Nematocyst | upgraded to chimaera 4 last night. changed my desktop background to an image stored on my desktop. i don't know how to get back the default that I had before. creating a new user, it's coming from /usr/share/desktop-base, yet there are no image files there. don't know how to get them selected | 05:56 |
---|---|---|
golinux | Nematocyst: Try /usr/share/images/desktop-base/ | 06:02 |
Nematocyst | golinux, thanks. that worked | 06:03 |
golinux | Excellent | 06:03 |
golinux | I am also happy to hear that the default pleases you | 06:04 |
Nematocyst | i'm surprised that location isn't a default (like backgrounds) if you want to return to them easily. must have been a brain fart (i had seen usr/share/desktop-base when checking the new user... must have actually been the directory you mentioned) | 06:06 |
golinux | desktop-base formats all components of the desktop theme | 06:08 |
golinux | I still get confused by some of the locations | 06:10 |
Nematocyst | fyi, it was a seriously painless dist-upgrade. did it yesterday, started at 6:30pm backing up with rsync. rsync done at 7:44. updated with everything working save a couple minor details by 8:40. never seen anything like that | 06:10 |
Nematocyst | it's one thing to be that solid if you install v3, update to v4 immediately... but been running beowulf since it arrived | 06:11 |
GoatAvenger | Backuping up before a dist-upgrade... Good idea. I should do that.. | 07:40 |
hyrcanus | i'm against software paternalism, but apt-get dist-upgrade could at least encourage that in a message to the user | 07:41 |
djph | hyrcanus: huh? | 12:25 |
hyrcanus | making a backup before upgrade or dist-upgrade | 12:26 |
djph | a | 12:29 |
djph | *ah | 12:29 |
hyrcanus | would you reccommend a backup before an unexperienced person performs upgrade or dist-upgrade? | 12:36 |
hyrcanus | it's been a long time since i ran into serious problems myself | 12:36 |
djph | hyrcanus: I mean, I always do, but because I know if I don't, the machine spirit will know, and cause something to go wrong. | 12:44 |
hyrcanus | :) | 12:44 |
djph | it's the same reason printers get a blood-sacrifice when changing the fuser drum | 12:44 |
linux_ | hello. for some reason, my 4.0 installation is attempting to upgrade some stable packages to the backport ones while using apt. where to start with this? | 14:13 |
linux_ | if I comment the backports line on sources.list, the system appears to be fully upgrading (i'm only using the bpo kernel as far as I can tell) | 14:14 |
linux_ | fully upgraded* | 14:14 |
linux_ | https://paste.debian.net/hidden/b2d38f10/ | 14:18 |
freem | Hi | 17:14 |
gnarface | Hi, freem. it's a slow channel, just ask your question and wait | 17:17 |
freem | I know | 17:17 |
freem | I'm just reading stuff before, to formulate it better. | 17:17 |
gnarface | if that linux_ person comes back tell them that's expected behavior because backports packages are versioned to look newer than the stable ones | 17:18 |
buZz | ah good point | 17:18 |
hyrcanus | this is a nice place to learn apt-y things | 17:18 |
hyrcanus | i am also very grateful to all the work debian and devuan maintainers do to keep things trouble free | 17:19 |
hyrcanus | thank you! | 17:19 |
freem | Well, basically, there's a conjonction between a very old project of mine, and old memory of mine of devuan having tools to make forking easier, and an increased boredom, and I was wondering if I could get infos on some of the very lower level layers of debian/devuan here. Still have to formulate it correctly to not waste other people's time | 17:20 |
hyrcanus | freem talks of forks, and then Pie arrives. and now I'm hungry. | 17:24 |
freem | heh | 17:24 |
freem | good forks are a useful tools to eat Pie indeed | 17:24 |
hyrcanus | since we're blathering anyway, can you try with your on-topic question? | 17:25 |
freem | (also, I don't think a toy stuff can be called a fork) | 17:25 |
freem | well, I was wondering about how much the pre/post inst/rm scripts are vital. | 17:26 |
hyrcanus | i think some packages need them to do some housecleaning, others do not. | 17:26 |
freem | I know and understand their roles, mind you. Currently, I'm reading at kernel's ones (good old dpkg-deb is useful for that) | 17:26 |
