I have an annoying problem. I run a two-monitor setup with xrandr, one monitor in portrait and the other in landscape mode. I set xrandr in the .xinitrc file, and it works well. The only problem is that when I quit the window manager (in this case goomwwm), the screen goes to black and I'm pretty sure I never return to the console (blind typing produces no results). The computer itself keeps running; I can ssh in. Any advice or suggestions? I can't quite figure how to turn xrandr *off* once it's been set in .xinitrc. Oh, and if I boot without xrandr, I go straight back to the console when I quit X (as it should). -- Peter King peter.king@utoronto.ca Department of Philosophy 170 St. George Street #521 The University of Toronto (416)-978-4951 ofc Toronto, ON M5R 2M8 CANADA http://individual.utoronto.ca/pking/ ========================================================================= GPG keyID 0x7587EC42 (2B14 A355 46BC 2A16 D0BC 36F5 1FE6 D32A 7587 EC42) gpg --keyserver pgp.mit.edu --recv-keys 7587EC42
On 09/30/2014 08:36 PM, Peter King wrote:
I have an annoying problem. I run a two-monitor setup with xrandr, one monitor in portrait and the other in landscape mode. I set xrandr in the .xinitrc file, and it works well. The only problem is that when I quit the window manager (in this case goomwwm), the screen goes to black and I'm pretty sure I never return to the console (blind typing produces no results). The computer itself keeps running; I can ssh in. Any advice or suggestions? I can't quite figure how to turn xrandr *off* once it's been set in .xinitrc. Oh, and if I boot without xrandr, I go straight back to the console when I quit X (as it should).
Hi Peter, I'm curious to see how your setting up your screens in .xinitrc. Can you post it? Follow up question, do any other WMs have an issue exiting with your xrandr setup? For comparison, I'm attaching a script I use to setup various secondary monitors with my laptop. As this is a laptop I defined a 'default' that sets all other outputs but the LCD to off. If you run it like below, SETDISPLAYS_DBG=0 setdisplays rightof the script will output the command it would have run, so you can investigate it. xrandr --output LVDS --auto --primary --output VGA-0 --auto --right-of LVDS Your output names will differ based on your hardware and potentially your video driver. -- Scott Sullivan
On Wed, Oct 01, 2014 at 08:42:14AM -0400, Scott Sullivan wrote:
I'm curious to see how your setting up your screens in .xinitrc. Can you post it?
Here is .xinitrc: xmodmap /home/pking/.Xmodmap xset +fp /usr/share/fonts/corefonts xset +fp /usr/share/fonts/default xset +fp /usr/share/fonts/ttf-bitstream-vera xset +fp /usr/share/fonts/urw-fonts xset +fp /usr/share/fonts/proggy-fonts xset +fp /usr/share/fonts/terminus xset +fp /usr/share/fonts/TTF # for continuous landscape + landscape: xrandr --output DVI-I-2 --right-of DVI-I-1 # for portrait + landscape: xrandr --output DVI-I-1 --rotate left xsetroot -cursor_name left_ptr urxvtd -q -o -f exec goomwwm
Follow up question, do any other WMs have an issue exiting with your xrandr setup?
Good question. I'll try cwm and dwm the next time I'm near that computer.
For comparison, I'm attaching a script I use to setup various secondary monitors with my laptop. As this is a laptop I defined a 'default' that sets all other outputs but the LCD to off.
