golinux | The netinst iso is very limited. | 00:00 |
---|---|---|
Achylles | and not very intuitive. So that I used the netinstall.iso and did using the expert mode | 00:00 |
golinux | Maybe there's a meta package missing | 00:01 |
Achylles | The desktop-live.iso does not have an expert mode install option :( | 00:01 |
Achylles | which I am used since many years ago, using debian | 00:01 |
golinux | No. It rsyncs the files | 00:02 |
Achylles | That refractainstall is really a pain | 00:02 |
golinux | LOL! | 00:02 |
Achylles | I mean, the layout... | 00:02 |
golinux | Many find it much easier. | 00:02 |
Achylles | I find terrible | 00:02 |
Achylles | I prefer the expert mode | 00:03 |
Achylles | much easier | 00:03 |
golinux | Did ou try launching it from terminal? | 00:03 |
Achylles | yes | 00:03 |
Achylles | the two ways... | 00:03 |
golinux | The gui kind of sucks | 00:03 |
Achylles | It sucks too much I would say | 00:04 |
golinux | Hmmm and fsmithred will not be around much for the next few weeks . . . | 00:04 |
golinux | I wish I had the magic answer. | 00:05 |
Achylles | Perhaps, it could be changed to something else | 00:05 |
golinux | Feel free to provide an alternative. | 00:05 |
Achylles | Sometimes people want to make things easier to help noobs, but they make things more complicated | 00:06 |
Achylles | My alternative is to add the expert mode (graphical and text) to the desktop-live.iso | 00:06 |
golinux | That is a whole other can of worms | 00:07 |
Achylles | Yeah. So, leave it | 00:08 |
Achylles | I just need to figure out how to configure my grub image from the blue one to the devuan default | 00:09 |
Digit | nope. this is getting irritating now. cant start any game without all cores filling up, and much over-burdened fan. i think this is new ever since upgrading. | 06:13 |
Digit | i wonder if i ... ~ nope, my brains thinking up gentoo style solutions. | 06:14 |
Digit | M-x tetris | 06:17 |
* Digit thanks emacs | 06:17 | |
Centurion_Dan | Digit: Are you sure your not now using the nouveau driver instead of nvidia? | 06:53 |
Centurion_Dan | lsmod should help expose that... | 06:55 |
Digit | lsmod | egrep 'nvidia|nouveau' returns only lines mentioning nvidia. none mention nouveau. though, i did at first check to see if nouveau is installed, and to my surprise it is. | 06:56 |
Digit | you did have me going thinking, wondering, if that could have explained some quirks i saw. | 06:56 |
Digit | i presume it's safe to uninstall libdrm-nouveau2 and xserver-xorg-video-nouveau | 06:57 |
gnarface | you shouldn't have to | 06:59 |
gnarface | it should be sufficient to blacklist the nouveau driver | 06:59 |
gnarface | which is probably already happening if the nvidia one is loading successfully | 06:59 |
gnarface | your problem is likely something else | 06:59 |
gnarface | did you ever try vsync like i suggested? | 07:00 |
Digit | i poked around in the nvidia-settings looking for it, wasnt sure i found it, but changed a couple things to no effect. | 07:00 |
gnarface | unfortunately there isn't just one place to set it | 07:01 |
gnarface | there is one place in nvidia-settings | 07:01 |
gnarface | also your window manager/compositor could have vsync settings of their own independent of that | 07:01 |
gnarface | and while the combinations may be educational for debugging different ways they can override or conflict with each other, | 07:02 |
gnarface | the most likely method to be reliable and not horrendously detrimental to your performance | 07:02 |
gnarface | is to just shut them all off and only enable the in-game settings for each opengl game having this problem | 07:03 |
gnarface | and this may not be the problem either, but you're never gonna narrow it down without trying | 07:03 |
Digit | i did change some settings in minetest, turned everything off basically. still same problem | 07:04 |
gnarface | the driver should also be able to obey the environment variable __GL_SYNC_TO_VBLANK=0 | 07:04 |
gnarface | if you run glxgears with that, set to 0 and 1 separately, you should be able to see a stark difference in the framerate output | 07:05 |
