox 5.1.6 is still installed. It can stay that way, since I don't use that install for anything other than testing. For actual work, I use vbox on my 64-bit Mageia 5 install on the same hardware, and that's working fine for what I use it to do. Update to my last comment: In order to provide a good test for server kernel 4.4.22 and Virtualbox 5.1.2, I did a complete re-install of 32-bit Mageia 5 on this system, then installed vbox 5.1.2 before trying the kernel update. Before updating the kernel, I tried vbox with the old Mageia 5 guest, and I saw the same symptoms as I saw with vbox 5.1.6. However, it WOULD boot the Mageia 5 iso to create another guest. After some investigation, I discovered that PAE had been enabled for the guest, one of the things I tried when I was trying to get 5.1.6 working. I unchecked that box, and then the guest worked with vbox 5.1.2. This leads me to wonder if there was some change regarding PAE in 32-bit vbox 5.1.4, where I first saw these symptoms, that was carried on to 5.1.6. I'll leave it to others more knowledgeable than me to answer that. So, I couldn't resist... Virtualbox 5.1.8 has landed, maybe it works better: SRPMS: kmod-vboxadditions-5.1.8-1.mga5.src.rpm kmod-virtualbox-5.1.8-1.mga5.src.rpm virtualbox-5.1.8-1.mga5.src.rpm i586: dkms-vboxadditions-5.1.8-1.mga5.noarch.rpm dkms-virtualbox-5.1.8-1.mga5.noarch.rpm python-virtualbox-5.1.8-1.mga5.i586.rpm vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.8-1.mga5.i586.rpm vboxadditions-kernel-4.4.22-desktop586-1.mga5-5.1.8-1.mga5.i586.rpm vboxadditions-kernel-4.4.22-server-1.mga5-5.1.8-1.mga5.i586.rpm vboxadditions-kernel-desktop586-latest-5.1.8-1.mga5.i586.rpm vboxadditions-kernel-desktop-latest-5.1.8-1.mga5.i586.rpm vboxadditions-kernel-server-latest-5.1.8-1.mga5.i586.rpm virtualbox-5.1.8-1.mga5.i586.rpm virtualbox-devel-5.1.8-1.mga5.i586.rpm virtualbox-doc-5.1.8-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.8-1.mga5.i586.rpm virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.8-1.mga5.i586.rpm virtualbox-kernel-4.4.22-desktop586-1.mga5-5.1.8-1.mga5.i586.rpm virtualbox-kernel-4.4.22-server-1.mga5-5.1.8-1.mga5.i586.rpm virtualbox-kernel-desktop586-latest-5.1.8-1.mga5.i586.rpm virtualbox-kernel-desktop-latest-5.1.8-1.mga5.i586.rpm virtualbox-kernel-server-latest-5.1.8-1.mga5.i586.rpm x11-driver-video-vboxvideo-5.1.8-1.mga5.i586.rpm x86_64: dkms-vboxadditions-5.1.8-1.mga5.noarch.rpm dkms-virtualbox-5.1.8-1.mga5.noarch.rpm python-virtualbox-5.1.8-1.mga5.x86_64.rpm vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.8-1.mga5.x86_64.rpm vboxadditions-kernel-4.4.22-server-1.mga5-5.1.8-1.mga5.x86_64.rpm vboxadditions-kernel-desktop-latest-5.1.8-1.mga5.x86_64.rpm vboxadditions-kernel-server-latest-5.1.8-1.mga5.x86_64.rpm virtualbox-5.1.8-1.mga5.x86_64.rpm virtualbox-devel-5.1.8-1.mga5.x86_64.rpm virtualbox-doc-5.1.8-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.8-1.mga5.x86_64.rpm virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.8-1.mga5.x86_64.rpm virtualbox-kernel-4.4.22-server-1.mga5-5.1.8-1.mga5.x86_64.rpm virtualbox-kernel-desktop-latest-5.1.8-1.mga5.x86_64.rpm virtualbox-kernel-server-latest-5.1.8-1.mga5.x86_64.rpm x11-driver-video-vboxvideo-5.1.8-1.mga5.x86_64.rpm Summary:
VirtualBox 5.1.6 =>
VirtualBox 5.1.8 I removed this one temporarily so it wont block the critical kernel update coming... Whiteboard:
(none) =>
feedback Could you put this back (just what's built from the virtualbox SRPM, not the kmods), so those of us using dkms-virtualbox and dkms-vboxadditions can still test 5.1.8? I'd like to test Cauldron again and want to see if 5.1.8 finally fixes the audio issues (which the ChangeLog suggests it might). Switching this to security as it closes serveral CVEs, but it will still wait until after kernel-4.4.26-1.mga5 is validated and pushed: http://www.oracle.com/technetwork/security-advisory/cpuoct2016-2881722.html#AppendixOVIR Component:
RPM Packages =>
