libera/#devuan/ Monday, 2021-03-15

nemoBit of a mystery with a home devuan machine04:29
nemoit consists of a Dell XPS with 2 HDs - a spinning rust setup as /home, and a built-in SSD setup for UEFI boot04:30
nemoIt had been running devuan for several years, but recently the spinning rust /home was failing, causing an increasing number of write errors and confusing a bunch of apps as a result04:30
nemoI finally got a replacement authorise, and as I rebooted to do a dd if=/fail of=/new  I was shocked by a GRUB> prompt04:31
nemoI guessed that maybe some grub update got interrupted due to file system errors on the /home mount (odd but not impossible I guess) so I ignored it and went on with the copy04:31
nemoonce /home was rescued and clean (which took about a dozen tries due to having to skip past sections - I blame the purchase of a Seagate)04:32
nemoI try rebooting. GRUB> prompt.04:32
nemoFine.  I grab Devuan Bewulf live image (fail right there 'cause it can't even detect the UEFI SSD)...04:33
nemoGrab Devuan Beowulf netinst instead, which I guess is what I should have done first time around.  Select rescue, go to the disc04:33
nemoeverything seems fine04:33
nemo / mounted on /dev/nvme01p7   and /boot/efi  on p1.   try grub update, grub install. everything looks normal. config seems fine04:34
nemoreboot. GRUB> prompt04:34
nemook... maybe... a kernel update got interrupted... somehow?  apt install --reinstall linux-image-amd64... reboot. GRUB>04:34
nemoummm grub config file corruption?  apt install --reinstall  every bloody grub-efi package04:35
nemoreboot. GRUB>04:35
nemoso... here's where it gets weird04:35
nemoI do:04:35
nemols (hd3,gpt1)/efi/devuan/grub.cfg04:35
nemocat (hd3,gpt1)/efi/devuan/grub.cfg04:36
nemothe grub.cfg has 3 lines.04:36
nemosearches for the UUID that corresponds to p704:36
nemosets it to root04:36
nemothen loads that grub.cfg04:36
nemoI retype, manually, every line in that grub.cfg that I just catted04:36
nemoand my Devuan grub splash pops up.  And boot proceeds04:37
nemonow I would rather like to not have to type out those 3 lines every time I boot this machine. not to mention kinda weird this happened in the first place04:37
nemooh. I almost forgot. important04:37
nemoI tried:04:37
gnarfacemake sure there's a newline at the end of the file04:38
nemoconfigfile (hd3,gpt1)/efi/devuan/grub.cfg04:38
nemognarface: kk one moment04:38
* rrq would look for/at (hd3,gpt1)/efi/debian/grub.cfg as well04:38
nemorrq: you're suggesting I copy it there?04:39
rrqyes... that's the one the uei boot is using04:39
nemognarface: 0a04:39
nemognarface: does it need dos format?04:39
gnarfacei don't think so04:39
rrqthe uefi boot image has the hardcoded and is then signed04:40
nemokaaaaay04:40
nemohm. don't think I have secure boot enabled04:40
nemoalso devuan shows up fine in the uefi boot menu04:40
nemobut can't hurt to try I suppose04:40
nemowould like to note this system has been working for a long time. I kind of assume this is related to the failing spinning rust. but now I'm not sure04:40
rrqyes, hte boot program is in /efi/devuan so that gives the name, but the program looks for /efi/debian/grub.cfg04:41
nemorrq: why did configfile (hd3,gpt1)/efi/devuan/grub.cfg   fail?04:41
nemoit was like grub wasn't reading it04:41
rrqexactly.04:41
nemobut I mean. I gave an explicit path04:41
nemorrq: is this related to the boothole vulnerability fixes?04:42
nemofrom last fall?04:42
rrqthe boot program whatever.efi has /efi/debian/grub.cfg hard coded, and that program is signed04:42
nemokk04:42
nemo/boot/efi/EFI/devuan# strings *.efi | grep debian04:43
nemo/EFI/debian04:43
nemolooks like you are exactly correct04:43