freem | I think those scripts are one of the major reasons to have bash classified as essential | 17:27 |
freem | in some cases, I think they could be not necessary, for a less general distro, and was thinking about just repackaging some packages, to start building a base for that toy project | 17:28 |
freem | I also noted, in the debian10's lifecycle, an increase in RAM usage for stuff like initrd | 17:29 |
hyrcanus | oh, so the debian packages do things that require bashisms beyond a posix shell? | 17:30 |
freem | which I think is related, since I'm pretty sure those scripts are directly responsible for regenarating initrd (which, IIRC, now requires more than 250MiB to boot, and yes, that's a lot) | 17:30 |
freem | that's what I suspect | 17:30 |
freem | I can't see any other reason to have bash considered essential | 17:30 |
freem | OTOH Debian is rather... prudent (and I like that fact) so maybe its just some artefact of long forgotten bashisms | 17:31 |
gnarface | yes, leading cause of "can't boot after switching init shell to dash" is old bashisms in init scripts (often obsolete or out-of-repo ones) | 17:32 |
hyrcanus | i can only generally agree that dependencies that can be avoided should be. but maybe the manpower and energy isn't available to remove that dependency | 17:32 |
freem | in any cases, I think those scripts are... mostly annoying, and might be avoided. I would not bother anyone with that, it's better to do and then show | 17:32 |
gnarface | however, i'm certain you don't need more than 250MiB to boot the current initrd | 17:32 |
freem | hyrcanus: exactly. I'm not asking to anyone to do the job, I'm just searching for infos to toy with it myself | 17:32 |
freem | lemme check my last stack of VMs then | 17:33 |
gnarface | total used memory after a clean boot on ceres amd64 is around 160MB but i'm pretty sure it will boot with less than that, maybe as little as 64MB | 17:33 |
freem | yes... *after* a clean boot :p | 17:33 |
hyrcanus | maybe there's a program that can scan for bashisms. try #bash maybe | 17:33 |
gnarface | freem: if you're looking for the devuan image build scripts they're on the gitlab | 17:33 |
freem | I will I think | 17:33 |
freem | the problem is, that waste of memory prevent having many VMs running | 17:34 |
freem | basically, at boot, my systems usually require less than 70MiB | 17:34 |
freem | without graphical user session. And I'm large. | 17:34 |
gnarface | heh, well if 250MB is a lot to your vms and the init system is actually blocking that much of it from them, you should consider a different VM solution that can share system ram better like linux-vservers | 17:35 |
freem | I never spoke about init | 17:35 |
freem | I spoke about initrd | 17:35 |
gnarface | well, that doesn't change my advice | 17:35 |
freem | initramfs if you prefer (longer word, easier to differentiate) | 17:35 |
freem | yes, indeed | 17:35 |
Tenkawa | initrd doesnt change your memory footprint at all.. it just moves it | 17:36 |
gnarface | custom initrd isn't a bad idea, either, since probably it is mostly drivers you're not even using | 17:36 |
freem | Tenkawa: at loading time, you need enough RAM to "dezip" it for sure | 17:36 |
freem | what intrigues me is that this use is pretty big | 17:36 |
Tenkawa | and you need to know "much" more of the interdependenies of what you want loaded | 17:36 |
freem | yes | 17:37 |
Tenkawa | freem: what are you "seeing" thats big though? | 17:37 |
freem | that's more or less the kind of stuff I'm trying to learn. I can setup my own init by myself without problems | 17:37 |
freem | wdym? | 17:37 |
Tenkawa | init is "not" a big process | 17:37 |
freem | I did NOT talked about *init* | 17:37 |
Tenkawa | I know.. you keep avoiding our question | 17:38 |
freem | which was? | 17:38 |
Tenkawa | "what is? | 17:38 |
Tenkawa | what do you define as a "big process" | 17:39 |
freem | that depends on the task at hands | 17:39 |
Tenkawa | no... you made an absolute statement... its taking up too much memory | 17:39 |
freem | but having an initramfs 4 times bigger than a working system seems excessive to me, yes | 17:39 |
Tenkawa | whats "its" | 17:39 |