gnarface | if not, then you know whether it's stuck on or stuck off, at least | 07:05 |
Digit | no notable difference both in the 250-270 range | 07:11 |
Digit | 260-280 i meant. ~sleepy. | 07:11 |
gnarface | Digit: that suggests it is not working in either case. something is going wrong. probably one of the myriad vsync settings is conflicting with another. | 07:38 |
gnarface | still no way to tell if that's the actual main source problem, but it is definitely a problem. | 07:38 |
gnarface | (also 280fps is way too low to be hardware accelerated by any nvidia card that is supported by the current drivers) | 07:40 |
gnarface | (so that's two things clearly wrong, at minimum. you had better examine your Xorg log) | 07:41 |
Digit | ah yup. re-investigating Xorg, searching for EE, i see a likely culprit. "Failed to initialize the GLX module". i'll leave it at this point to pick up from after sleep. great thanks for the pointers. :) | 07:44 |
Centurion_Dan | ah, sound like your missing the nvidia glx libs. | 07:50 |
gnarface | probably, but go so far as to suggest someone should bundle all that nvidia crap into a single task to avoid repeating this groundhog day over and over again, and everyone loses their shit | 08:02 |
gnarface | Digit: it could be a lot of things going wrong, but unless you mixed repos, chances are you're just missing the package "libgl1-nvidia-glx" | 08:03 |
Digit | was just advised to mix repos earlier, for eudev from ascii, to avoid the one from ceres. | 08:04 |
gnarface | interesting | 08:04 |
gnarface | well i wouldn't have suggested doing that | 08:04 |
gnarface | but mixing devuan versions isn't as dangerous as mixing distros, at least | 08:04 |
gnarface | the most important thing is that all the "*nvidia*" have the same version number | 08:05 |
Digit | n i'm leaving synaptic all multi-coloured, and with "mark additional required changes" up showing what the next step of mess of removing n installing would be if i tried install a different glx nvidia lib | 08:05 |
Digit | until i've slept, n have a fresh head. | 08:06 |
gnarface | that's probably a wise choice | 08:06 |
* Digit cyclops-zombying | 08:06 | |
golinux | Digit: You still here? | 08:50 |
golinux | sgfxi takes the pain out of an Nvidia install. | 08:51 |
golinux | At least it used to. I use nouveau now with onboard graphics. | 08:51 |
golinux | You can find it here: https://smxi.org/ | 08:52 |
golinux | It takes care of blacklisting nouveau and installs from Nvidia iirc. | 08:53 |
golinux | Sets everything up 1 2 3 | 08:54 |
skyroveRR | Hey guys, I'm using devuan ASCII 2.0.0 on an intel, it's a standard install. My monitor is capable of 1920x1080, but the kernel for some reason is setting it to only 1024x768. I can't find any other higher setting in Settings > Display in XFCE. Any ideas? | 09:16 |
gnarface | see how this stuff is just like a laugh track on a loop? | 09:18 |
gnarface | skyroveRR: check your Xorg log, it's probably defaulting to vesa or generic framebuffer drivers | 09:18 |
gnarface | there could be other reasons | 09:19 |
gnarface | but you have to see the Xorg log to find out | 09:19 |
gnarface | it will probably either be in /var/log/ or ~/.local/share/xorg/ | 09:19 |
skyroveRR | Ok, let me have a look. | 09:20 |
gnarface | it should be called Xorg.0.log but there are cases where the "0" might iterate | 09:21 |
skyroveRR | gnarface: I'm looking at the log file, and it says "Falling back to old probe method for vesa", and it goes on to load some kind of a sub module called glamoregl. | 09:23 |
skyroveRR | Let me pastebin the log, one moment. | 09:23 |
skyroveRR | gnarface: http://termbin.com/kxyz | 09:24 |
gnarface | skyroveRR: paste it to me privately, or use paste.debian.net and i'll look at it. you should be specifically looking for lines with "(EE)" on them, but what you've said already supports my initial hypothesis | 09:26 |
skyroveRR | Ok, one moment. | 09:27 |
skyroveRR | There are no errors (EE) lines in the log file, gnarface | 09:29 |
