Description of problem: Xorg version: 1.19.5 After the xorg updates, X refuses to start in vbox with error (Screen not found) If xorg.conf from /etc/X11 is removed, X starts OK, but with limited resolution If drakx11 is used to regenerate the config file, selects vboxvideo as preferred driver and loads v4l module but X is unable to start if X -configure is used, the X server shows error message in the tty, and sometimes (without vboxadditions installed) using that config file makes X start The 4.14.20 kernel with vboxadditions shows in journal an error ("This module taints the kernel") This may be is the answer to the vboxvideo driver not found that appears in Xorg.0.log when I try to boot with that kernel All the vboxadditions rpm writes into /var/lib/dkms-binary the kernel objects But the 5.2.6 vbox video drivers for X are installed out of /usr/lib/xorg/modules.... are installed in /usr/lib.... when X tries to search for the video driver, specially if using vboxadditions removed from kernel and the config file asks for vboxvideo... the module is not found making fail the X
I found the vbox Xorg driver error viewing the files option in MCC/Add or remove packages I will move the files to the correct folder and write other comment if this resolves the issue
Created attachment 10012 [details] Xorg driver from MGA5 with vboxvideo compiled In the mga6 core release vbox driver, the drivers folder does not exists... may be we are relying in the vboxvideo driver from vboxadditions package
(In reply to Mauricio Andrés Bustamante Viveros from comment #0) > Description of problem: > > Xorg version: 1.19.5 > After the xorg updates, X refuses to start in vbox with error (Screen not > found) x11-server-xorg has version 1.19.5 in Mageia 6 since November last year... are you sure you updated the x11-server packages right before hitting this issue?
CC: (none) => marja11, tmb
I am sure, David Hodgins had the same issue and removing the xorg.conf seudo resolved this
(In reply to Mauricio Andrés Bustamante Viveros from comment #4) > I am sure, David Hodgins had the same issue and removing the xorg.conf seudo > resolved this CC'ing him
CC: (none) => davidwhodgins
IIRC the problem I had was caused by having both dkms-vboxadditions and vboxadditions-kernel-desktop-latest installed. Fixed mine by going with just vboxadditions-kernel-desktop-latest in that install.
I reinstalled the MGA6X32 in vbox 5.1.34 I found: X is configured with the modesetting driver (by XFdrake) MGA6 default kernel (4.9.35) starts X without problems Tests: 1. Installed the 4.9.50 kernel 2. Installed the 4.9.56 kernel 3. Installed the 4.14.10 kernel 4. Installed the 4.14.13 kernel 5. Installed the 4.9.40 kernel All of above kernels except 4.9.40 and 4.14.1? that made segfault, did not start X The Xorg.0.log of every failed Xstart contains the same error "Modesetting driver for modesetting kernel drivers: kms Falling back to old probe method for modesetting Open /dev/dri/card0: No such file or directory Screen 0 deleted: No matching config section Unloadmodule: modesetting Device detected but none match de config file Fatal error: No screen found" If I use MCC to configure X or use the drakx11 command, both configures the X video driver as "vboxvideo" If I delete the X configuration file, the X starts with video driver "fbdev" If I change from modesetting to fbdev the video driver section, X starts with slowness in every kernel that failed to start. With the modesetting driver every kernel that fails to start X presents the glitch seen in the video https://drive.google.com/open?id=1UiBv-M9eSr-oUr5LS01nb_lg40n8xAtq This issue is seen in MAG5 too, but reviewing the MGA5 x11-drivers-vboxdriver packages I see that the dri, and /usr/lib/xorg/modules were available, from an specific update, these files not exists I only tested the kernels, to avoid insert failures in my testing I will make room for snapshots and will try updating the X packages, from 1.19.3 to 1.19.5
@ tmb Sorry, but I don't know what/whom to assign this bug report to.