Security Kernel got pushed, so here comes virtualbox packages again: SRPMS: kmod-vboxadditions-5.1.8-2.mga5.src.rpm kmod-virtualbox-5.1.8-2.mga5.src.rpm virtualbox-5.1.8-1.mga5.src.rpm i586: dkms-vboxadditions-5.1.8-1.mga5.noarch.rpm dkms-virtualbox-5.1.8-1.mga5.noarch.rpm python-virtualbox-5.1.8-1.mga5.i586.rpm vboxadditions-kernel-4.4.26-desktop-1.mga5-5.1.8-2.mga5.i586.rpm vboxadditions-kernel-4.4.26-desktop586-1.mga5-5.1.8-2.mga5.i586.rpm vboxadditions-kernel-4.4.26-server-1.mga5-5.1.8-2.mga5.i586.rpm vboxadditions-kernel-desktop586-latest-5.1.8-2.mga5.i586.rpm vboxadditions-kernel-desktop-latest-5.1.8-2.mga5.i586.rpm vboxadditions-kernel-server-latest-5.1.8-2.mga5.i586.rpm virtualbox-5.1.8-1.mga5.i586.rpm virtualbox-devel-5.1.8-1.mga5.i586.rpm virtualbox-doc-5.1.8-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.8-1.mga5.i586.rpm virtualbox-kernel-4.4.26-desktop-1.mga5-5.1.8-2.mga5.i586.rpm virtualbox-kernel-4.4.26-desktop586-1.mga5-5.1.8-2.mga5.i586.rpm virtualbox-kernel-4.4.26-server-1.mga5-5.1.8-2.mga5.i586.rpm virtualbox-kernel-desktop586-latest-5.1.8-2.mga5.i586.rpm virtualbox-kernel-desktop-latest-5.1.8-2.mga5.i586.rpm virtualbox-kernel-server-latest-5.1.8-2.mga5.i586.rpm x11-driver-video-vboxvideo-5.1.8-1.mga5.i586.rpm x86_64: dkms-vboxadditions-5.1.8-1.mga5.noarch.rpm dkms-virtualbox-5.1.8-1.mga5.noarch.rpm python-virtualbox-5.1.8-1.mga5.x86_64.rpm vboxadditions-kernel-4.4.26-desktop-1.mga5-5.1.8-2.mga5.x86_64.rpm vboxadditions-kernel-4.4.26-server-1.mga5-5.1.8-2.mga5.x86_64.rpm vboxadditions-kernel-desktop-latest-5.1.8-2.mga5.x86_64.rpm vboxadditions-kernel-server-latest-5.1.8-2.mga5.x86_64.rpm virtualbox-5.1.8-1.mga5.x86_64.rpm virtualbox-devel-5.1.8-1.mga5.x86_64.rpm virtualbox-doc-5.1.8-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.8-1.mga5.x86_64.rpm virtualbox-kernel-4.4.26-desktop-1.mga5-5.1.8-2.mga5.x86_64.rpm virtualbox-kernel-4.4.26-server-1.mga5-5.1.8-2.mga5.x86_64.rpm virtualbox-kernel-desktop-latest-5.1.8-2.mga5.x86_64.rpm virtualbox-kernel-server-latest-5.1.8-2.mga5.x86_64.rpm x11-driver-video-vboxvideo-5.1.8-1.mga5.x86_64.rpm Whiteboard:
feedback =>
(none) It's now consistent. $ cat /etc/modprobe.d/vboxvideo.conf blacklist vboxvideo That's now required on both i586 and x86_64 guests. Previously, it was only needed on x86_64 guests. (In reply to Dave Hodgins from comment #66) > It's now consistent. > > $ cat /etc/modprobe.d/vboxvideo.conf > blacklist vboxvideo > > That's now required on both i586 and x86_64 guests. Previously, it was only > needed on x86_64 guests. It shouldn't be, if you use nomodeset at the kernel command line. (In reply to David Walser from comment #67) > (In reply to Dave Hodgins from comment #66) > > It's now consistent. > > > > $ cat /etc/modprobe.d/vboxvideo.conf > > blacklist vboxvideo > > > > That's now required on both i586 and x86_64 guests. Previously, it was only > > needed on x86_64 guests. > > It shouldn't be, if you use nomodeset at the kernel command line. Removed the blacklisting and added the kernel option nomodeset for both the i586 and x86_64 Mageia guests, and X is working properly again. Thanks for the suggestion. Would be nice if we could automatically ensure it's added for all Mageia guests under VirtualBox. I'll have to test a new install (will wait for 5.1 iso images), to see if it's something I mistakenly removed at some time in the past. Installed on x86_64 UEFI with nvidia GTX 770, 352.79 kernel 4.4.26 32bit guest running fine with the equivalent kernel. On Intel i3-based x86_64 host, both 64-bit and 32-bit Mageia guests start ok with 5.1.6 guest additions in place. Once they are updated to 5.1.8 guest additions, both need the nomodeset kernel option to boot correctly. A 32-bit Windows XP guest was OK with guest additions downloaded from the vendor, except that it complained about my anti-malware program might be out of date. (In reply to Dave Hodgins from comment #68) > Removed the blacklisting and added the kernel option nomodeset for both the > i586 and x86_64 Mageia guests, and X is working properly again. Thanks for > the suggestion. > > Would be nice if we could automatically ensure it's added for all Mageia > guests under VirtualBox. Nope, it's a workaround... We really need to figure out why it breaks in default install It's a workaround, but due to upstream regressions, it's a needed one. It really would help if we could have our installer automatically add nomodeset for VirtualBox VMs. I'd rather fix it properly in vbox... I have some ideas, now I just need to find the time/energy to try it It would be nice to fix if possible, and if you can do that, more power to you. In the meantime, using nomodeset doesn't seem to have any downsides. (In reply to Thomas Andrews from comment #70) > On Intel i3-based x86_64 host, both 64-bit and 32-bit