nemorrq: sooo... how'd I get a devuan in there. from an earlier installer?04:43
nemoI guess I need to fix that across the board04:43
nemorrq: and. maybe on other devuan machines too..04:43
nemolet's see what the one I'm typing on here has04:44
rrqcopy from /efi/devuan ...04:44
nemorrq: I will I will! you clearly know what you're talking about.  I just don't want to have to do manual stuff every grub update04:44
nemoand right now grub seems to write to EFI/devuan04:44
rrqI thought it was so generic it wouldn't ever change ... but that was wrong apparently04:44
Xenguy.oO( It's sometimes not easy being on the fringe )04:45
rrqthe alternative would have been to fork enough to make a proper devuan boot loader04:46
nemoXenguy: heh. at least my new system didn't bootloop due to RDRAND systemd bug ☺04:46
nemorrq: so... right now devuan grub seems to write to efi/devuan04:46
Xenguybeen under a rock04:46
nemorrq: so I need to manually perform this copy after every upgrade?04:47
nemorrq: so is this related to the boothole fixes? and does this break every Devuan UEFI out there?04:47
nemorrq: sounds like it is something brand new and completely unrelated to my HD woes, just bad timing for me that I have 2 issues at once ☺04:48
nemohttps://www.debian.org/security/2020-GRUB-UEFI-SecureBoot/  that is04:48
nemolike they hardcoded the config path to prevent alternates04:48
nemoah. good. machine I'm typing on right now isn't even using UEFI so no concerns there04:49
nemocouldn't remember at all what I'd setup04:49
rrqyes; grub is clever enough to install under EFI/devuan, but it installs the debian binary which refers to EFI/debian ... all on a FAT file system so linking is not possible04:50
nemorrq: heh. I'd mentally ruled out linking for that reason04:50
nemogah04:50
nemoso annoying04:50
nemorrq: so... this is kinda bad then. I'm not the only one hitting this. just the first to notice/complain04:51
nemothankfully that grub.cfg is crazy-generic, so I guess I probably don't need to run a cp every update at least04:51
rrqyes, that grub.cfg looked very generic so wouldn't need to change ..04:52
nemowelp... thanks for your help. time for me to fix every UEFI devuan I have04:53
nemogood luck with a proper solution04:53
nemoa grub shim to do the copy? ☺04:53
rrqfelt like too much work to me :) learning how efi signing works04:54
nemorrq: hope you don't need like some cross-sign from microsoft ☹04:55
rrqI understood it to be two levels: first signing of the programs (always), and then an application level verification signing, the latter being the secure boot04:56
rrqand that latter can be include time bomb as well04:57
rrqI suppose we should work out a fix-it script to add to update-grub's post-processing scripts05:01
nemodefinitely sounds like fastest and easiest05:02
nemoand fewer people trying fruitless rescue boots the better05:02
nemoI mean, only sheer bullheadeness made me try manually typing in contents of configfile after attempting the configfile command05:03
DashiePiehi nemo05:39
nemorrq: any news on "fixits" ? was going to extend my mitigation to warning coworker who also uses devuan17:03
fsmithrednemo, did you try removing grub-efi-amd64-signed? Just keep the unsigned version. That alone will probably fix it. (It did for me in two cases)17:25
nemofsmithred: oh. I had not tried that yet.17:27
nemofsmithred: so this *is* related to the boothole mitigations?17:27
fsmithredthe what?17:27
nemo23:42 < nemo> rrq: is this related to the boothole vulnerability fixes?17:28
nemo23:48 < nemo> https://www.debian.org/security/2020-GRUB-UEFI-SecureBoot/  that is17:28
fsmithredI'm talking about the new version of grub thta does not boot17:28
nemoyes. I know17:28