The "module taint" is an "out of tree taint", meaning the driver is not in upstream kernel, so it has not gone through kernel validation so it can be merged upstream... Now this is finally changing as vboxvideo is now ABI stable and landed in upstream 4.15 series kernels... the vboxsf and other bits from vboxadditions should be landing in 4.16. and x11-driver-video-vboxvideo DDX has became part of upstream xorg maintenance so that is also there... Now some of the issues you see comes from mixed drivers... if the host is vbox 5.1.x it will get some trouble as the bits in mga6 is 5.2.x, something that usually not works that good... (even if it might work at some levels) Them there is the fact that current mga virtualbox 5.2.6 has known problems with atleast 3d acceleration, something that is fixed in 5.2.8 currently in testing: https://bugs.mageia.org/show_bug.cgi?id=22686 So my suggestion is to get the host up to 5.2.8, install the 5.2.8 guest bits from testing and see how that works out...
Good afternoon/night Reading the above comment from tmb, I did a little review and found that if I select/install a kernel update (but not using the kernel-lastest meta package)in MCC install or remove packages, the vboxadditions that are required for that kernel are not selected/installed Test: 1. Install the vboxadditions for 4.14.20 kernel 2. Review if vboxadditions were installed when I installed the others kernel Results: Installed the vboxadditions for 4.14.20 kernel and X start correctly but, takes more than expected to get the plasma desktop Running the 4.9.35 and 4.14.20 kernels in runlevel 3, both kernels boots fast, but the 4.9.35 is more faster than 4.14.20 to get the plasma desktop (maybe melt and spectre patches) The vboxadditions for the others kernel were not installed when I did the install from MCC https://drive.google.com/open?id=13HeyyVIh-kpi-3c65Lsz-iMg3Yr_NR12 PD: I upgraded the vbox software in the host to 5.2.8 after today tests and I got the same results from the 4.14.20 kernel https://drive.google.com/open?id=1HzlRLSYbTTsG09PwaM6H7vpeyXEz56EI The test until this moment were without updates testing and updates from mga core release I will be out until thursday (helping to others in an hospital) and will be updating the Xorg
(In reply to Mauricio Andrés Bustamante Viveros from comment #10) > > The test until this moment were without updates testing and updates from mga > core release > > I will be out until thursday (helping to others in an hospital) and will be > updating the Xorg It's always good to help others, thanks :-) How did updating go, can this bug report be closed?
Hi Marja, QA Team and others This can be closed, but some sidenotes for documentation (if applicable) 1. If the end user without using mageiaupdate, wants to select a new kernel (not the newest) and the mageia system is in virtualbox, the end user must select manually the virtualbox packages related to the selected kernel before install or reboot 2. If the end user without using mageiaupdate, wants the newest kernel, is recommended that select the kernel lastest package to avoid this issue Thanks
(In reply to Mauricio Andrés Bustamante Viveros from comment #12) > Hi Marja, QA Team and others > > This can be closed, but some sidenotes for documentation (if applicable) > > 1. If the end user without using mageiaupdate, wants to select a new kernel > (not the newest) and the mageia system is in virtualbox, the end user must > select manually the virtualbox packages related to the selected kernel > before install or reboot > > 2. If the end user without using mageiaupdate, wants the newest kernel, is > recommended that select the kernel lastest package to avoid this issue > > Thanks Do you have specific wiki or other pages in mind where this should be done? Reassigning to documentation team
Summary: Xorg fails in vbox as guest, using windows vbox version 5.1.12 and later (not 5.2 branch) after mga5 updates or mga6 updates => Improve documentation? (Was: Xorg fails in vbox as guest, using windows vbox version 5.1.12 and later (not 5.2 branch) after mga5 updates or mga6 updates)Assignee: bugsquad => doc-bugs
Hello, Done here: https://wiki.mageia.org/en/VirtualBox#On_guests Please check that this is correct, or edit it if needed.
CC: (none) => yves.brungard_mageia
(In reply to papoteur from comment #14) > Hello, > Done here: https://wiki.mageia.org/en/VirtualBox#On_guests > Please check that this is correct, or edit it if needed. @ Mauricio Can you please check it and close this report when it's correct? If you need edit rights in the wiki, please say so, we can add you to the whitelist when you give your username. Thanks :-)
Documentation as requiere Closing as resolved
Status: NEW => RESOLVEDResolution: (none) => FIXED