[gentoo-user] Re: newish vbox gentoo guest crashes from startx
Harry Putnamwrites: > [ebuild R ] x11-base/xorg-server-1.18.4:0/1.18.4::gentoo USE="doc > glamor suid udev xorg -dmx -ipv6 -kdrive -libressl -minimal (-selinux) > -static-libs -systemd -tslib -unwind -wayland -xephyr -xnest -xvfb" 0 > KiB > [ebuild R] x11-drivers/xf86-input-mouse-1.9.2::gentoo 0 KiB > [ebuild R ] x11-drivers/xf86-video-virtualbox-5.1.10::gentoo USE="dri" > PYTHON_TARGETS="python2_7" 0 KiB > [ebuild R] x11-drivers/xf86-input-evdev-2.10.4::gentoo 0 KiB > [ebuild R] x11-drivers/xf86-input-keyboard-1.9.0::gentoo 0 KiB > > I'm proceeding with those rebuilds now. As posted previously there was no change in behavior following above activity still `startlxde' gets no response, just a returned prompt after a very brief pause (maybe 1.5 sec). startx still causes a full bodied crash. Further... sometimes a feeze happens when none of the logins are active. Can't tell what activity brings it about. So, I'm guessing there is some underlying problem that has yet to be uncovered. I'm mostly out of ideas. Still hoping someone will have some notions about how to further diagnose this install. I'm checking logs and watching behavior. I'll keep the install as command line only for mail for now...
Re: [gentoo-user] Re: newish vbox gentoo guest crashes from startx
Other thing to keep in mind is if you recently installed/updated virtbox is to reboot your host to make sure the host drivers have installed correctly, and make sure your kernel for the guest have the correct modules built. On Thu, Dec 8, 2016, 12:10 Harry Putnamwrote: > Harry Putnam writes: > > [...] > > >> The only thing I can think of left to check is > >> emerge @x11-module-rebuild > >> and see what it will install > > > > > > [ebuild R ] x11-base/xorg-server-1.18.4:0/1.18.4::gentoo USE="doc > > glamor suid udev xorg -dmx -ipv6 -kdrive -libressl -minimal (-selinux) > > -static-libs -systemd -tslib -unwind -wayland -xephyr -xnest -xvfb" 0 > > KiB > > [ebuild R] x11-drivers/xf86-input-mouse-1.9.2::gentoo 0 KiB > > [ebuild R ] x11-drivers/xf86-video-virtualbox-5.1.10::gentoo USE="dri" > > PYTHON_TARGETS="python2_7" 0 KiB > > [ebuild R] x11-drivers/xf86-input-evdev-2.10.4::gentoo 0 KiB > > [ebuild R] x11-drivers/xf86-input-keyboard-1.9.0::gentoo 0 KiB > > > > I'm proceeding with those rebuilds now. > > Rebuilds finished... still: > startlxde has no effect whatever > > startx causes a full on crash > > Just a point of interest and possible diagnostic data. > > The lists below show a group of x11 pkgs that are uniq on a previous > gentoo install that is successfully running lxde. That is, the listed > pkgs are on the successful host but not on the problem host. > > I did not think any of them looked like possible causes of my problem. > > The second list (of one) below shows a pkg on the problem host that is > not on the succesfull host. Again does not appear to be a possible > cause of my problem > > > uniq on successful lxde running host gv > > x11-apps/bdftopcf > x11-apps/luit > x11-apps/xcursorgen > x11-apps/xkill > x11-apps/xlsfonts > x11-apps/xmessage > x11-apps/xmodmap > x11-apps/xpr > x11-apps/xsetroot > x11-apps/xwd > x11-libs/libXScrnSaver > x11-libs/libXaw3d > x11-libs/pangox-compat > x11-misc/xbindkeys > x11-terms/xterm > x11-themes/sound-theme-freedesktop > > > Uniq on problem host g0 > > x11-apps/xrefresh > > >
[gentoo-user] Re: newish vbox gentoo guest crashes from startx
Harry Putnamwrites: [...] >> The only thing I can think of left to check is >> emerge @x11-module-rebuild >> and see what it will install > > > [ebuild R ] x11-base/xorg-server-1.18.4:0/1.18.4::gentoo USE="doc > glamor suid udev xorg -dmx -ipv6 -kdrive -libressl -minimal (-selinux) > -static-libs -systemd -tslib -unwind -wayland -xephyr -xnest -xvfb" 0 > KiB > [ebuild R] x11-drivers/xf86-input-mouse-1.9.2::gentoo 0 KiB > [ebuild R ] x11-drivers/xf86-video-virtualbox-5.1.10::gentoo USE="dri" > PYTHON_TARGETS="python2_7" 0 KiB > [ebuild R] x11-drivers/xf86-input-evdev-2.10.4::gentoo 0 KiB > [ebuild R] x11-drivers/xf86-input-keyboard-1.9.0::gentoo 0 KiB > > I'm proceeding with those rebuilds now. Rebuilds finished... still: startlxde has no effect whatever startx causes