gnarface | well wait, here's how i got confused, you started out talking about the forking tools used to make devuan and then went into how bash was taking up too much memory | 17:40 |
freem | yes | 17:40 |
freem | I don't know how but stuff went to this | 17:40 |
Tenkawa | gnarface: exactly | 17:40 |
hyrcanus | it's a different subject now. | 17:40 |
freem | I only try to get infos on those tools | 17:40 |
freem | originally | 17:40 |
Tenkawa | he made an absolute statement and is avoiding it | 17:40 |
hyrcanus | initramfs being 250MB? | 17:40 |
gnarface | usually the people complaining about bash taking up too much memory are complaining specifically about it during init | 17:40 |
freem | around that value, yes | 17:41 |
freem | I was not talking about init, never | 17:41 |
hyrcanus | ^ clarified then | 17:41 |
freem | or at least about init as PID1 | 17:41 |
gnarface | there's a flag you can set in /etc/ somewhere to make it build default initrd with only drivers for detected hardware | 17:41 |
freem | yes | 17:41 |
freem | but I'm still interested in understanding the why, instead of just working around it | 17:41 |
freem | this involves playing with stuff | 17:42 |
freem | this also involves understanding how stuff are generated, and I am, also, interested in those pre/post inst/rm scripts | 17:42 |
Tenkawa | initramfs size can be easily set..there is no why.. it doesnt "take up that much memory" it only uses up whats allocated during use | 17:42 |
gnarface | well you can try changing /etc/initramfs-tools/initramfs.conf:20 from MODULES=most to MODULES=dep and then rebuild the initramfs | 17:43 |
freem | so, those are the 2 topics I'm mostly interested in: the tooling devuan built, to *play* with it, and the diving into the scripts in packages, to see how much they are really needed | 17:43 |
gnarface | er, the initrd, i mean | 17:43 |
freem | not to ask to anyone to do anything: I don't care, I'm adult, I can work my systems myself | 17:43 |
gnarface | the pre/post scripts are mostly gonna be the same as debian; they handle package installation and removal details | 17:43 |
freem | exact | 17:43 |
freem | they trigger all events at install/removal/upgrade/downgrade of package | 17:44 |
freem | they are the heart of distro | 17:44 |
gnarface | it's good to look into them, you might find mistakes, i doubt you'll find much ram savings though | 17:44 |
freem | well, the packages | 17:44 |
freem | it's more about understanding more my systems | 17:44 |
gnarface | apt-get source [package] | 17:44 |
gnarface | that's a good place to start | 17:45 |
freem | it's ok, I suppose to rant about stuff one have seen in particular situations. It's better to understand why | 17:45 |
gnarface | there should be a devuanized new maintainer's guide in the works somewhere but i'm not sure where | 17:45 |
freem | my intent is to master my systems, not to ask the world to reduce size because it's too big for old systems no one should care about anymore | 17:45 |
gnarface | you can still refer to the debian one for most of this stuff, but really only chapters 6 and 8 are super useful and even then there's only like one page from each that isn't diversionary | 17:46 |
freem | and, well, I thought that people which forked debian may have built a knowledge base | 17:46 |
freem | which would avoid the dive in source, which is costly in time | 17:46 |
gnarface | fsmithered did have a link to one but i forget where and he's not here to ask at the moment | 17:46 |
freem | I know C, I know C++, I know shell, I have basics of perl, I could dive. | 17:46 |
freem | reading doc is still faster :) | 17:46 |
gnarface | it is probably on the forum somewhere... does anyone have that? someone started writing up a condensed devuanized nmg? | 17:47 |
gnarface | out of habit i've still been referring to the debian docs a lot, but i know they're slowly scrubbing it of non-systemd stuff | 17:48 |
gnarface | the basics of how packages work didn't change and most the packages are themselves literally unchanged so any pre-systemd, or non-systemd-specific documentation from debian should still be accurate | 17:49 |