Mageia guests start ok > with 5.1.6 guest additions in place. Once they are updated to 5.1.8 guest > additions, both need the nomodeset kernel option to boot correctly. A 32-bit > Windows XP guest was OK with guest additions downloaded from the vendor, > except that it complained about my anti-malware program might be out of date. Things didn't go quite as well with my Athlon X2/nvidia machine, but like David I think that may be because of something I did along the way. Suffice it to say that I gave up on the existing Mageia 5 guests, and imported copies of those on my Intel machine. They worked as they did on the Intel machine, needing nomodeset to boot, but acting fine once booted. My i586 install on the same hardware didn't go well at all. After the upgrade to kernel-server 4.4.26, Mageia i586 guests would not boot in vbox 5.1.2 unless I reduced the memory allocated to them to less than 1.6GB in one case, and 2GB in the other. This lead me to believe that the server kernel was at fault, so I replaced it with the desktop kernel. After that, they would only boot if memory allocation was 1.5GB or below. Trying the upgrade to vbox 5.1.8, neither would boot no matter what I did. I will suspend all tests of vbox on this hardware until 5.1 comes out, after which I will create new hosts and new guests and try again. Maybe by then the welt from beating my head on this wall will have diminished a bit. Package(s) under test:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
default install of:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
[root@localhost wilcal]# uname -a
Linux localhost.localdomain 4.4.26-desktop-1.mga5 #1 SMP Thu Oct 20 09:30:12 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.26-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.2-8.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.2-1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.2-8.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Created clients:
M5 i586 Gnome Live-CD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
install kernel under test from updates_testing using:
urpmi --search-media Testing kernel-desktop-latest kernel-desktop-devel-latest
[root@localhost wilcal]# urpmi --search-media Testing kernel-desktop-latest kernel-desktop-devel-latest
No package named kernel-desktop-latest
No package named kernel-desktop-devel-latest
The following 9 packages are going to be installed:
- dkms-virtualbox-5.1.8-1.mga5.noarch
- perl-URPM-5.06.3-1.mga5.x86_64
- vboxadditions-kernel-4.4.26-desktop-1.mga5-5.1.8-2.mga5.x86_64
- vboxadditions-kernel-desktop-latest-5.1.8-2.mga5.x86_64
- virtualbox-5.1.8-1.mga5.x86_64
- virtualbox-guest-additions-5.1.8-1.mga5.x86_64
- virtualbox-kernel-4.4.26-desktop-1.mga5-5.1.8-2.mga5.x86_64
- virtualbox-kernel-desktop-latest-5.1.8-2.mga5.x86_64
- x11-driver-video-vboxvideo-5.1.8-1.mga5.x86_64
System boots to a working desktop. Common apps work. Screen dimensions are correct.
Virtualbox still runs just fine
Install Virtualbox update:
The following 9 packages are going to be installed:
- dkms-virtualbox-5.1.6-1.mga5.noarch
- meta-task-5-28.1.mga5.noarch
- vboxadditions-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64
- vboxadditions-kernel-desktop-latest-5.1.6-1.mga5.x86_64
- virtualbox-5.1.6-1.mga5.x86_64
- virtualbox-guest-additions-5.1.6-1.mga5.x86_64
- virtualbox-kernel-4.4.22-desktop-1.mga5-5.1.6-1.mga5.x86_64
- virtualbox-kernel-desktop-latest-5.1.6-1.mga5.x86_64
- x11-driver-video-vboxvideo-5.1.6-1.mga5.x86_64
[root@localhost wilcal]# uname -a
Linux localhost.localdomain 4.4.26-desktop-1.mga5 #1 SMP Thu Oct 20 09:30:12 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.26-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.8-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.8-2.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.8-1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.8-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.8-2.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.8-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Previously created client:
M5 i586 Gnome Live-CD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
Created clients:
M5 x86_64 Gnome Live-DVD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
M5 x86_64 KDE Live-DVD installed runs and updates as a Vbox client
Common apps work. Screen dimensions are correct.