a full on crash Just a point of interest and possible diagnostic data. The lists below show a group of x11 pkgs that are uniq on a previous gentoo install that is successfully running lxde. That is, the listed pkgs are on the successful host but not on the problem host. I did not think any of them looked like possible causes of my problem. The second list (of one) below shows a pkg on the problem host that is not on the succesfull host. Again does not appear to be a possible cause of my problem uniq on successful lxde running host gv x11-apps/bdftopcf x11-apps/luit x11-apps/xcursorgen x11-apps/xkill x11-apps/xlsfonts x11-apps/xmessage x11-apps/xmodmap x11-apps/xpr x11-apps/xsetroot x11-apps/xwd x11-libs/libXScrnSaver x11-libs/libXaw3d x11-libs/pangox-compat x11-misc/xbindkeys x11-terms/xterm x11-themes/sound-theme-freedesktop Uniq on problem host g0 x11-apps/xrefresh
[gentoo-user] Re: newish vbox gentoo guest crashes from startx
Alan McKinnonwrites: > On 08/12/2016 02:26, Harry Putnam wrote: >> Alan McKinnon writes: >> [...] Theoretically the install should be ready now to run X and lxde. However, when I run `startlxde' nothing whatsoever happens. At least not at the command line or visibly. Prompt just returns lamely. However, if I run startx. The vm crashes clear to power off or sometimes just becomes inoperable with these error messages showing. (Note: nothing can be clicked moved or etc at this point) >>> >>> >>> >>> Your first problem seems to be you don't have any input or output >>> devices. Please post >>> >>> - eix xorg-server >>> - values of INPUT_DEVICES and VIDEO_CARDS from make.conf >> >>> - confirm that you built the kernel with evdev and the kernel module is >>> loaded >> >> harry > eix xorg-server >> [I] x11-base/xorg-server [...] >> harry > grep VIDEO /etc/portage/make.conf >> VIDEO_CARDS="virtualbox" >> >> I don't have anthing in make.conf about Input devices nor have I had >> on 2 other gentoo installs that successfully ran lxde. >> >> I've used the same make.conf with little variation on all three >> >> --- --- ---=--- --- --- >> >> harry > cd /lib/modules >> >> harry > grep -r evdev.ko * >> Binary file 4.8.8-gentoo/modules.dep.bin matches >> 4.8.8-gentoo/modules.order:kernel/drivers/input/evdev.ko >> 4.8.8-gentoo/modules.dep:kernel/drivers/input/evdev.ko: >> >> --- --- ---=--- --- --- >> >> harry > sudo lsmod >> Module Size Used by >> vboxguest 171071 0 >> evdev 8704 0 > > > > All that looks OK. > AFAIR xorg uses evdev as the default, so if you don't specify it > explicitly, it's what you get. Is there something I should be specify explicitly? I guess you mean in make.conf with INPUT_DEVICES= > Just for fun, do you have x11-drivers/xf86-input-evdev emerged? harry > eix x11-drivers/xf86-input-evdev [I] x11-drivers/xf86-input-evdev Available versions: 2.9.1 2.9.2 2.10.3 (~)2.10.4 Installed versions: 2.10.4(22:57:23 11/18/16) Homepage:https://www.x.org/wiki/ Description: Generic Linux input driver I'm currently trying to compare what is installed on another gentoo guest on a solaris host with what I have on this one in the x11 department, using qlop -l but it gets a bit difficult with updates showing the same base-names repeatedly on that host since its been in business quite a bit longer. Thought about using uniq and sort in someway to ascertain things but have totally forgotten how to use it, I even wrote an awk script once to uniqify lists, but of course it appears to have gone the way of the doodoo bird. AAR, it appears so far to have the same stuff on each. So there still must be something basic different. > The only thing I can think of left to check is > emerge @x11-module-rebuild > and see what it will install [ebuild R] x11-base/xorg-server-1.18.4:0/1.18.4::gentoo USE="doc glamor suid udev xorg -dmx -ipv6 -kdrive -libressl -minimal (-selinux) -static-libs -systemd -tslib -unwind -wayland -xephyr -xnest -xvfb" 0 KiB [ebuild R] x11-drivers/xf86-input-mouse-1.9.2::gentoo 0 KiB [ebuild R] x11-drivers/xf86-video-virtualbox-5.1.10::gentoo USE="dri" PYTHON_TARGETS="python2_7" 0 KiB [ebuild R] x11-drivers/xf86-input-evdev-2.10.4::gentoo 0 KiB [ebuild R] x11-drivers/xf86-input-keyboard-1.9.0::gentoo 0 KiB I'm proceeding with those rebuilds now.