Thanks for posting your script! -- Peter King peter.king@utoronto.ca Department of Philosophy 170 St. George Street #521 The University of Toronto (416)-978-4951 ofc Toronto, ON M5R 2M8 CANADA http://individual.utoronto.ca/pking/ ========================================================================= GPG keyID 0x7587EC42 (2B14 A355 46BC 2A16 D0BC 36F5 1FE6 D32A 7587 EC42) gpg --keyserver pgp.mit.edu --recv-keys 7587EC42
I have made a bit of progress on the problem -- but only a bit. (1) The keymap seems to be part of the problem: all the modifier keys are, as far as I can tell, dead. That's why keyboard-oriented tiling WMs are more or less stuck (goomwwm, cwm, dwm, etc.) (2) However, twm works as well as it ever did. (3) I followed the gentoo guide to setting up kernel modesetting using the intel graphics driver, and it made things appreciably worse: when the kernel boots and the graphical framebuffer is supposed to take over, the screen simply goes black. I can ssh in, though, and see that according to dmesg the inteldrmfb has successfully loaded, so what gives? (4) I also got no joy whatsoever at blind-typing into the console to start X, which, again, I could see by ssh was supposedly running but only gave a blank screen. All this on the internal 1680x1050 laptop screen. So while it was blank, I thought, why not, and plugged in an external VGA monitor. Voila! The console came up on the external and the internal screens. Better yet, so did X with twm. It seems to respond to xrandr, too. For one last oddity, I dropped to the console, unplugged the external monitor, and the internal screen still runs console/X just fine. Go figure. This is all on a Lenovo 3000 n100 0768 model, which I recall was finicky the last time I set it up (back in 2006 or so). But I don't recall dead keymaps, and, while I had to use the i915resolution hack back then, that should all have been superseded by kernel modesetting... -- Peter King peter.king@utoronto.ca Department of Philosophy 170 St. George Street #521 The University of Toronto (416)-978-4951 ofc Toronto, ON M5R 2M8 CANADA http://individual.utoronto.ca/pking/ ========================================================================= GPG keyID 0x7587EC42 (2B14 A355 46BC 2A16 D0BC 36F5 1FE6 D32A 7587 EC42) gpg --keyserver pgp.mit.edu --recv-keys 7587EC42
| From: Peter King <peter.king@utoronto.ca> | I have made a bit of progress on the problem -- but only a bit. | | (1) The keymap seems to be part of the problem: all the modifier keys are, | as far as I can tell, dead. That's why keyboard-oriented tiling WMs | are more or less stuck (goomwwm, cwm, dwm, etc.) Modifier keys dead? I assume you mean shift, alt, ctrl. That's pretty odd. It doesn't sound like something to do with lack of hardware support -- those are pretty standardized. | (2) However, twm works as well as it ever did. | | (3) I followed the gentoo guide to setting up kernel modesetting using the | intel graphics driver, and it made things appreciably worse: when the | kernel boots and the graphical framebuffer is supposed to take over, | the screen simply goes black. I can ssh in, though, and see that | according to dmesg the inteldrmfb has successfully loaded, so what | gives? Sometimes there is a proprietary thingee that gates the display. For example, I've had notebooks that had no backlight until something was, for lack of a more accurate word, turned on. Can you see a faint image? | (4) I also got no joy whatsoever at blind-typing into the console to start | X, which, again, I could see by ssh was supposedly running but only | gave a blank screen. SSH from another machine should let you do a lot. You can look at /var/log/Xorg.0.log You can actually run xrandr, for example, by setting DISPLAY: DISPLAY=:0 xrandr That probably assumes that you are logged in on X or the ownership will be wrong. Perhaps you can log in via blind typing. | All this on the internal 1680x1050 laptop screen. So while it was blank, | I thought, why not, and plugged in an external VGA monitor. Voila! The | console came up on the external and the internal screens. Better yet, so | did X with twm. It seems to respond to xrandr, too. Did the modifier keys work? See if /var/log/Xorg.0.log or dmesg or something are different in the two cases. | This is all on a Lenovo 3000 n100 0768 model, which I recall was finicky | the last time I set it up (back in 2006 or so). Google? I'm sure you've done that a lot. Try some live distros for variety? | But I don't recall dead | keymaps, and, while I had to use the i915resolution hack back then, that | should all have been superseded by kernel modesetting... Ask it to use VGA instead of intel?