nemoI was curious as to why they had done this17:28
nemoI'm trying your suggestion now17:28
fsmithredI'm curious as to how it affect those who claim they are using grub-pc17:28
nemofsmithred: after removing that package still see:17:29
nemo# strings grubx64.efi | grep debian17:29
nemo/EFI/debian17:29
nemoI'll see if that actually means anything after reboot though17:30
nemofsmithred: should I have done a grub update too?17:30
fsmithredmaybe17:30
fsmithredor dpkg-reconfigure grub17:30
nemofsmithred: my non-UEFI machine does not seem to have any issues booting though17:30
fsmithredsame here with legacy bios.17:30
fsmithredseems fine17:30
nemono change in that "strings" after dpkg-reconfigure grub-efi17:31
nemogonna try an update-grub17:31
fsmithredmaybe grub-install first17:31
nemofsmithred: that did it17:31
nemofsmithred: strings is now clean17:31
fsmithredcool17:31
fsmithredthat won't work if you use secure boot17:32
nemofsmithred: ok. so I have a nicer mitigation than "manually copy something over" to recommend to people17:32
nemofsmithred: yeah, not doing that anyway17:32
nemofsmithred: this machine is inside my house, and if anyone breaks in to haxor it I have far far far bigger problems17:32
fsmithredyeah, I understand17:32
fsmithredI have a sword for such events17:33
nemohee17:33
nemofsmithred: I never did progress very far in iaido or in tameshigiri practice.17:33
nemojust enough to probably not slice my own fingers off17:34
fsmithredgood start17:34
fsmithredtry three-section staff. There's no way to learn it without clobbering yourself.17:34
fsmithredI sure hope we don't have to fork grub17:35
nemoI've done a fair amount of naginata. and a few lessons of bojutsu 'cause my old kendo sensei did that class after kendo17:36
nemothat's limit of my stick stuff17:36
nemofsmithred: I wonder what other debian derivatives are doing17:36
nemofsmithred: you guys might want to at least do a blog announcement on the 2 mitigations for people who want to get ahead of this17:37
fsmithredwe have a blog?17:37
nemoummm site announce page. dunno what you call that 😝17:37
nemo# strings grubx64.efi | grep debian17:38
nemo/EFI/debian17:38
nemooups17:38
nemomkdir /boot/efi/EFI/debian;cp /boot/efi/EFI/devuan/grub.cfg /boot/efi/EFI/debian/  | apt remove grub-efi-amd64-signed;grub-install;update-grub17:38
nemohttps://mailinglists.dyne.org/cgi-bin/mailman/listinfo/devuan-announce17:39
fsmithredoh, I didn't even bother with the debian dir17:39
* fsmithred is still looking up japanese words17:39
nemofsmithred: well. first one was recommended by rrq and would work with secure boot presumably17:39
nemo2nd one would be cleaner but would not17:39
nemoand I guess a quick fix would be you guys hacking in a script modification to do the first one on every grub update17:40
nemohttps://dev1galaxy.org/viewforum.php?id=12  here too17:40
nemo(places to warn people)17:40
nemoAPic: you're on every channel I'm on?17:47
APicCould be17:48
nemo😝17:48
APic17:48
tuxd3vdoes any one knows the package name for the ascii cam?18:17
tuxd3vthe One jaromil were working on :)18:18
tuxd3vhasciicam18:19
tuxd3v:)18:19
tuxd3vdoes anyonhe managed to have it working?18:25
tuxd3vI am getting:18:25
tuxd3vhttps://paste.debian.net/1189503/18:26
nemofsmithred: https://old.reddit.com/r/debian/comments/lkeogh/devuan_31_released_debian_fork_now_offers_runit/gnxzbq9/18:31
nemofsmithred: is emorrp1 correct? he's insisting I shouldn't be bothering with devuan18:31
nemofor me you guys were protecting me from random package breakages18:31
nemolike tomcat18:31
fsmithredwho what?18:31
nemofellow in that reddit comment18:32
nemohttps://old.reddit.com/r/debian/comments/lkeogh/devuan_31_released_debian_fork_now_offers_runit/gnje12e/18:32