freem | yes, but, well, what I want to understand is not how they are built in itself. Writting the various files (except for those scripts, ok) and running the tools is tedious, but not that hard | 17:52 |
freem | well, at least, to produce a binary package from binaries. Writting a source package is more complicated, clearly | 17:53 |
freem | there was a reason I was not asking directly. I was afraid of not asking stuff correctly... and I was proven true. Guess I should not have joined that early. | 17:54 |
freem | FWIW, I already have built a set of VMs to do automatic installs through PXE + scripts based on debootstrap | 17:55 |
freem | that's even how I noticed that initrd memory increase btw | 17:55 |
gnarface | yea really try that MODULES=dep thing | 17:56 |
gnarface | try that first | 17:56 |
freem | yep, might work | 17:56 |
gnarface | the pre/post scripts aren't magic they're just shell scripts and they're hand-made by every single packager so you just pick a package and read its scripts | 17:56 |
freem | but I was diverted by that RAM question, as I always tend to. The real thing is, I want to understand how installing a kernel will generate an initrd | 17:57 |
freem | actually, from what I remember, they are *not* completely hand-written | 17:57 |
gnarface | one of those package scripts does it but by calling system tools like update-initramfs | 17:58 |
freem | IIRC, in source packages, you can declare triggers, which will partially generate them | 17:58 |
freem | that trigger thingy is rather complex, and I have not dug in it since... long. | 17:58 |
gnarface | uh, there might be some boilerplate available to be generated by the tools yes, i have never used it because mostly i modify other packages | 17:58 |
freem | debian's documentation about packaging is... sorry, but rather messy | 17:59 |
freem | there's also a shitload of tools | 17:59 |
gnarface | yea, skip to chapter 6 | 17:59 |
freem | which are sometimes duplicated | 17:59 |
gnarface | yea, sometimes there's the "debian one" and then there's the one everyone else uses, and usually they're both usable | 17:59 |
gnarface | sometimes debian will have an "old skool" one and a new guard one | 18:00 |
freem | as I said, I'm perfectly able to create a binary archive from a set of files, using the dpkg-deb command | 18:00 |
hyrcanus | i use dpkg-buildpackage and have no idea what i'm doing | 18:00 |
freem | I even wrote some scripts at last work to allow colleagues to deploy stuff with less hassle, while avoiding me to manually check all the shit they'd have wrote | 18:00 |
freem | hyrcanus: I believe `dpkg-deb -b ` is the lowest level you can go. | 18:01 |
hyrcanus | ah | 18:01 |
gnarface | i usually use dpkg-buildpackage for stock packages but i generally use make-kpkg for kernel packages and checkinstall for out-of-distro undebianized sources | 18:01 |
freem | this basically takes a set of files and put them into the "1 version file, 1 archive for DEBIAN, 1 archive for /" end file | 18:02 |
hyrcanus | oh thanks again gnarface - i'll learn checkinstall a bit | 18:02 |
freem | dpkg... that's just the tool on which everything rely, at some point or another | 18:02 |
freem | (and there's only one maintainer for this from what I've read somewhere in the MLs... sad) | 18:02 |
freem | I would not call myself a sysadmin or a maintainer: I'm not. But I do have some basics about how my fav (since more than 10 years) distro works internally | 18:04 |
freem | still trying to increase knowledge though, about the internals, and the why. Not about the tooling itself | 18:04 |
freem | on those, debian, and all derivated distros, are not really nicely documented. It's hard to document those things, of course, especially considering the age of the project | 18:05 |
freem | anyway, I'm not here to give lectures. I probably know less than a lot after all. | 18:06 |
freem | I'll ask more specific questions when I'll have some | 18:06 |
* freem have man pages to read | 18:07 | |
freem | for example, if I asked the question "what creates the `/lib/modules/$version/.fresh-install` files?" I doubt I'd find many to answer that :) | 18:08 |
* freem needs to teach vim to open the scripts in a meaningful order | 18:09 | |
furrymcgee | https://codesearch.debian.net/search?q=fresh-install | 18:11 |