Re: [gentoo-user] Re: newish vbox gentoo guest crashes from startx
On 08/12/2016 02:26, Harry Putnam wrote: > Alan McKinnonwrites: > >> On 07/12/2016 17:34, Harry Putnam wrote: >>> Setup: >>> >>>gentoo 32bit vbox guest on win 10 64bit host >>>Installed xorg-server, lxde Meta pkgs and deps >>>(along with many other pkgs ... to many to list >>>here but included at the end) >>> >>> uname -a: >>> Linux g0 4.8.8-gentoo #2 Fri Nov 18 20:16:14 EST 2016 i686 Intel(R) >>> Xeon(R) CPU X5450 @ 3.00GHz GenuineIntel GNU/Linux >>> >>> Theoretically the install should be ready now to run X and lxde. >>> However, when I run `startlxde' nothing whatsoever happens. At least >>> not at the command line or visibly. Prompt just returns lamely. >>> >>> However, if I run startx. The vm crashes clear to power off or >>> sometimes just becomes inoperable with these error messages showing. >>> (Note: nothing can be clicked moved or etc at this point) >> >> >> >> Your first problem seems to be you don't have any input or output >> devices. Please post >> >> - eix xorg-server >> - values of INPUT_DEVICES and VIDEO_CARDS from make.conf > >> - confirm that you built the kernel with evdev and the kernel module is >> loaded > > harry > eix xorg-server > [I] x11-base/xorg-server > Available versions: 1.12.4-r5(0/1.12.4) 1.12.4-r7(0/1.12.4) > 1.15.2-r2(0/1.15.2) 1.15.2-r4(0/1.15.2) 1.16.4(0/1.16.1) 1.16.4-r5(0/1.16.1) > 1.17.4(0/1.17.4) (~)1.18.0(0/1.18.0) (~)1.18.1(0/1.18.1) (~)1.18.2(0/1.18.2) > (~)1.18.3(0/1.18.3) 1.18.4(0/1.18.4) [M](~)1.19.0(0/1.19.0) **(0/) > {dmx doc glamor ipv6 kdrive libressl minimal nptl selinux static-libs +suid > systemd tslib +udev unwind wayland xephyr xnest xorg xvfb} > Installed versions: 1.18.4(21:29:32 12/06/16)(doc glamor suid udev xorg > -dmx -ipv6 -kdrive -libressl -minimal -selinux -static-libs -systemd -tslib > -unwind -wayland -xephyr -xnest -xvfb) > Homepage:https://www.x.org/wiki/ > Description: X.Org X servers > > ---- ---=--- - > > harry > grep VIDEO /etc/portage/make.conf > VIDEO_CARDS="virtualbox" > > I don't have anthing in make.conf about Input devices nor have I had > on 2 other gentoo installs that successfully ran lxde. > > I've used the same make.conf with little variation on all three > > --- --- ---=--- --- --- > > harry > cd /lib/modules > > harry > grep -r evdev.ko * > Binary file 4.8.8-gentoo/modules.dep.bin matches > 4.8.8-gentoo/modules.order:kernel/drivers/input/evdev.ko > 4.8.8-gentoo/modules.dep:kernel/drivers/input/evdev.ko: > > --- --- ---=--- --- --- > > harry > sudo lsmod > Module Size Used by > vboxguest 171071 0 > evdev 8704 0 All that looks OK. AFAIR xorg uses evdev as the default, so if you don't specify it explicitly, it's what you get. Just for fun, do you have x11-drivers/xf86-input-evdev emerged? The only thing I can think of left to check is emerge @x11-module-rebuild and see what it will install > > >> There's a decent X11 install page on the gentoo wiki, did you follow it? >> Installing on vbox is not too different from installing on metal except >> you don't use the video driver for the physical card, you use the vbox >> one (it's virtualized). > > I used the gentoo lxde wiki, which basically says that once you've got > a basic gentoo install AND the xorg-server is installed you are ready > to install lxde. It recommends using the meta package which I did. > > I assumed emerge would take care of dependants. You assume correct > >> Lastly, why a 32 bit guest on a 64 bit host? There's little benefit from >> doing that, and amd64 is far better supported in general than x86_64. >> Plus you just create more work for the host. > > Does that have bearing on my problem? Not really. I noticed and wondered. We can ignore it for the rest of the discussion > > There were two reasons: > 1) I hadn't yet found the BIOS switch to turn on Virtualization and >the newer VBox since late 4.XX I think, do not even offer 64 bit >if it is not set in bios. So the vbox version I used didn't have >that on offer. > > 2) I found the discussions about the new approach with 2 sets of >libraries and so forth to be pretty confusing... don't recall now >exactly what things confused me... so I went straight 32 bit with the >idea it would be less confusing. And for 2 gentoo installs that >seemed to be born out. > > Since I had no trouble with those, thought I would stay with the tried > and true approach. But apparently I missed something important or > some such. > > -- Alan McKinnon alan.mckin...@gmail.com