I still have to use the "nomodeset" option on the kernel
Looks good here.
Lewis Smith
2016-10-24 09:16:50 CEST
CC:
lewyssmith =>
(none)
David Walser
2016-10-25 20:21:44 CEST
URL:
(none) =>
http://lwn.net/Vulnerabilities/704586/ VM crashed while the 5.1.8 guest packages were installing, after rebooting, had double versions of two of them installed and had to get rid of the 5.1.6 ones from the RPM database with rpm -e. On this particular VM, I hadn't been needing nomodeset for some reason, but do need it with 5.1.8. Choppy audio issues unfortunately still haven't been fixed. Mageia 5 x86_64 host, Mageia 5 i586 guest. Whiteboard:
(none) =>
feedback Tested in 32 bits MGA5, with a Mandriva One 2007 LiveCD. Boot failed with 5.1.6, all is Ok with 5.1.8. So I'd put MGA5-32-OK ... Since this still ha issues, I moved the rpms in comment 65 to backports_testing for now so they wouldn't interfere / block the kernel security update just added: https://bugs.mageia.org/show_bug.cgi?id=19725 As soon as kernel 4.4.32 in https://bugs.mageia.org/show_bug.cgi?id=19791 is validated/pushed, I will push a new 5.1.10 based vbox package that should hopefully fix audio issues as upstream has reverted to codecs used in 5.0 series until they can properly fix up the rewrite Time to try again: SRPMS: kmod-vboxadditions-5.1.10-1.mga5.src.rpm kmod-virtualbox-5.1.10-1.mga5.src.rpm virtualbox-5.1.10-1.mga5.src.rpm i586: dkms-vboxadditions-5.1.10-1.mga5.noarch.rpm dkms-virtualbox-5.1.10-1.mga5.noarch.rpm python-virtualbox-5.1.10-1.mga5.i586.rpm vboxadditions-kernel-4.4.32-desktop-1.mga5-5.1.10-1.mga5.i586.rpm vboxadditions-kernel-4.4.32-desktop586-1.mga5-5.1.10-1.mga5.i586.rpm vboxadditions-kernel-4.4.32-server-1.mga5-5.1.10-1.mga5.i586.rpm vboxadditions-kernel-desktop586-latest-5.1.10-1.mga5.i586.rpm vboxadditions-kernel-desktop-latest-5.1.10-1.mga5.i586.rpm vboxadditions-kernel-server-latest-5.1.10-1.mga5.i586.rpm virtualbox-5.1.10-1.mga5.i586.rpm virtualbox-devel-5.1.10-1.mga5.i586.rpm virtualbox-doc-5.1.10-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.10-1.mga5.i586.rpm virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.10-1.mga5.i586.rpm virtualbox-kernel-4.4.32-desktop586-1.mga5-5.1.10-1.mga5.i586.rpm virtualbox-kernel-4.4.32-server-1.mga5-5.1.10-1.mga5.i586.rpm virtualbox-kernel-desktop586-latest-5.1.10-1.mga5.i586.rpm virtualbox-kernel-desktop-latest-5.1.10-1.mga5.i586.rpm virtualbox-kernel-server-latest-5.1.10-1.mga5.i586.rpm x11-driver-video-vboxvideo-5.1.10-1.mga5.i586.rpm x86_64: dkms-vboxadditions-5.1.10-1.mga5.noarch.rpm dkms-virtualbox-5.1.10-1.mga5.noarch.rpm python-virtualbox-5.1.10-1.mga5.x86_64.rpm vboxadditions-kernel-4.4.32-desktop-1.mga5-5.1.10-1.mga5.x86_64.rpm vboxadditions-kernel-4.4.32-server-1.mga5-5.1.10-1.mga5.x86_64.rpm vboxadditions-kernel-desktop-latest-5.1.10-1.mga5.x86_64.rpm vboxadditions-kernel-server-latest-5.1.10-1.mga5.x86_64.rpm virtualbox-5.1.10-1.mga5.x86_64.rpm virtualbox-devel-5.1.10-1.mga5.x86_64.rpm virtualbox-doc-5.1.10-1.mga5.noarch.rpm virtualbox-guest-additions-5.1.10-1.mga5.x86_64.rpm virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.10-1.mga5.x86_64.rpm virtualbox-kernel-4.4.32-server-1.mga5-5.1.10-1.mga5.x86_64.rpm virtualbox-kernel-desktop-latest-5.1.10-1.mga5.x86_64.rpm virtualbox-kernel-server-latest-5.1.10-1.mga5.x86_64.rpm x11-driver-video-vboxvideo-5.1.10-1.mga5.x86_64.rpm Summary:
VirtualBox 5.1.8 =>
VirtualBox 5.1.10 Something appears to have gone wrong during the build. The x11-driver-video-vboxvideo package is missing the /usr/lib/xorg/modules/drivers/vboxvideo_drv.so file, so X doesn't start. Whiteboard:
(none) =>
feedback nope, it was intentionally disabled in this build according to upstream developer suggestion... (I can flip it back on with a simple buildflag change) if you change /etc/X11/xorg.conf from vboxvideo to modesetting, does it work then ? did you test on 32bit or 64bit ? Ouch. So many upstream regressions in vboxvideo that they've given up? I hope all of the integration features still work with modesetting (Shlomi is usually good at catching when they don't, so hopefully he will help test this). I will have to remove nomodeset as well. We'll need to document all this in the advisory. I am testing a 64-bit host and 32-bit VM. according to the dev it's a planned direction as the modesetting ddx has improved a