freem | furrymcgee: thanks, but... I just needed to open the scripts in a meaningful order :) | 18:12 |
freem | I was really just looking at them in the wrong order. | 18:13 |
freem | hmm | 18:13 |
freem | https://git.devuan.org/devuan-sdk/libdevuansdk/src/branch/master/docs | 18:13 |
freem | "This documentation is obsolete and should be updated." | 18:14 |
freem | I suppose this is done by compiling something with "python-markdown ruby-ronn" ? | 18:14 |
freem | I'm failing to understand what the SDK is really for tbh | 18:15 |
freem | (when reading the readme. Not sure if there is compiled doc available somewhere?) | 18:16 |
gnarface | if there is it's probably somewhere else on that gitlab | 18:16 |
gnarface | but i think we're behind on documentation for some of this stuff | 18:16 |
freem | doc is always to annoying to maintain... | 18:17 |
gnarface | if you want to look for something productive and educational to do, porting debian docs probably isn't a bad idea | 18:17 |
* freem understands so much | 18:17 | |
freem | tbh, I think debian's doc is not that great | 18:18 |
Tenkawa | freem: Debian's documents are perfectly fine | 18:18 |
freem | ok. | 18:18 |
freem | It's perfectly fine. | 18:18 |
freem | You're absolutely true. | 18:19 |
gnarface | freem: they straight up told me they purposefully wrote the first 5 chapters of the NMG as a barrier to entry. hopefully our documentation won't be that way. | 18:19 |
freem | this is likely why, for so many years, I was using gentoo and then arch's doc for debian stuff | 18:19 |
freem | but yeah, debian's doc perfectly fine. | 18:20 |
gnarface | heh, yea gentoo's docs, now that's actually a good effort | 18:20 |
gnarface | debian's NMG is delivered with maximum contempt for new users | 18:20 |
gnarface | like i said; just skip to chapter 6 | 18:21 |
freem | okay | 18:21 |
furrymcgee | the sdk is to build iso images, I tried this once https://furrymcgee.github.io/Devuan.html | 18:21 |
freem | only to build iso images? | 18:21 |
freem | if that's that... well, I already have a set of dash scripts doing it (and network installs and... not the topic)... | 18:22 |
freem | I guess no need to take time on this then | 18:22 |
* hyrcanus claps with joy! c-only, no dependencies. https://github.com/kevinboone/epub2txt2 | 18:23 | |
Tenkawa | if you also look at the repo its 4 years out of date | 18:23 |
freem | which repo? | 18:24 |
Tenkawa | https://git.devuan.org/devuan-sdk/libdevuansdk/src/branch/master/docs | 18:24 |
freem | I see commit 6 months ago. But right, most stuff is old. Better source? | 18:24 |
freem | the commit to say stuff it outdated is 11 months old btw | 18:25 |
freem | hyrcanus: I'll forward this link to other chans! | 18:27 |
Tenkawa | no that was saying the docs were out of date | 18:27 |
Tenkawa | besides the one bug fix 6 months ago (which looks incomplete btw) its all 4 years since the last change | 18:28 |
* freem have not looked at full history | 18:28 | |
freem | so, any alternative to suggest? | 18:28 |
Tenkawa | I'm looking at the "functionality" | 18:29 |
freem | ACK | 18:29 |
Tenkawa | in chimaera its going to probably need to be changed again | 18:29 |
Tenkawa | (depending on partitioning capabilities) | 18:31 |
razor[m] | hi | 19:00 |
user3614 | During boot I have got a lot of errors "i915 0000:00:02.0: [drm] *ERROR* Fault errors on pipe A: 0x00000080". | 19:00 |
user____ | That's normally a PCI error. Overclocking anything user3614 ? | 19:03 |
razor[m] | as far as i know i915 is intel graphics gfx built into kernel isnt it | 19:03 |
user____ | Yes, but the error is a PCI pipe error | 19:04 |
user____ | Which can occur when things are overclocked in the bios. | 19:04 |
user____ | OR he has a second AGP card and the bios selection is not liked by the kernel, which thinks output is to go to i915 | 19:05 |
user3614 | I haven't overclocked anything, and this is the first time I've seen it. | 19:05 |
user____ | user3614: does it boot properly? | 19:05 |
user3614 | Besides the error messages everything looks OK. | 19:06 |
user____ | https://duckduckgo.com/?t=ffsb&q=%5Bdrm%5D+*ERROR*+Fault+errors+on+pipe+A%3A+0x00000080%22&ia=web | 19:07 |