gnarface | how about (WW) ? | 09:29 |
gnarface | i assume there's a lot of (WW), but we're primarily interested in the first ones, and the unique ones. we're not interested in hundreds of repeats of variations of resolutions and refresh rates being invalidated; we already expected that to be happening | 09:30 |
skyroveRR | gnarface: http://paste.debian.net/1034333/ | 09:31 |
Centurion_Dan | skyroveRR: It's an intel gpu it seems... can you send send a paste of the output of lspci?? | 09:32 |
skyroveRR | g There are three WWs, one is The directory "/usr/share/fonts/X11/cyrillic" does not exist., other is Falling back to old probe method for fbdev and the last one is Falling back to old probe method for vesa . | 09:32 |
skyroveRR | Centurion_Dan: sure. | 09:32 |
gnarface | yea, i've run into the same issue before with at least a dozen different machines | 09:33 |
gnarface | probably all it needs is for you to specifically add the call to the intel driver in an xorg.conf snippet | 09:33 |
gnarface | i don't know why it's not smart enough to recognize it's own | 09:33 |
skyroveRR | Centurion_Dan: http://paste.debian.net/1034335/ | 09:34 |
skyroveRR | gnarface: alright. | 09:35 |
skyroveRR | gnarface: I think udev itself hasn't probed my GPU yet, and that's causing it to try other modules? | 09:36 |
gnarface | skyroveRR: a potential cause, sure, and just as likely mundane permissions, but if you tell it specifically to try the intel driver first then you'll know for sure | 09:37 |
gnarface | lately i've been seeing many more instances of it just trying vesa and fbdev first and just deciding "ok, good enough" to one or the other without even getting as far as trying the Intel one | 09:38 |
Centurion_Dan | also make sure that you have the package xserver-xorg-video-intel installed | 09:39 |
gnarface | yea, definitely. although, Xorg *should* have grabbed *all* the packages whether it thought you needed them or not | 09:40 |
skyroveRR | Centurion_Dan: I'll do that, one moment | 09:40 |
skyroveRR | Centurion_Dan: gnarface: I do have intel_drv.so inside /usr/lib/xorg/modules/drivers directory.. | 09:42 |
skyroveRR | Meaning it did grab all the packages, xserver-xorg-video-*. | 09:42 |
skyroveRR | ati, intel, nouveau, vmware... all drv.so files are there. | 09:43 |
skyroveRR | In which sensible directory inside devuan can I put my new xorg.conf file? | 09:44 |
gnarface | you should put it in /etc/X11/ if it is a whole config file, or /etc/X11/xorg.conf.d/ if you are just making a snippet | 09:45 |
gnarface | you probably just need a snippet | 09:45 |
gnarface | probably just the driver section | 09:45 |
Centurion_Dan | it's definitely intel... you might have to force loading the intel module - can you paste the output of lscpi -vvv | 09:46 |
skyroveRR | Centurion_Dan: sure | 09:46 |
skyroveRR | Centurion_Dan: only pasted the output of the VGA section; if you want the complete output (which will be pain to copy to paste.debian.net), let me know: http://paste.debian.net/1034339/ | 09:49 |
Centurion_Dan | make that lspci -n -vvv - need the pci-id | 09:51 |
skyroveRR | Hm? | 09:51 |
skyroveRR | lspci -n -vvv ? | 09:51 |
Centurion_Dan | try modprobe i915 for a start and restart your DE | 09:52 |
Centurion_Dan | yup. | 09:52 |
skyroveRR | i915 is already loaded in lsmod | 09:52 |
Centurion_Dan | see what xorg says then... | 09:52 |
skyroveRR | gnarface: there's no xord.conf.d inside /etc/X11... do I create it? Are devuan's X startup files written in any way to parse that directory? | 09:55 |
Centurion_Dan | skyroveRR: don't do that... you shouldn't need to ever create and xorg.conf | 09:57 |
skyroveRR | Then? | 09:57 |
Centurion_Dan | did the i915 in the lsmod output have a 0 or a 1 next to it/ | 09:58 |
Centurion_Dan | ? | 09:58 |
skyroveRR | It had 4 next to it | 09:59 |
Centurion_Dan | btw what version you running? | 09:59 |
skyroveRR | Devuan version? | 09:59 |
Centurion_Dan | yeah and kernel | 09:59 |
skyroveRR | ASCII 2.0.0, kernel umm... Linux devuan 4.9.0-6-amd64 #1 SMP Debian 4.9.88-1+deb9u1 (2018-05-07) x86_64 GNU/Linux | 10:00 |