* fsmithred stops reading about japanese female warriors18:32
nemohe's in #libregamenight telling me that Devuan offers "nothing new" that they have "given up on a fork" and all developers focus on "contributing to upstream Debian"18:32
nemoand. I'm open to this possibility18:32
nemoforks are tiresome18:32
nemoI just don't want to get screwed on a key update18:32
fsmithredwtf? We have not given up.18:32
fsmithredWe have pushed some changes upstream.18:33
nemofsmithred: well. lemme dump history of the chat18:33
fsmithredWe believe everyone will benefit if debian and devuan cooperate with each other18:33
nemoI may be misinterpreting him18:33
nemomy understanding was always that Devuan was focused on trying to stay close to upstream debian in terms of resources and maintainability18:33
fsmithredour goal is to provide debian without systemd18:34
fsmithredso yeah, we try not to change too much18:34
fsmithredwhat does "nothing new" mean?18:35
nemohttps://m8y.org/tmp/emorrp1.txt18:36
nemofsmithred: I could simply be arguing from ignorance here18:36
nemobut I'm open to using upstream debian as he pushes here and in his reddit posts. but at present it seems risky18:36
nemow/ chance of random strips that Devuan at least tries to protect me from18:37
nemobut since he was speaking for Devuan devs, I figured I'd ask you whether he was blowing smoke, since he's not here.18:37
Tenkawaimho. if an app/game needs to be infrastructure centric... there are bigger issues18:37
Tenkawainfrastructure being systemd vs sysvinit18:37
Tenkawain this case18:38
nemothere was no reason to delete the tomcat script18:38
nemoof all the apps under debian, tomcat, as basically java VM, was the most-standalone18:38
nemoserver apps18:38
nemobut apparently, doing so does not violate the debian vote according to him18:38
nemoso what I don't see is... if a functioning script in a major package can be deleted for no reason, with no repercussions, and get pushed to stable.18:39
nemothen in what way does the vote mean anything18:39
nemoand in what way can I count on Debian's sysv guarantees18:39
Tenkawanemo: which group are we talking that wants to "vote"?18:39
nemoTenkawa: the Debian vote on retaining support for sysv18:40
nemoTenkawa: the one heavily quoted by him in the chat above18:40
fsmithredlast Technical Committe vote18:40
Tenkawathat's just the way they are moving18:40
fsmithredthe one with five choices to drop support for sysvinit and one or two to keep it18:40
TenkawaI disagree with it myself18:40
fsmithredmy numbers may be off18:40
nemoTenkawa: that's fine... what I'm trying to understand is emorrp1 seems to be simultaneously arguing that deleting sysv from packages is fine, and does not violate the vote18:40
nemoand also simultaneously, that Debian is actively supporting sysv, and Devuan serves no purpose18:41
nemoas a user, I'd be happier with a unified Debian. it's easier for me, no randomly breaking grub for example18:41
fsmithredanyway, as I understand it, you can now have debian without systemd and without devuan, because of elogind, which debian got from us.18:41
nemobut... I don't see how I could trust it not to screw me tomorrow18:41
Tenkawait serves no purpose "to them"18:41
nemofsmithred: yeah, that's nice of them, (and of you) but if they have no guarantess of not pushing breakage to stable...18:41
fsmithrednemo, we are not closing up shop yet.18:41
nemogood!18:42
nemo'cause I would be very nervous every time I ran apt update on debian18:42
fsmithredthey need us18:42
fsmithredto be a thorn in their side18:42
nemoAre his comments largely correct though? in the 2 reddit posts linked? (top comments, short to read)18:42
nemofsmithred: 😃18:42