user____ | user3614: your fault seems to be well known | 19:07 |
user____ | this claims there is a solution https://www.linuxquestions.org/questions/slackware-14/first-boot-was-ok-but-second-boot-ends-with-errors-4175585395/?__cf_chl_jschl_tk__=pmd_Iys..QdEVcwEy7aFHPLB321w2RcYwWpfa3qBB0FYmMU-1634922466-0-gqNtZGzNArujcnBszQY9 | 19:08 |
user____ | "I found the answer: | 19:10 |
user____ | I simply had to disable the onboard graphics (Intel)." | 19:10 |
user____ | What I said above. Do you have an AGP card in addition to the onboard video i915 user3614 ? | 19:10 |
user____ | If so, disable the i915 internal VGA from BIOS. | 19:11 |
user3614 | I don't know, but the graphics card is really Intel. | 19:11 |
user____ | Looks like it's not possible to run 2-card mode with an i915 currently. | 19:11 |
gnarface | if it's one of those optimus laptops things could be more complicated | 19:11 |
user____ | i915 is a chipset internal one normally. Do you have 2 VGA connectors on the back? | 19:11 |
gnarface | (may be 1 vga and 1 hdmi) | 19:12 |
gnarface | (also, there could be an internal nvidia card piggy-backing on the same output) | 19:12 |
user____ | https://github.com/FreeBSDDesktop/kms-drm/issues/39 -- even FreeBSD is affected | 19:12 |
user3614 | I only have a HDMI connector besides the laptop screen. There is nothing conmected there. | 19:12 |
user____ | gnarface: i915 based external / AGP cards are unheard-of. i915 is not exactly stellar speed. | 19:12 |
user____ | Ah, laptop! | 19:13 |
user____ | Ok, on laptops, the switching from internal graphica and agp is more complex. | 19:13 |
gnarface | user____: i915 based external cards are almost unheard-of, but nvidia internal cards are somewhat more common | 19:13 |
user____ | Do you have a nvidia or similar APG? | 19:13 |
user____ | gnarface: i915 is currently present only as "low end" inside chipsets afaik. | 19:13 |
gnarface | user____: yes, but it's also common for users who have bought one of these dual-gpu hybrid laptops to not know | 19:14 |
user____ | Yes. But then lshw could help. Run from boot disk. | 19:14 |
gnarface | yea i'd just run: lspci |grep VGA | 19:14 |
gnarface | user3614: how many lines does this command return? lspci |grep VGA | 19:15 |
user____ | That's after booting. Alas, he can't. So there are boot iso's which include lshw | 19:15 |
user____ | https://wiki.archlinux.org/title/Laptop/MSI this is an Arch example of laptop with i915 + nvidia | 19:15 |
user3614 | It returns two lines. | 19:15 |
gnarface | user3614: lemme guess, one of them has the word NVidia in it? | 19:15 |
user3614 | 00:02.0 VGA compatible controller: Intel Corporation Skylake GT2 [HD Graphics 520] (rev 07) | 19:16 |
user3614 | 00:1f.5 Non-VGA unclassified device: Intel Corporation Device 9d24 (rev 21) | 19:16 |
gnarface | oh, weird... | 19:16 |
user____ | Right, you want to use the Skylake HD | 19:16 |
gnarface | user____: did they ever make hybrid graphics with 2 intel GPUS?? that's what it seems like i'm seeing here | 19:16 |
user____ | gnarface: yes, it's the same story, internal HD and external [other] -- which can also be Intel. | 19:17 |
gnarface | never seen that one before | 19:17 |
debdog | not read the backlog bot sometimes it does not have a "vga" in it: | 19:17 |
debdog | 01:00.0 3D controller: NVIDIA Corporation GK107M [GeForce GT 750M] (rev a1) | 19:17 |
user____ | The graphics card in laptops is just another PCIe card. | 19:17 |
user3614 | I have been using this laptop for several years, and this is the first time I have seen this error message. | 19:17 |
gnarface | hmm, good to note, debdog .... | 19:17 |
user____ | https://forums.linuxmint.com/viewtopic.php?t=355208&p=2055269 user3614 try this | 19:18 |
gnarface | user3614: more likely signs of something now almost working that didn't used to work at all | 19:18 |
user____ | yes that too. | 19:18 |
user____ | Or unseated PCIe connector inside. | 19:18 |
gnarface | heh, yea | 19:19 |
gnarface | but if it's just getting the wrong one you might be able to override it by manually specifying the PCIID in the xorg.conf | 19:19 |
user____ | user3614: in BIOS what selections do you have for VGA? | 19:19 |
gnarface | it usually just loads whatever one is first on the bus which isn't always a safe assumption in this case | 19:19 |