It was a combination of four separate problems, which is why it took some while to run it down. First, it turned out that my kernel configuration was not building in the backlighting module for laptops (thanks for the suggestion Hugh!). Once I recompiled the kernel with backlighting, the console reappeared. Second, the Lenovo 3000 n100 0768, for reasons unknown to me, must have a strange hardware keymap, because the (left) Windows key, the (right) menu key, and the PrtScr and Pause keys are all completely dead, according to showkey. This, despite enabling "windows keys" in the keymap. Since the tiling window managers I was using used the left Windows key and the main modifier, and exited X using the Pause key, no wonder it all seemed dead. Third, I had somehow forgotten basic X facts, and in my .xinitrc had set an xterm to run but without putting it in the background -- which meant that startx never got to the line running the window manager, and so I had X up with no wm, no surprise this didn't work out too well. Fourth, I needed to run mkfontdir on several files in /usr/share/fonts, so X would have something with which to draw into an xterm. No cure for the dead keys, as far as I know, but I can set the wm to use different modifier keys and work around them. Now, if I can only figure out how to turn off the LCD screen and leave the external monitor on... Thanks to one and all for the suggestions! -- Peter King peter.king@utoronto.ca Department of Philosophy 170 St. George Street #521 The University of Toronto (416)-978-4951 ofc Toronto, ON M5R 2M8 CANADA http://individual.utoronto.ca/pking/ ========================================================================= GPG keyID 0x7587EC42 (2B14 A355 46BC 2A16 D0BC 36F5 1FE6 D32A 7587 EC42) gpg --keyserver pgp.mit.edu --recv-keys 7587EC42
| From: Peter King <peter.king@utoronto.ca> | Second, the Lenovo 3000 n100 0768, for reasons unknown to me, must have a | strange hardware keymap, because the (left) Windows key, the (right) menu | key, and the PrtScr and Pause keys are all completely dead, according to | showkey. Under X, have you tried the xev command? Are any events created for those keys?
Well, about Nouveau, the last stable kernel has a few improvments, including nouveau... Linus Torvalds has announced the latest stable release of the Linux kernel, version 3.17. Below is a short list compiling some notable highlights of this release. It’s by no means exhaustive. 1. Microsoft Xbox One controller support (without vibration) 2. Additional improvements to Sony SIXAXIS support 3. Toshiba “Active Protection Sensor” support 4. New ARM support includes Rockchip RK3288 and AllWinner A23 SoCs 5. “Cross-thread filter setting” for secure computing facility 6. Broadcom BCM7XXX-based board support (used in various set-top boxes) 7. Enhanced AMD Radeon R9 290 support 8. Misc. Nouveau driver improvements, including Kepler GPU fixes 9. Audio support includes Wildcatpoint Audio DSP on Intel Broadwell Ultrabooks. =================================================== Marcelo Cavalcante Rocha - Kalib Pós-Graduando em Governança de Tecnologia da Informação - ESTÁCIO/FIC Graduado em Sistemas de Informações - ESTÁCIO/FIC Usuário Linux #407564 | Usuário Asterisk #1148 Fortaleza - Ceará - Brazil Celular: +55 085 87620983 Certificações: ITIL V3 <http://www.itil-officialsite.com/Qualifications/ITILV3QualificationLevels/ITILV3FoundationQualificationinITServiceManagement.aspx> | CSM <http://www.scrumalliance.org/> | LPI-C1 <http://www.lpi.org/> | LPI-C2 <http://www.lpi.org/> | LPI-C3 <http://www.lpi.org/> | Novell CLA <http://www.novell.com/training/certinfo/cla/> Minha Pessoa: Blog <http://www.marcelocavalcante.net> Projetos: Tux-CE <http://www.tux-ce.org> | Archlinux-br <http://www.archlinux-br.org> | Chakra <http://www.chakra-project.org/> | KDE Brasil <http://br.kde.org> | TLUG <http://gtalug.org/wiki/Main_Page> | PUG-CE <http://pug-ce.python.org.br/blog/> =================================================== Proteja meu endereço como estou protegendo o seu. Não revele e-mail dos correspondentes: use Cco (Copia Carbonada Oculta). Retire os endereços antes de reenviar. Dificulte assim a disseminação de vírus e spam. 2014-10-07 18:30 GMT-03:00 D. Hugh Redelmeier <hugh@mimosa.com>:
| From: Peter King <peter.king@utoronto.ca>
| Second, the Lenovo 3000 n100 0768, for reasons unknown to me, must have a | strange hardware keymap, because the (left) Windows key, the (right) menu | key, and the PrtScr and Pause keys are all completely dead, according to | showkey.
Under X, have you tried the xev command? Are any events created for those keys?
--- GTALUG Talk Mailing List - talk@gtalug.org http://gtalug.org/mailman/listinfo/talk
These 'improvements' suggest that game boxen, smart phones and game playing are major thrust areas. I use my computer for work related tasks and and am looking toward process control systems. So how do I model a multi-card multi-monitor system that includes real time support so that it looks like a game which could also be accessed from a smart phone - - - then lots of developers will want to work on the idea? Dee On Wed, Oct 8, 2014 at 12:42 PM, Marcelo Cavalcante <kalibslack@gmail.com> wrote:
Well, about Nouveau, the last stable kernel has a few improvments, including nouveau...