lot, so the boxvideo xorg driver is mostly only inteded for xorg 1.17 and older... if it works nicely with modesetting we could maybe do a s/vboxvideo/modesetting/ in a versioned trigger to make the change automatic... Well we have xorg 1.16 in Mageia 5, so maybe we should keep it enabled there. I'll test it with modesetting though. It has worked fine for me before, but we'll need Shlomi or someone to check out all of the integration features. I don't resize my VM window or do clipboard sharing or any of that kind of stuff. heh, my bad then... I've messed up versions :/ for some reason I checked we have 1.18, but thats since I upgraded my test laptop to cauldron and forgot about it... I'll respin with vboxvideo enabled Just FYI, the console isn't quite right with modesetting right now. When KMS kicks in during the boot (before X starts) the console goes small and turns mostly purple with a little bit of black and mageia smeared across the bottom. X actually then works fine, but if you switch to a virtual console it's small (looks fine though). Audio has gone back to how it was before the regressions. I had to switch from HD Audio (which doesn't work at all now, but that's how it used to be) back to AC97 and the stuttering is gone and it sounds fine. So that's progress.
Adrien D
2016-11-26 19:54:05 CET
CC:
ghbdtn =>
(none) vboxvideo ddx now re-enabled, so new rpm list: SRPMS: kmod-vboxadditions-5.1.10-1.1.mga5.src.rpm kmod-virtualbox-5.1.10-1.1.mga5.src.rpm virtualbox-5.1.10-1.1.mga5.src.rpm i586: dkms-vboxadditions-5.1.10-1.1.mga5.noarch.rpm dkms-virtualbox-5.1.10-1.1.mga5.noarch.rpm python-virtualbox-5.1.10-1.1.mga5.i586.rpm vboxadditions-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.i586.rpm vboxadditions-kernel-4.4.32-desktop586-1.mga5-5.1.10-1.1.mga5.i586.rpm vboxadditions-kernel-4.4.32-server-1.mga5-5.1.10-1.1.mga5.i586.rpm vboxadditions-kernel-desktop586-latest-5.1.10-1.1.mga5.i586.rpm vboxadditions-kernel-desktop-latest-5.1.10-1.1.mga5.i586.rpm vboxadditions-kernel-server-latest-5.1.10-1.1.mga5.i586.rpm virtualbox-5.1.10-1.1.mga5.i586.rpm virtualbox-devel-5.1.10-1.1.mga5.i586.rpm virtualbox-doc-5.1.10-1.1.mga5.noarch.rpm virtualbox-guest-additions-5.1.10-1.1.mga5.i586.rpm virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.i586.rpm virtualbox-kernel-4.4.32-desktop586-1.mga5-5.1.10-1.1.mga5.i586.rpm virtualbox-kernel-4.4.32-server-1.mga5-5.1.10-1.1.mga5.i586.rpm virtualbox-kernel-desktop586-latest-5.1.10-1.1.mga5.i586.rpm virtualbox-kernel-desktop-latest-5.1.10-1.1.mga5.i586.rpm virtualbox-kernel-server-latest-5.1.10-1.1.mga5.i586.rpm x11-driver-video-vboxvideo-5.1.10-1.1.mga5.i586.rpm x86_64: dkms-vboxadditions-5.1.10-1.1.mga5.noarch.rpm dkms-virtualbox-5.1.10-1.1.mga5.noarch.rpm python-virtualbox-5.1.10-1.1.mga5.x86_64.rpm vboxadditions-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.x86_64.rpm vboxadditions-kernel-4.4.32-server-1.mga5-5.1.10-1.1.mga5.x86_64.rpm vboxadditions-kernel-desktop-latest-5.1.10-1.1.mga5.x86_64.rpm vboxadditions-kernel-server-latest-5.1.10-1.1.mga5.x86_64.rpm virtualbox-5.1.10-1.1.mga5.x86_64.rpm virtualbox-devel-5.1.10-1.1.mga5.x86_64.rpm virtualbox-doc-5.1.10-1.1.mga5.noarch.rpm virtualbox-guest-additions-5.1.10-1.1.mga5.x86_64.rpm virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.x86_64.rpm virtualbox-kernel-4.4.32-server-1.mga5-5.1.10-1.1.mga5.x86_64.rpm virtualbox-kernel-desktop-latest-5.1.10-1.1.mga5.x86_64.rpm virtualbox-kernel-server-latest-5.1.10-1.1.mga5.x86_64.rpm x11-driver-video-vboxvideo-5.1.10-1.1.mga5.x86_64.rpm Source RPM:
virtualbox-5.1.10-1.mga5 =>
virtualbox-5.1.10-1.1.mga5 Looks like we have a winner. Mageia 5 x86_64 host, Mageia 5 i586 guest. It's working fine. I also have the new kernel running on both guest and host, as well as an i586 laptop. Whiteboard:
(none) =>