skyroveRR | Wonder which one it is.. | 10:01 |
skyroveRR | Oh, it's 4.9.0-6. | 10:01 |
skyroveRR | Wonder why there's 4.9.88-1+deb9u1 in uname -a.. | 10:01 |
Centurion_Dan | 4.9.0-6 is the package name, and 4.9.88-1+* is the debian package version | 10:08 |
skyroveRR | Centurion_Dan: so where do I modify the xorg.conf? | 10:11 |
Centurion_Dan | you shouldn't need to... | 10:14 |
skyroveRR | ... | 10:16 |
gnarface | skyroveRR: sorry, wandered away, yes, create it. if you look at the top of the log file, it should specifically tell you all the places it is checking, in fact | 10:17 |
gnarface | i'm not sure there wasn't also a module to modprobe the last time i had to do this... | 10:17 |
skyroveRR | I created /usr/share/X11/xorg.conf.d/20-intel.conf as per https://wiki.archlinux.org/index.php/intel_graphics and added the 4 lines. | 10:19 |
gnarface | i like to keep all my configs in /etc just so they are easier to back up | 10:20 |
skyroveRR | Yeah, but the X server in devuan likes to see things in /usr/share/X11 :) | 10:20 |
skyroveRR | Or... does it look in /etc/X11 also? | 10:20 |
skyroveRR | brb | 10:21 |
gnarface | like i said, the log should tell you where it looks | 10:26 |
gnarface | it won't actually matter which location you choose, other than in the way it affects the order your file is parsed in respect to others | 10:27 |
gnarface | man intel | 10:36 |
gnarface | the "intel" man page is actually the man page for this driver | 10:37 |
gnarface | there are other options, but you most likely won't need to touch any of them | 10:37 |
gnarface | there wasn't non-free firmware for intel video, was there? | 10:39 |
Centurion_Dan | not that I'm aware of... but I haven't seen much intel hardware lately... | 10:41 |
Centurion_Dan | skyroveRR: /etc/X11/xorg.d/<name> and /etc/X11/xorg.conf - I'd suggest using the former.. | 10:43 |
Centurion_Dan | see the xorg.conf manpage | 10:44 |
skyroveRR | Hmm, reading the man page and making changes hardly made a difference, gnarface and Centurion_Dan.... I tried adding Option "ReprobeOutputs" "true" in the hope that it would work, but it didn't. I remember that when screen resolution was set correctly at first, during my initial install of devuan 1 JESSIE, the resolution was set in the boot process itself, right from the GRUB to the automated textual | 12:46 |
skyroveRR | boot process to SLIM login manager. My point is, the resolution is not being set from GRUB itself like it used to in my case. | 12:46 |
rafalcpp | how to configure static eth.. names for eth cards, using udev rules, or using /etc/iftab ? | 13:18 |
gnarface | skyroveRR: did the log file say it at least loaded the intel driver successfully? | 13:19 |
gnarface | rafalcpp: the file is /etc/network/interfaces | 13:19 |
rafalcpp | gnarface: hm? network/interfaces is used to rename eth.. cards device names? | 13:24 |
gnarface | rafalcpp: sorry, no that would be udev; look in /etc/udev/rules.d/70-persistent-net.rules | 13:25 |
skyroveRR | gnarface: yes it did. | 13:25 |
gnarface | well that's definitely no good | 13:26 |
gnarface | you can go further with the config | 13:27 |
gnarface | you can add a monitor section, to specify horzontal and veritical refresh ranges | 13:28 |
gnarface | maybe even a custom modeline | 13:28 |
gnarface | you really shouldn't have to though | 13:28 |
gnarface | since it is just a regular LCD panel right? | 13:28 |
gnarface | skyroveRR: ^ | 13:29 |
rafalcpp | ok | 13:30 |
gnarface | rafalcpp: *-persistent-net.rules; i don't know if it's gonna be "70" for you or not | 13:32 |
skyroveRR | gnarface: yup. | 13:32 |
skyroveRR | gnarface: but my observation is, if things were right, they would be, right from the boot process itself. From stage 1 grub. | 13:33 |
gnarface | i'm not sure if they changed that about grub | 13:34 |
gnarface | you can check in /etc/default/grub to see if the comments have an example of how to set it | 13:34 |
gnarface | maybe it will help if you can | 13:35 |