Linus Torvalds has announced the latest stable release of the Linux kernel, version 3.17. Below is a short list compiling some notable highlights of this release. It’s by no means exhaustive.
1. Microsoft Xbox One controller support (without vibration) 2. Additional improvements to Sony SIXAXIS support 3. Toshiba “Active Protection Sensor” support 4. New ARM support includes Rockchip RK3288 and AllWinner A23 SoCs 5. “Cross-thread filter setting” for secure computing facility 6. Broadcom BCM7XXX-based board support (used in various set-top boxes) 7. Enhanced AMD Radeon R9 290 support 8. Misc. Nouveau driver improvements, including Kepler GPU fixes 9. Audio support includes Wildcatpoint Audio DSP on Intel Broadwell Ultrabooks.
===================================================
Marcelo Cavalcante Rocha - Kalib
Pós-Graduando em Governança de Tecnologia da Informação - ESTÁCIO/FIC Graduado em Sistemas de Informações - ESTÁCIO/FIC Usuário Linux #407564 | Usuário Asterisk #1148 Fortaleza - Ceará - Brazil Celular: +55 085 87620983 Certificações: ITIL V3 <http://www.itil-officialsite.com/Qualifications/ITILV3QualificationLevels/ITILV3FoundationQualificationinITServiceManagement.aspx> | CSM <http://www.scrumalliance.org/> | LPI-C1 <http://www.lpi.org/> | LPI-C2 <http://www.lpi.org/> | LPI-C3 <http://www.lpi.org/> | Novell CLA <http://www.novell.com/training/certinfo/cla/> Minha Pessoa: Blog <http://www.marcelocavalcante.net> Projetos: Tux-CE <http://www.tux-ce.org> | Archlinux-br <http://www.archlinux-br.org> | Chakra <http://www.chakra-project.org/> | KDE Brasil <http://br.kde.org> | TLUG <http://gtalug.org/wiki/Main_Page> | PUG-CE <http://pug-ce.python.org.br/blog/>
===================================================
Proteja meu endereço como estou protegendo o seu. Não revele e-mail dos correspondentes: use Cco (Copia Carbonada Oculta). Retire os endereços antes de reenviar. Dificulte assim a disseminação de vírus e spam.
2014-10-07 18:30 GMT-03:00 D. Hugh Redelmeier <hugh@mimosa.com>:
| From: Peter King <peter.king@utoronto.ca>
| Second, the Lenovo 3000 n100 0768, for reasons unknown to me, must have a | strange hardware keymap, because the (left) Windows key, the (right) menu | key, and the PrtScr and Pause keys are all completely dead, according to | showkey.
Under X, have you tried the xev command? Are any events created for those keys?
--- GTALUG Talk Mailing List - talk@gtalug.org http://gtalug.org/mailman/listinfo/talk
--- GTALUG Talk Mailing List - talk@gtalug.org http://gtalug.org/mailman/listinfo/talk
On Tue, Sep 30, 2014 at 7:36 PM, Peter King <peter.king@utoronto.ca> wrote:
I have an annoying problem. I run a two-monitor setup with xrandr, one monitor in portrait and the other in landscape mode. I set xrandr in the .xinitrc file, and it works well. The only problem is that when I quit the window manager (in this case goomwwm), the screen goes to black and I'm pretty sure I never return to the console (blind typing produces no results). The computer itself keeps running; I can ssh in. Any advice or suggestions? I can't quite figure how to turn xrandr *off* once it's been set in .xinitrc. Oh, and if I boot without xrandr, I go straight back to the console when I quit X (as it should).
I don't really know how to modify xrandr but I can tell you that part of your problem is that xrandr does NOT like multi-monitor setups. This I found out the hard way when I tried to setup my multi-monitor system. I had to not enable xrandr as part of my solution. Perhaps some time in the future the authors of xrandr will update their solution to better reflect the real world. Dee
On Wed, Oct 01, 2014 at 08:57:45AM -0500, o1bigtenor wrote:
I don't really know how to modify xrandr but I can tell you that part of your problem is that xrandr does NOT like multi-monitor setups. This I found out the hard way when I tried to setup my multi-monitor system. I had to not enable xrandr as part of my solution. Perhaps some time in the future the authors of xrandr will update their solution to better reflect the real world.
