libera/#devuan-dev/ Sunday, 2021-01-31

fsmithredrrq I was using the desktop-live today00:16
fsmithredthen switched over to refracta-10.2 (November 2020)00:16
fsmithredI need to do it again because the last install didn't format the partition00:16
fsmithredso I got refracta+desktop live00:17
fsmithredpretty weird00:17
fsmithredearlier today I tried booting the netinstall iso and it did not work with secure-boot00:17
rrqright. do you know if there's a qemu test option for that?00:18
fsmithredI don't know00:19
rrqI suppose it'll need some variant qemu bios ... and then we'll need someone with secure-boot blessing skills00:23
rrqhmm 2012 article... https://lwn.net/Articles/503820/ .. seems still relevant00:31
rrqhmm the ovmf and ovmf-ia32 packages seem useful here00:44
fsmithredthat'll give you uefi00:46
fsmithreddo they also emulate secure boot?00:46
rrqnot sure .. I just started at https://wiki.debian.org/SecureBoot/VirtualMachine00:56
fsmithredthe list of files in ovmf suggests that it does something with secure boot00:58
fsmithredovmf: /usr/share/OVMF/OVMF_CODE_4M.secboot.fd00:59
fsmithredthere's also one without the 4M00:59
fsmithredbut not all the files listed by apt-file are installed with ovmf. wtf?01:01
fsmithredthose files must be in a later version. Maybe the one in backports01:02
fsmithredI'm getting better results when I use a different usb stick. Refracta install worked. That means the live installer is working.01:07
rrq.. but how do we make installer-isos acceptable for secure-boot h/w ? maybe all those people here interested in that could form a team and make this happen?01:23
fsmithredI thought including the signed grub would be enough01:28
rrqI would have thought it needs both a scrambled code file to execute and a key to first unscramble that code file ... is that what "signed grub" provides?01:45
fsmithredit provides additional files that go in /boot/efi/EFI/<name>01:47
fsmithredname is determined by the distro01:47
fsmithredI just installed again from desktop-live01:48
fsmithredinstaller works, grub appears to work (no error) but doesn't boot01:48
fsmithredI was able to boot from grub command line. Change to ID=debian in /etc/os-release and then grub-install and update-grub, reboot and it works.01:49
fsmithredso I need to get rid of the grub-signed package in the desktop-live01:49
fsmithredthen it'll work like Refracta does01:49
fsmithredI can try a uefi netinstall tomorrow01:50
rrqok... I guess the isos should have their EFI partition look the same way as the installed EFI partition02:03
fsmithredI think the isos use a special name02:06
fsmithredfor removable media bootx64.efi02:06
rrqyes that's what there is currently "/efi/boot/bootx64.efi" as the only thing on the EFI partition02:09
fsmithredthe laptop has it as Boot/bootx64.efi02:11
fsmithredthat's some kind of failsafe bootloader that all uefi should recognize02:12
rrqright; and I think the spec is not case sensitive; probably suggests all uppercase02:14
fsmithredpretty sure I've seen all upper02:14
rrqwell, it's a fat file system02:15
fsmithredok, too cold to sit still here. I'm gonna visit the wood stove.02:15
fsmithred-12 C outside02:15
fsmithredI'm not checking inside02:15
rrq:)02:16
* onefang sends a few degrees to fsmithred.02:18
onefangrrq could probably spare a few as well.02:18
bb|hcbfsmithred: -12 out is fine, -12 inside would be bad02:22
golinuxEven I could spare a few today  . . .02:24
* Xenguy offers fsmithred the 3rd degree02:24
bb|hcbi only know the legacy boot process, no experience with efi and secure boot :( but can burn a usb and test on some laptop?02:27
rrqhmm uefi spec is 2557 pages ...02:27
rrqbut "secure boot" section is just 33 pages02:36
fsmithredI'll play with the netinstall on uefi hardware tomorrow and see what it does02:38
fsmithredbb|hcb, having more people test is always a good thing02:41
fsmithredThis is probably better than the 2557 pages: https://www.rodsbooks.com/efi-bootloaders/index.html02:42
fsmithredthe specs only work if the manufacturers stick to them02:43
bb|hcbfsmithred: sure, drop me a note when it runs ok in lab, i have couple of machines within hand reach...02:46
rrq(mmm spec pages 1723-1729 seem to be the critical bit)02:47
masonRod's Books is a fantastic resource.04:06
rrqyeah.. had lunch and lost interest :)04:08
fsmithredemdete, no, the bug in change-username regarding config files with spaces in the names is not known. Or wasn't known until just now when I saw your post.15:37
fsmithredI didn't know there were config files with spaces in the names.15:38
fsmithredis $i just picking up the first word? If so, I'll have to figure out how to get quotes around the whole file name. Oh, maybe this is because the shell handles spaces in filenames differently.15:40
fsmithredIf you have more info about what's happening or suggestions to fix it, please ping me. Thanks.15:41
emdetefsmithred: as far as i remember chromium had those (spaces in filenames)21:59
emdetefsmithred: will check that again - the passwd geckoes code was something to improve as well21:59
fsmithredwhat is geckoes?22:00
masonGECOS?22:01
fsmithredI'm picturing lizards22:01
fsmithredand expecting that to be wrong22:01
masonI like to picture lizards. I'll join you.22:01
masonBut once we're done, https://en.wikipedia.org/wiki/Gecos_field22:02
masonBut not... just... yet: https://duckduckgo.com/?t=ffab&q=iguana&atb=v237-1&iax=images&ia=images22:03
mason(No offense, but iguanas have more dignity and gravitas.)22:04
fsmithredok, chfn noted22:04
onefangThe local water dragons might disagree.22:05
* onefang wanders off.22:05
fsmithredemdete, yeah in .config/chromium there are files and directories with spaces in their names.22:08
fsmithredand I figured out how to deal with the spaces22:19
fsmithredmaybe. Haven't tested in situ yet.22:22

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