gnarface | i really don't know | 13:35 |
gnarface | i don't remember that ever being necessary for me though | 13:35 |
skyroveRR | I think if the kernel can't properly set the resolution at grub itself, there's very little point in playing around with X. | 13:38 |
skyroveRR | I'll try playing with GRUB_GFXMODE by setting it to 1920x1080 like you said in /etc/default/grub. Let's see. | 13:39 |
gnarface | eh, you'd think that, but the kernel may be using some vesa text mode or KMS to set the resolution of the system virtual console; xorg may just have coincidentally been inheriting it as a default in jessie | 13:39 |
gnarface | believe it or not, Xorg has no requirements to obey that setting or preserve it | 13:40 |
gnarface | the nvidia binary drivers can't even make it match the desktop resolution | 13:40 |
gnarface | they don't play nice with KMS | 13:40 |
skyroveRR | I've yet to be an nvidia customer :) | 13:40 |
gnarface | it could be the kernel too | 13:42 |
gnarface | that is still a possibility | 13:42 |
gnarface | what do you get as the output for "lsmod |grep i915" ? | 13:42 |
gnarface | i'm curious what it depends on there | 13:43 |
skyroveRR | http://paste.debian.net/1034366/ | 13:44 |
gnarface | i dunno, seems normal | 13:45 |
Digit | well, that was unpleasant. sgfxi has not proven smooth n helpful, this time. | 15:02 |
Digit | pffff. i really dont have the energy for this. ~ though am very glad i can still magically boot into X despite all the nvidia version mismatch warnings now. ... TTY are sideways for me, with my portrait mounted monitors, so that adds to the not-fun. | 15:04 |
Digit | in trying to retreat from the errors n failed exit sgfxi gave, i'm told to run sudo dpkg --configure -a, but upon its first question, "blah blah blah. Would you like to run `nvidia-xconfig --restore-original-backup` to attempt restoration of the original X configuration file?", either way i answer, Y or N (just hit enter for default) it just seems to hang. | 15:07 |
Digit | oh, at least Ctrl-C has an effect in my gui term. it just seemed to be stuck in tty. | 15:08 |
Digit | was it working, silently? ~ | 15:08 |
sruiz_ | Hi! a question is there a Devuan page showing plates, processors or PCs "friendly" with the system? | 15:24 |
djph | should be anything that Debian works on for the most part | 15:26 |
* Digit ready to table flip, as even in gui, the "Welcome to the NVIDIA Software Installer for Unix/Linux" that came up amidst apt'ing, has hung... locking up apt/synaptic/dpkg again. | 15:29 | |
* Digit impatient kills dpkg, (unduly(?)) certain it's not doing anything | 15:32 | |
* Digit wrestles his way out of the loop, finding points of interjection he can deviate from, says no to the first question asking if wanting to remove some nvidia xorg blah blah that when reading it he had been saying yes to | 15:34 | |
skyroveRR | That's a lot of self-talking.... | 15:36 |
Digit | sorry. | 15:37 |
Digit | well, sit-rep: cant do much with apt. can play games again without them chewing up my cpu causing overheating. | 15:37 |
]BFG[ | nvidia what a bummer, they have so much money why cant they opensource their shitty drivers 100% they would still make billions from hw sales | 15:48 |
]BFG[ | anyways way better than ati with non/existent broken drivers | 15:48 |
unmy | from the same reason like most of the companies :/ they don't wanna share with own super secret code! | 15:50 |
fsmithred | or they are worried that others will notice that the secret code was borrowed/stolen | 15:51 |
unmy | or broken and/or with many shit hacks :D | 15:52 |
fsmithred | lol, that, too. | 15:54 |
MinceR | they probably licensed a ton of patents and maybe even code from other sources and aren't allowed to show the code even if they wanted to | 15:57 |
unmy | and not always legaly using other sources... :P | 16:02 |
MinceR | also they might not want to show what they're doing so it isn't so easy for patent abusers to find a case against them | 16:03 |
unmy | brutal reality! | 16:08 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!