MGA5-64-OK HP Probook 6550b, Intel i3, RAM newly upgraded to 8GB, Intel graphics. Original vbox 5.1.2, running on 64-bit desktop kernel 4.4.32. x86_64 and i586 guests, neither of which needs nomodeset to start and run. Audio and video are choppy in the guests, unless hardware acceleration is disabled. Update to vbox 5.1.10-1.1 proceeded without incident. Extension pack downloaded and installed using vbox's "Check for Updates." x86 guest seemed to update to 5.1.10 guest additions and vboxvideo driver without incident, but would not reboot unless "nomodeset" was used. The x86_64 guest video became distorted and completely unreadable partway through the guest additions/vboxvideo update, and had to be shut down. Afterward, it would only reboot if "nomodeset" is used. I did not attempt to update to the 4.4.35 kernel that is on the repos, either in host or guests. On the plus side, audio and video "stuttering" is reduced from what it was, with none noticeable at all when hardware acceleration is disabled. MGA5-64 on HP Pavilion dv7 KDE AMD free driver I followed the procedure for virtualbox testing with the 4.4.32 kernel : - test that install/upgrade works. No problem - verify that older virtual disks/systems still work. I used an old virtual machine running CentOS, no problem. - try to install new virtual machines. I installed and updated a Debian Iso to test, everything is working with the additional guest additions. - check that when you start to create a new virtual machine, it lists Mageia and Mageia (64 bit) in os list and that the matching icons are shown. It's ok, when I create a new virtual machine, I can see Mageia and Mageia 64 bits in os list with the icons shown. CC:
(none) =>
youpburden on mga5-64 Updates installed: - virtualbox-5.1.10-1.1.mga5.x86_64 - virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.x86_64 - virtualbox-kernel-desktop-latest-5.1.10-1.1.mga5.x86_64 Packages installed cleanly Extension pack upgraded cleanly Guests: winxp - OK guest additions updated cleanly mga5-32 updates installed: - vboxadditions-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.i586 - vboxadditions-kernel-desktop-latest-5.1.10-1.1.mga5.i586 - virtualbox-guest-additions-5.1.10-1.1.mga5.i586 - x11-driver-video-vboxvideo-5.1.10-1.1.mga5.i586 Packages installed cleanly On rebooting, I had to switch from vboxvideo to modesetting to get X to start. Otherwise, OK CC:
(none) =>
jim Host system: ASRock motherboard, Athlon X2 7750 processor, 8GB RAM, Nvidia 9800GT video card. Host had been updated to virtualbox 5.1.8 for a previous test. Guests: One each 64-bit and 32-bit Mageia 5 systems, created months ago and fully-updated before beginning this test, including guest additions for vbox 5.1.8. Updates installed on host: dkms-virtualbox-5.1.10-1.1.mga5.noarch.rpm virtualbox-5.1.10-1.1.mga5.x86_64.rpm virtualbox-doc-5.1.10-1.1.mga5.noarch.rpm virtualbox-kernel-4.4.32-server-1.mga5-5.1.10-1.1.mga5.x86_64.rpm virtualbox-kernel-server-latest-5.1.10-1.1.mga5.x86_64.rpm All packages seemed to update cleanly, without apparent error. the virtualbox package seemed to take a long time to install, as if the kernel modules were being built locally. New extension pack downloaded and installed easily. On each guest, nomodeset had been necessary to boot with vbox 5.1.8 guest additions. Still necessary to boot from vbox 5.1.10. Each guest was then updated with the appropriate 5.1.10 guest additions, dkms-guest additions, vboxvideo driver, and kernel modules. Again installation of guest additions package itself took a long time, as if the kernel module was being built locally. Each guest still required nomodeset to boot. Simply changing "vboxvideo" in xorg.conf to "modesetting" resulted in a system that would not start. I had to learn the procedure for using nano in safe mode to restore it. (not exactly within my comfort zone.) After restarting each guest once more, again using nomodeset, a search in MCC revealed that the X11 modesetting video driver is NOT installed - the reason why the edit failed. I have no idea why the modesetting driver was left out, or how to install it properly in a working system, so that is where I'm leaving things at this time. The systems that were started using nomodeset seem to be working just fine. (In reply to Thomas Andrews from comment #96) > After restarting each guest once more, again using nomodeset, a search in > MCC revealed that the X11 modesetting video driver is NOT installed - the > reason why the edit failed. I have no idea why the modesetting driver was > left out, or how to install it properly in a working system, so that is > where I'm leaving things at this time. The systems that were started using > nomodeset seem to be working just fine. I should have known. MCC is the tool to use. Just because you don't see a driver at first glance doesn't mean it isn't there. One sometimes needs to look further, in a different spot on a slightly more obscure list. After using MCC to select the modesetting driver in each guest, and removing the "nomodeset" kernel option, each guest started with no problems. Package(s) under test:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