gnarface | yea if you can just disable the weak one in the bios, all the better, but that's actually a rare feature in this class of hardware | 19:20 |
gnarface | or you could just ignore it if everything seems fine :) | 19:20 |
user3614 | In BIOS so far I have only found options for boot order, and viewing logs. | 19:20 |
user____ | No, it is present in 99% of BIOSes. | 19:20 |
user____ | user3614: wow. Phoenix? What brand is this laptop? | 19:20 |
* gnarface maybe has selection bias from sitting in a support channel all day | 19:20 | |
user3614 | This is a Librem 13v2. | 19:21 |
gnarface | user3614: i'd try the other boot order as a test, then maybe do some video benchmarks | 19:21 |
user____ | Interesting beast you have there user3614 | 19:21 |
user____ | So it seems to have Intel HD520 graphics, plus the "legacy" chipset internal i915 | 19:22 |
user____ | The idea would be to tell it to ignore the i915 | 19:22 |
user____ | user3614: you could rename i915.ko in /lib/modules/* | 19:23 |
gnarface | hmm, reviews make it look pretty nice. i'm jealous | 19:24 |
user____ | Also, it seems HD520 graphics require the nonfree? Intel graphics drivers. You do have nonfree enabled user3614 ? | 19:24 |
gnarface | hd520 requires non-free drivers for real? i thought they'd put all their firmware in the main vanilla kernel | 19:25 |
user____ | I am not sure, but from what I am reading, that might be so. Not drivers per se but some blob. | 19:25 |
user3614 | I have only two non-free packages: "firmware-atheros" and "firmware-misc-nonfree". | 19:25 |
user3614 | They are for wifi and bluetooth. | 19:26 |
gnarface | you probably want intel-microcode too for the cpu vulnerability patches | 19:26 |
user____ | try this: sudo updatedb; locate i915.ko; sudo mv /.../i915.ko ~/i915.ko.backup | 19:26 |
user____ | then try to start x, see what happens | 19:26 |
gnarface | ... although maybe not because supposedly those fixes come with a performance hit | 19:26 |
user____ | BEFORE REBOOTING: sudo update-initramfs | 19:27 |
user3614 | X is already running without all this. | 19:27 |
user____ | Looks like i915 is a little buggy | 19:27 |
user____ | Oh, now you tell us user3614 ? !! | 19:27 |
user3614 | I am in this channel on that laptop, so rebooting means I'll be back soon. | 19:29 |
gnarface | i already gathered it was working fine, i was just about helping it work faster and getting rid of that weird error | 19:29 |
user____ | user3614: stop, new instructions, did not know x11 works: https://forums.linuxmint.com/posting.php?mode=quote&f=59&p=1737618 | 19:29 |
gnarface | too late | 19:29 |
user____ | Tant pis. I hope he can handle cli undoing what he did to reboot with gui again... | 19:30 |
user3614 | I've just rebooted and haven't seen the error. I haven't changed anything on the system. | 19:31 |
user____ | :) | 19:32 |
user____ | So now you are running Windows, and it pretends to fix itself by rebooting? | 19:32 |
gnarface | i actually have seen this behavior on systems with both windows and linux in some cases; where in a weird situation the linux drivers will only work after the windows drivers have set them up then you warm-reboot | 19:33 |
user____ | Duck is full of i915 linux driver errors. Looks like it could use "some" attention. Not new messages either, lots are 5-6 yo | 19:33 |
user____ | gnarface: yes that is a known problem. Also seen on sound cards sometimes. | 19:34 |
user3614 | There is only Devuan on this laptop. :) | 19:34 |
gnarface | good to know | 19:37 |
gnarface | and yea that was primarily something i've also seen only with soundcards, particularly on Dells | 19:37 |
gnarface | hey user3614 would you mind pasting the entire output of lspci to paste.debian.net? i'm just curious | 19:38 |
user3614 | https://paste.debian.net/1216431/ | 19:40 |
gnarface | thanks | 19:47 |
gnarface | huh, yea that's a strange thing i've never seen before | 19:48 |
gnarface | cool | 19:48 |
gnarface | i have to assume "non-VGA unclassified device" is secret code for actually a VGA device just not one you can attach to directly | 19:49 |