fsmithredI did not read all his comments18:42
Tenkawanemo: let me read it for a sec18:43
TenkawaI havent read it yet18:43
fsmithredbut he does not speak for any devuan devs that I know18:43
fsmithredand I think I know them all18:43
nemoTenkawa: skimming the discussion in #libregamenight would be appreciated too18:43
nemoTenkawa: I realise it is a frivolous game channel. but this all shocked me. and he sounds like a debian-ite18:43
nemohttps://old.reddit.com/r/debian/comments/lkeogh/devuan_31_released_debian_fork_now_offers_runit/gnxzbq9/ https://old.reddit.com/r/debian/comments/lkeogh/devuan_31_released_debian_fork_now_offers_runit/gnje12e/ https://m8y.org/tmp/emorrp1.txt18:43
nemothe 3 links rejoined18:43
Tenkawayou are talking about emorrp1 right?18:44
Tenkawahis opinions?18:45
fsmithred13:08 < emorrp1> what do you mean? Devuan is basically defunct, the work is being done inside debian proper these days18:45
Tenkawayeah18:45
Tenkawahe's just ranting18:45
fsmithred^^^ absolutley false statement I just posted18:46
nemoyes. that kind of thing. I was shocked, but also trying to be open to the possibility that I should investigate using Debian directly18:46
nemoI just didn't trust them18:46
fsmithredI wish we did not have to keep forking packages18:46
fsmithredI wish we really did not have to exist, but in fact we do.18:46
fsmithredanyway, this is not a support discussion18:47
Tenkawanod18:47
nemofsmithred: well. it was18:47
nemofsmithred: I was honestly wondering if Debian was safe to switch back to, and if his set of scripts for "making Debian work like Devuan" was an accurate reflection18:47
nemoI was not trying to start an argument, just wondering if he was right, and it was similar/same18:48
fsmithredum...18:48
DashiePiehi nemo18:48
nemohi DashiePie18:48
nemofsmithred: if so, I was literally going to try switching one of my sets of apt sources back on a non-critical devuan to initiate process18:48
fsmithredif he's not rebuilding packages to remove dependency on systemd, then I guess he has systemd installed, like mxlinux has18:48
nemofsmithred: ok. that sounds rather useless18:48
fsmithreddpkg -l |grep devuan18:49
fsmithredwill give you an idea of what packages we forked18:49
Tenkawarather dangerous to try to go backwards too18:49
nemofsmithred: that is quite a lot. but his claim was it had all been reintegrated upstream ☺18:50
nemothat was the surprising bit to me. so I was like... huh... maybe I should re-examine the... 10 machines I currently have on devuan18:50
fsmithredno it hasn't. We still have to fork packages.18:50
fsmithredsome debian packages now require systemd|elogind18:51
fsmithredbut not all18:52
nemos/10/11/18:52
nemofsmithred: ok. good to know. IRC is a place of random smoke blowing, but he sounded savvy18:52
nemoshall not worry about it18:52
fsmithredsimilar discussion happened on the forum recently18:52
fsmithredhttps://dev1galaxy.org/viewtopic.php?id=414518:53
nemofsmithred: ah. links to the reddit thread above. which was mostly him18:54
nemothanks. will read through18:54
tuxd3vany one knows about a camera application that can do more than 640x480?19:13
tuxd3vcheese limits me to 640x48019:14
nemoO_o19:14
nemoseriously?19:14
Tenkawastop getting that fake stuff19:14
Tenkawalol19:14
tuxd3vwell the box advertize FullHD 1080P @30fps19:15
Tenkawabut hmm... I wonder if you dont have rights to the right /dev/*19:15
tuxd3vboth are in the video group19:15
Tenkawatuxd3v: do you have the right access to /dev/video*19:15
TenkawaI know that can burn you19:15
tuxd3vat least to /dev/video0 yes19:16
Tenkawawell darn19:16
tuxd3vbut to video1 I really don't know19:16
tuxd3vdoes you guys get bigger resolutions on cheese?19:18