The following 12 packages are going to be installed:
- dkms-virtualbox-5.1.2-1.mga5.noarch
- lib64qt5opengl5-5.4.2-1.1.mga5.x86_64
- lib64qt5printsupport5-5.4.2-1.1.mga5.x86_64
- nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64
- vboxadditions-kernel-4.4.32-desktop-1.mga5-5.1.2-11.mga5.x86_64
- vboxadditions-kernel-desktop-latest-5.1.2-11.mga5.x86_64
- virtualbox-5.1.2-1.mga5.x86_64
- virtualbox-doc-5.1.2-1.mga5.noarch
- virtualbox-guest-additions-5.1.2-1.mga5.x86_64
- virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.2-11.mga5.x86_64
- virtualbox-kernel-desktop-latest-5.1.2-11.mga5.x86_64
- x11-driver-video-vboxvideo-5.1.2-1.mga5.x86_64
default install of:
kernel-desktop-latest
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
nvidia-current-kernel-desktop-latest
root@localhost wilcal]# uname -a
Linux localhost 4.4.32-desktop-1.mga5 #1 SMP Tue Nov 15 09:08:15 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.32-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.2-11.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.2-1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.2-11.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.2-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Created client:
M5 i586 Gnome Live-CD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
install:
urpmi --search-media Testing kernel-desktop-latest kernel-desktop-devel-latest
System reboots to a working desktop. Common apps work. Screen dimensions are correct.
Virtualbox still runs just fine
Install Virtualbox update:
install of:
virtualbox vboxadditions-kernel-desktop-latest dkms-virtualbox
virtualbox-guest-additions virtualbox-kernel-desktop-latest x11-driver-video-vboxvideo
from updates_testing
The following 8 packages are going to be installed:
- dkms-virtualbox-5.1.10-1.1.mga5.noarch
- vboxadditions-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.x86_64
- vboxadditions-kernel-desktop-latest-5.1.10-1.1.mga5.x86_64
- virtualbox-5.1.10-1.1.mga5.x86_64
- virtualbox-guest-additions-5.1.10-1.1.mga5.x86_64
- virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.x86_64
- virtualbox-kernel-desktop-latest-5.1.10-1.1.mga5.x86_64
- x11-driver-video-vboxvideo-5.1.10-1.1.mga5.x86_64
[root@localhost wilcal]# uname -a
Linux localhost 4.4.35-desktop-1.mga5 #1 SMP Sat Nov 26 12:26:53 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@localhost wilcal]# urpmi kernel-desktop-latest
Package kernel-desktop-latest-4.4.35-1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox
Package virtualbox-5.1.10-1.1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi vboxadditions-kernel-desktop-latest
Package vboxadditions-kernel-desktop-latest-5.1.10-1.1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi dkms-virtualbox
Package dkms-virtualbox-5.1.10-1.1.mga5.noarch is already installed
[root@localhost wilcal]# urpmi virtualbox-guest-additions
Package virtualbox-guest-additions-5.1.10-1.1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi virtualbox-kernel-desktop-latest
Package virtualbox-kernel-desktop-latest-5.1.10-1.1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi x11-driver-video-vboxvideo
Package x11-driver-video-vboxvideo-5.1.10-1.1.mga5.x86_64 is already installed
[root@localhost wilcal]# urpmi nvidia-current-kernel-desktop-latest
Package nvidia-current-kernel-desktop-latest-352.79-10.mga5.nonfree.x86_64 is already installed
[root@localhost wilcal]# lspci -k
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 440] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 3518
Kernel driver in use: nvidia
Kernel modules: nvidiafb, nouveau, nvidia_current
Previously created client:
M5 i586 Gnome Live-CD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
Created clients:
M5 x86_64 Gnome Live-DVD runs as a Vbox client.
Common apps work. Screen dimensions are correct.
M5 x86_64 KDE Live-DVD installs, runs and updates as a Vbox client
Common apps work. Screen dimensions are correct.
I still have to use the "nomodeset" option on the kernel
Looks good here.