How did you set up your multimonitor system without xrandr? I'm perfectly willing to do it some other way -- I'm committed to the solution, not the means. -- Peter King peter.king@utoronto.ca Department of Philosophy 170 St. George Street #521 The University of Toronto (416)-978-4951 ofc Toronto, ON M5R 2M8 CANADA http://individual.utoronto.ca/pking/ ========================================================================= GPG keyID 0x7587EC42 (2B14 A355 46BC 2A16 D0BC 36F5 1FE6 D32A 7587 EC42) gpg --keyserver pgp.mit.edu --recv-keys 7587EC42
On Wed, Oct 1, 2014 at 10:42 AM, Peter King <peter.king@utoronto.ca> wrote:
On Wed, Oct 01, 2014 at 08:57:45AM -0500, o1bigtenor wrote:
I don't really know how to modify xrandr but I can tell you that part of your problem is that xrandr does NOT like multi-monitor setups. This I found out the hard way when I tried to setup my multi-monitor system. I had to not enable xrandr as part of my solution. Perhaps some time in the future the authors of xrandr will update their solution to better reflect the real world.
How did you set up your multimonitor system without xrandr? I'm perfectly willing to do it some other way -- I'm committed to the solution, not the means.
On Wed, Oct 1, 2014 at 10:42 AM, Peter King <peter.king@utoronto.ca> wrote:
On Wed, Oct 01, 2014 at 08:57:45AM -0500, o1bigtenor wrote:
I don't really know how to modify xrandr but I can tell you that part of your problem is that xrandr does NOT like multi-monitor setups. This I found out the hard way when I tried to setup my multi-monitor system. I had to not enable xrandr as part of my solution. Perhaps some time in the future the authors of xrandr will update their solution to better reflect the real world.
How did you set up your multimonitor system without xrandr? I'm perfectly willing to do it some other way -- I'm committed to the solution, not the means.
Sorry for the blank message!! (to the list!!) I'm running nvidia graphics cards (2 connected to monitors) so its a royal pain but I'm running the nvidia drivers. Its not that easy to set up the first time (I spent a lot of time the first time) but now its a painful process but its quite doable. You need to remove the nouveau driver. Means that first you need to get your system running on something else and then you black list nouveau and then you install the nvidia drivers and then you need to setup all the goodies and then you can turf the nouveau drivers and then you have OK performance. Dee
On 10/01/2014 01:14 PM, o1bigtenor wrote:
I'm running nvidia graphics cards (2 connected to monitors) so its a royal pain but I'm running the nvidia drivers. Its not that easy to set up the first time (I spent a lot of time the first time) but now its a painful process but its quite doable.
Sounds like your doing multi-card, not specifically multi-monitor, am I mistaken? Xrandr works great for multi-monitor, but does indeed assume single card. The assumption is not Xrandr's fault, support for multi-card been a target feature of Xorg to for the last decade. Multi-monitor single cards have been "good enough" for just as long.
You need to remove the nouveau driver. Means that first you need to get your system running on something else and then you black list nouveau and then you install the nvidia drivers and then you need to setup all the goodies and then you can turf the nouveau drivers and then you have OK performance.
Fought with this today actually. Add to your kernel boot line: modprobe.blacklist=nouveau This gets it gone before it's used to by the console. -- Scott Sullivan
On 10/06/2014 09:09 PM, Scott Sullivan wrote:
On 10/01/2014 01:14 PM, o1bigtenor wrote:
I'm running nvidia graphics cards (2 connected to monitors) so its a royal pain but I'm running the nvidia drivers. Its not that easy to set up the first time (I spent a lot of time the first time) but now its a painful process but its quite doable.
Sounds like your doing multi-card, not specifically multi-monitor, am I mistaken?
Xrandr works great for multi-monitor, but does indeed assume single card. The assumption is not Xrandr's fault,
support for multi-card been _has not been_ a target feature
of Xorg to for the last decade. Multi-monitor single cards have been "good enough" for just as long.