tuxd3vthe max to choose of cheese for me is 640x480 puff19:24
tuxd3v2 years waiting for a webcam and shows up this s*t19:25
nemoumm19:29
nemofine. hang on19:29
nemolet me install cheese19:29
nemotbh hadn't really needed anything but firefox for video up until now19:29
fsmithredtuxd3v, will vlc do what you want?19:42
tuxd3vfsmithred, good idea :)19:42
tuxd3vwith cheese is to forget :(19:42
tuxd3vthis is the information I could get from it till now: https://paste.debian.net/1189519/19:47
nemotuxd3v: I'm getting a LOT more than 640x480 in cheese19:54
nemoone moment19:54
nemotuxd3v: 1280x72019:55
tuxd3vho my god..19:56
tuxd3vso its a fault from my webcam..damm19:56
nemo*shrug*19:56
nemoor how it exposes itself and how cheese is parsing that19:56
nemoor like webcam having low res vid and high res static or...19:56
nemodunno19:56
nemoBus 001 Device 006: ID 1b3f:1167 Generalplus Technology Inc. WEB CAM19:57
nemowhat I'm using19:57
nemojust some cheapo knockoff thing. seems to work well... once or twice a month it screws up gamma and I have to unplug and replug USB19:57
nemobut apart from that19:57
tuxd3vwait... wait... smoke is on the horizon19:58
nemomic and vid seem fine19:58
tuxd3vdo you see now there a 1280*1080 ? https://paste.debian.net/1189520/19:59
tuxd3vlets test19:59
tuxd3v:)19:59
nemooh look. the matricites are back20:00
DPADoes anyone know what could cause apt-get upgrade and apt-get dist-upgrade to disagree like this: https://pastebin.com/HRryueHC20:04
buZztuxd3v: mplayer -vf screenshot tv://20:04
buZz^ can do any resolution20:04
buZzpress S for a photo20:04
masonDPA: Could you use bpaste.net or termbin.com or something that doesn't want quite so much javascript?20:05
DPAhttps://bpa.st/DKTA20:06
masonDPA: Interesting. I assume you've run update again once or twice to rule out a race condition on update...?20:07
DPAYes20:08
masonDPA: Can you share apt info meson | grep Version | nc termbin.com 9999    ?20:09
masonand what version is installed now?20:09
DPAVersion: 0.56.1-1~bpo10+120:09
masonOh... Backports.20:10
masonDPA: Did you explicitly pull that from backports? Anyway, from there I'd see what uninstalling it would try to drag with it, and reinstall with newer stuff if you explicitly want it.20:11
DPAYes, I think I installed it when I needed a newer meson version to build mesa. I probably don't need the newer version anymore now though.20:12
masonIt'll tell you.20:14
DPAOh, now I see, it wanted a newer ninja-build version. Looks like new bpo meson would now need bpo ninja-build, and I had regular ninja-build installed.20:17
DPAI now removed bpo meson and installed regular meson again for now. I could probably install from bpo again too, it should figure the deps out if I tell it to take stuff from backports, but I don't need it right now anyway.20:19
masonYou're better off using the smallest possible set from backports generally.20:20
DPAI tend to need some quiet low level stuff from there. For example, qemu-user-static, I can't bootstrap chimaera or bullseye with the regular one.20:25
DPAFor chimaera, I even had to take debootstrap directly from chimaera because backports was too old. And mkfs.f2fs from f2fs-tools was missing some options I needed.20:25
DPAThere are probably some other things I don't remember right now too.20:25
DPAMaybe I should just switch to chimaera, it runs well on some other systems I have. (But those aren't regular desktops)20:27
buZzare they S/390 machines?20:29
DPANo, two arm64 machines (a smartphone and a devkit), and some lxc containers on x86_64.20:30
buZzah, pretty regular stuff20:30
buZztoo bad20:30
buZzi was hoping for something exotic20:31

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!