With the Probook 6550b from Comment 93, using kernel 4.4.32, modesettings driver will not work. Mageia 5 guests do not start, exhibiting the same symptoms as when using the vboxvideo driver without the nomodeset option. Vboxvideo WITH the nomodeset kernel option does allow Mageia 5 guests to start. Everyone: Please note that the intended method of operation for Mageia 5 guests is to use the vboxvideo module with nomodeset. The modesetting module (without nomodeset obviously) should only be used with Mageia 6 guests. (In reply to David Walser from comment #100) > Everyone: > Please note that the intended method of operation for Mageia 5 guests is to > use the vboxvideo module with nomodeset. The modesetting module (without > nomodeset obviously) should only be used with Mageia 6 guests. Appreciate the clarification. Thanks. Will go fix it now... (In reply to James Kerr from comment #95) My mga5-32 guest is also OK if I use vboxvideo with nomodeset The Mageia 5 guests on my Athlon X2-based machine still show audio and video dropouts when using VLC. However, as before I now believe this is due to inadequate hardware, as the Intel i3-based machine can be configured so they are not present. It's nothing that I believe should delay the update. Installed these updates on an x86_64 machine with nvidia GTX970 and 4k monitor on DisplayPort. Installed KDE4 from the mag5.1 i586 classic DVD iso and rebooted to the desktop with nomodeset. The screen resized to 2560x1600 which suits me because I prefer to work with PIP setting. Can somebody explain what 'unified outputs' means in monitor settings? Web-searching seems that it has something to do with dual monitors. Perhaps, running virtualbox in tandem? GNOME Live had been running fine on this vbox. KDE4 is not displaying any problems. Finished configuring media sources via MageiaWelcome and updated the system. Firefox working. LibreOffice writer functioning. Shall report any anomalies over the next 24 hours but it looks OK. updated advisory in svn with references to: - CVE-2016-2177 - CVE-2016-2178 - CVE-2016-2179 - CVE-2016-2180 - CVE-2016-2181 - CVE-2016-2182 - CVE-2016-2183 - CVE-2016-5501 - CVE-2016-5538 - CVE-2016-5605 - CVE-2016-5608 - CVE-2016-5610 - CVE-2016-5611 - CVE-2016-5613 - CVE-2016-6302 - CVE-2016-6303 - CVE-2016-6304 - CVE-2016-6305 - CVE-2016-6306 - CVE-2016-6307 - CVE-2016-6308 - CVE-2016-6309 - CVE-2016-7052 and rpms list in comment 91 Whiteboard:
MGA5-64-OK =>
MGA5-64-OK advisory I installed this on [brian@localhost ~]$ uname -a Linux localhost 4.4.32-desktop-1.mga5 #1 SMP Tue Nov 15 09:08:15 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux GNOME IDE for host. AMD Hardware and Video. It seems to be working fine. Was able to open up and use VM's for Linux instances. Anybody try a Windows OS? CC:
(none) =>
brtians1 I did briefly fire up a Windows 7 VM at work; it loaded fine. trying this on 32 bit [brian@localhost ~]$ uname -a Linux localhost 4.4.30-server-2.mga5 #1 SMP Fri Nov 4 21:10:28 UTC 2016 i686 i686 i686 GNU/Linux I get the following error when I just pick VirtualBox 5.1.10.1.1 Sorry, the following package cannot be selected: - virtualbox-5.1.10-1.1.mga5.i586 (due to conflicts with virtualbox-kernel-4.4.32-desktop-1.mga5-5.1.10-1.1.mga5.i586) hmm - received some new updates - now it will load. 32-bit Windows XP guest worked OK for me, on a 64-bit host. [brian@localhost ~]$ uname -a Linux localhost 4.4.32-server-1.mga5 #1 SMP Tue Nov 15 11:00:38 UTC 2016 i686 i686 i686 GNU/Linux [brian@localhost ~]$ After installing things in the right order, I was able to import a VM appliance (mageia i586) into the 32-bit VirtualBox install and start it. Used the browser and navigated the Mageia site. Works for me. Anything anybody want me to do with 32-bit hosted VBox? Whiteboard:
MGA5-64-OK advisory =>
MGA5-64-OK advisory mga5-32-ok for grins I loaded x86_64 client (imported the appliance). I was able to bring up the client and hit the mageia.org site as well. I can only imagine the weird memory address mapping the VM has to do. But it works on the same machine as listed above. 32-bit works. Why do it? Well, I'm sure someone somewhere sometime wants to. Host: $ cat /etc/release Mageia release 5 (Official) for x86_64 $ uname -a Linux marte 4.4.36-desktop-1.mga5 #1 SMP Fri Dec 2 17:48:04 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux $ VirtualBox --help | head -n 1 Oracle VM VirtualBox Manager 5.1.10 Guests: Windows XP 32bit Windows 7 64bit Windows 10 64bit No issues or regressions found. Validating. Keywords:
(none) =>
validated_update An update for this issue has been pushed to the Mageia Updates repository. http://advisories.mageia.org/MGASA-2016-0408.html Status:
NEW =>
RESOLVED LWN reference for CVE-2016-6309: https://lwn.net/Vulnerabilities/708242/ |