user3614 | gnarface, perhaps you'll find the Debian wiki page for this laptop useful: https://wiki.debian.org/InstallingDebianOn/Purism/Librem%2013 | 19:55 |
gnarface | hmm, interesting | 19:56 |
gnarface | says only the bluetooth needs non-free stuff | 19:56 |
gnarface | for some reason i think it must need it for wifi too... | 19:57 |
user3614 | I might remember wrongly, but for some reason I needed both non-free firmware packages. | 19:57 |
gnarface | yea my guess is that wifi "works" but not for anything WPA/WPA2 encrypted without the non-free firmware | 19:58 |
gnarface | iirc that was the situation with both atheros and intel's wifi stuff | 19:59 |
gnarface | but maybe not all models | 19:59 |
rwp | I haven't ever seen "Non-VGA unclassified device" before either. That's really surprising. | 20:00 |
rwp | gnarface, I will confirm that I have seen things that needed a windows driver to run first and then the free driver worked okay after. For me it was an audio device. | 20:01 |
user3614 | The website https://linux-hardware.org/?id=pci:8086-9d24 claims it to be a "Skylake-U/Y SPI Controller" found in many laptops. | 20:01 |
user3614 | Well, many Google and Purism machines, and few others. | 20:03 |
user____ | . | 20:09 |
user____ | Looks like daddy Intel took some liberties with the pci interface and id tables. Hey they are so big, they can do whatever Dell has been doing for ages <cough>. | 20:11 |
user____ | theregister calls Intel Chipzilla for a reason. | 20:12 |
user____ | https://assets.amuniversal.com/6df96c4009ba013a7d48005056a9545d | 20:12 |
user____ | <this is ot but funny> | 20:12 |
user____ | golinux: ? ^ :) | 20:12 |
Tenkawa | yes that machine needs atheros non-free firmware | 20:14 |
user____ | re: non free blobs: most embedded things like wlan cards have a proper cpu inside and no (!) flash. The nonfree blob is it's firmware, it needs to be loaded after every boot or cold wake from sleep. | 20:15 |
Tenkawa | wifi will work but bluetooth will not without it | 20:15 |
Tenkawa | if you need some useful reference for that box: https://wiki.debian.org/InstallingDebianOn/Purism/Librem%2013 | 20:18 |
Tenkawa | has some info on the keyboard too if you need the specific mappings | 20:18 |
gnarface | it was already linked above | 20:18 |
Tenkawa | gnarface: ahh missed it | 20:18 |
Tenkawa | those kb mappings are more helpful than anything | 20:19 |
user____ | heh. Who really needs the kbd mappings? | 20:19 |
gnarface | their icons don't seem to distinguish between "works 100% without non-free firmware" and "works with restricted functionality without non-free firmware" | 20:19 |
Tenkawa | user____: non us keyboard users?? | 20:19 |
user____ | gnarface: the bt icon's legend suggests otherwise | 20:19 |
gnarface | i'm saying the wifi is atheros too though and should be not be marked with a thumbs-up | 20:20 |
user____ | Tenkawa: once one has a wm running, it's a wrestling match with that one to just be able to enter shortcuts into any app | 20:20 |
user____ | gnarface: old(er) atheros works fine w/o too much blobbing. ath5 and 6 working here. | 20:20 |
gnarface | hmm, really? noted | 20:21 |
user____ | As I said, there is a cpu which needs fw in the adapter. The blob provides that. | 20:21 |
Tenkawa | user____: the "adapter" needs the fw not the cpu | 20:22 |
user____ | You call it tomaato I call it tomaeto... | 20:22 |
Tenkawa | it does make a difference.. | 20:22 |
Tenkawa | no | 20:22 |
Tenkawa | its a different component | 20:22 |
Tenkawa | a different chip on the motherboard | 20:22 |
Wonka | I think, user____ tried to say "the adapter" has its own CPU which needs firmware uploaded to it. | 20:23 |
Tenkawa | just like some laptop models that are called the exact thing have several revisions of internals | 20:23 |
user____ | https://wikidevi.wi-cat.ru/Qualcomm_Atheros -- see there. There's a chronology, now all are "SoC", just isused as "adapter and phy". Old ones had 2 chips. Same thing. | 20:24 |
golinux | user____: Hehehe . . . | 20:24 |
user3614 | I didn't succeed with the keyboard mapping for Librem13 from the Debian wiki. I have found claims that eudev are ignoring those rules: https://forum.artixlinux.org/index.php/topic,2815.0.html | 21:27 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!