On Mon, Oct 6, 2014 at 8:09 PM, Scott Sullivan <scott@ss.org> wrote:
On 10/01/2014 01:14 PM, o1bigtenor wrote:
I'm running nvidia graphics cards (2 connected to monitors) so its a royal pain but I'm running the nvidia drivers. Its not that easy to set up the first time (I spent a lot of time the first time) but now its a painful process but its quite doable.
Sounds like your doing multi-card, not specifically multi-monitor, am I mistaken?
2 cards and 4 monitors 1 monitor in portrait and 3 in landscape mode (wouldn't mind a 4th in landscape!)
Xrandr works great for multi-monitor, but does indeed assume single card. The assumption is not Xrandr's fault, support for multi-card been a target feature of Xorg to for the last decade. Multi-monitor single cards have been "good enough" for just as long.
My research showed that multi-card has been a 'hope' for a very long time. It must be because few people actually run more than 1 graphics card. I wish there were an easy way to slave cards together (I actually have 3 graphics cards - - - wanted them for serious numeric processing).
You need to remove the nouveau driver. Means that first you need to get
your system running on something else and then you black list nouveau and then you install the nvidia drivers and then you need to setup all the goodies and then you can turf the nouveau drivers and then you have OK performance.
Fought with this today actually. Add to your kernel boot line:
modprobe.blacklist=nouveau
This gets it gone before it's used to by the console.
One needs to NOT be running any x window stuff at that time too! The whole process isn't very straightforward. Its a pity that nouveau weren't just either seriously upgraded or perhaps deprecated. Dee
| From: o1bigtenor <o1bigtenor@gmail.com> | Its a pity that nouveau | weren't just either seriously upgraded or perhaps deprecated. Nouveau impresses me given that it is a product of reverse engineering. If you were to deprecate it, I would hope that you have an open source replacement.
On October 6, 2014 11:35:48 PM EDT, "D. Hugh Redelmeier" <hugh@mimosa.com> wrote:
| From: o1bigtenor <o1bigtenor@gmail.com>
| Its a pity that nouveau | weren't just either seriously upgraded or perhaps deprecated.
Nouveau impresses me given that it is a product of reverse engineering.
If you were to deprecate it, I would hope that you have an open source replacement.
--- GTALUG Talk Mailing List - talk@gtalug.org http://gtalug.org/mailman/listinfo/talk
o1bigtenor, Are you possibly confusing the old nv driver with the nouveau driver that replaced? http://www.x.org/wiki/nv/ http://nouveau.freedesktop.org/wiki/ -- Scott Sullivan
On Mon, Oct 6, 2014 at 10:51 PM, Scott Sullivan <scott@ss.org> wrote:
On October 6, 2014 11:35:48 PM EDT, "D. Hugh Redelmeier" <hugh@mimosa.com> wrote:
| From: o1bigtenor <o1bigtenor@gmail.com>
| Its a pity that nouveau | weren't just either seriously upgraded or perhaps deprecated.
Nouveau impresses me given that it is a product of reverse engineering.
If you were to deprecate it, I would hope that you have an open source replacement.
--- GTALUG Talk Mailing List - talk@gtalug.org http://gtalug.org/mailman/listinfo/talk
o1bigtenor,
Are you possibly confusing the old nv driver with the nouveau driver that replaced?
No - - - nouveau goes absolutely sick when you try to run more than one graphics card (unless that has changed in the last 4 to 6 months!). Dee
On Mon, Oct 6, 2014 at 10:35 PM, D. Hugh Redelmeier <hugh@mimosa.com> wrote:
| From: o1bigtenor <o1bigtenor@gmail.com>
| Its a pity that nouveau | weren't just either seriously upgraded or perhaps deprecated.
Nouveau impresses me given that it is a product of reverse engineering.
If you were to deprecate it, I would hope that you have an open source replacement.
As a user nouveau feels like it was created to be 'good enough' rather than to lead. It is because of the lack of support for multi-card and multi-monitor systems that I say this. In fairness I see a lot of businesses using a 17 or a 19 inch LCD monitor - - - someone is really missing the productivity increase that you can get from more screen real estate. My system is set up with 20 desktops where I have 16 defined for different tasks and/or needs. This available space makes it easy to switch from one task to another without taking a lot of time to find the various bits needed to do the work. Dee
participants (5)
-
D. Hugh Redelmeier -
Marcelo Cavalcante -
o1bigtenor -
Peter King -
Scott Sullivan