[gentoo-user] Re: newish vbox gentoo guest crashes from startx
Alan McKinnonwrites: > On 07/12/2016 17:34, Harry Putnam wrote: >> Setup: >> >>gentoo 32bit vbox guest on win 10 64bit host >>Installed xorg-server, lxde Meta pkgs and deps >>(along with many other pkgs ... to many to list >>here but included at the end) >> >> uname -a: >> Linux g0 4.8.8-gentoo #2 Fri Nov 18 20:16:14 EST 2016 i686 Intel(R) >> Xeon(R) CPU X5450 @ 3.00GHz GenuineIntel GNU/Linux >> >> Theoretically the install should be ready now to run X and lxde. >> However, when I run `startlxde' nothing whatsoever happens. At least >> not at the command line or visibly. Prompt just returns lamely. >> >> However, if I run startx. The vm crashes clear to power off or >> sometimes just becomes inoperable with these error messages showing. >> (Note: nothing can be clicked moved or etc at this point) > > > > Your first problem seems to be you don't have any input or output > devices. Please post > > - eix xorg-server > - values of INPUT_DEVICES and VIDEO_CARDS from make.conf > - confirm that you built the kernel with evdev and the kernel module is > loaded harry > eix xorg-server [I] x11-base/xorg-server Available versions: 1.12.4-r5(0/1.12.4) 1.12.4-r7(0/1.12.4) 1.15.2-r2(0/1.15.2) 1.15.2-r4(0/1.15.2) 1.16.4(0/1.16.1) 1.16.4-r5(0/1.16.1) 1.17.4(0/1.17.4) (~)1.18.0(0/1.18.0) (~)1.18.1(0/1.18.1) (~)1.18.2(0/1.18.2) (~)1.18.3(0/1.18.3) 1.18.4(0/1.18.4) [M](~)1.19.0(0/1.19.0) **(0/) {dmx doc glamor ipv6 kdrive libressl minimal nptl selinux static-libs +suid systemd tslib +udev unwind wayland xephyr xnest xorg xvfb} Installed versions: 1.18.4(21:29:32 12/06/16)(doc glamor suid udev xorg -dmx -ipv6 -kdrive -libressl -minimal -selinux -static-libs -systemd -tslib -unwind -wayland -xephyr -xnest -xvfb) Homepage:https://www.x.org/wiki/ Description: X.Org X servers ---- ---=--- - harry > grep VIDEO /etc/portage/make.conf VIDEO_CARDS="virtualbox" I don't have anthing in make.conf about Input devices nor have I had on 2 other gentoo installs that successfully ran lxde. I've used the same make.conf with little variation on all three --- --- ---=--- --- --- harry > cd /lib/modules harry > grep -r evdev.ko * Binary file 4.8.8-gentoo/modules.dep.bin matches 4.8.8-gentoo/modules.order:kernel/drivers/input/evdev.ko 4.8.8-gentoo/modules.dep:kernel/drivers/input/evdev.ko: --- --- ---=--- --- --- harry > sudo lsmod Module Size Used by vboxguest 171071 0 evdev 8704 0 > There's a decent X11 install page on the gentoo wiki, did you follow it? > Installing on vbox is not too different from installing on metal except > you don't use the video driver for the physical card, you use the vbox > one (it's virtualized). I used the gentoo lxde wiki, which basically says that once you've got a basic gentoo install AND the xorg-server is installed you are ready to install lxde. It recommends using the meta package which I did. I assumed emerge would take care of dependants. > Lastly, why a 32 bit guest on a 64 bit host? There's little benefit from > doing that, and amd64 is far better supported in general than x86_64. > Plus you just create more work for the host. Does that have bearing on my problem? There were two reasons: 1) I hadn't yet found the BIOS switch to turn on Virtualization and the newer VBox since late 4.XX I think, do not even offer 64 bit if it is not set in bios. So the vbox version I used didn't have that on offer. 2) I found the discussions about the new approach with 2 sets of libraries and so forth to be pretty confusing... don't recall now exactly what things confused me... so I went straight 32 bit with the idea it would be less confusing. And for 2 gentoo installs that seemed to be born out. Since I had no trouble with those, thought I would stay with the tried and true approach. But apparently I missed something important or some such.