[00:00] Oh O.K. [00:00] anyone know by chance which version of gcc is default on 21.04? [00:00] is it 9 or 10? [00:00] Right now the store made 2 partitions: OS partition with 200 GB & 265 GB for my files, I'm not happy with this, I'd like one large partition. [00:01] Thete: 10 [00:01] ty [00:02] how is ubuntu latest vs LTS? do you experience stability issues? [00:02] Theyou can check this for yourself on https://packages.ubuntu.com or with rmadison / dak ls [00:02] IRCMonkey: My system as an example of a tight working install: https://termbin.com/sp48 . [00:02] Thete: ^ [00:03] is there a flag for `apt get` that will auto-select `yes`? [00:03] ah, -y [00:03] nvm [00:03] Back -looking- [00:04] Thete: https://people.canonical.com/~ubuntu-archive/madison.cgi [00:04] I saw that, but for some reason my system was using gcc-9 by default [00:04] so I'm guessing I messed something up [00:04] IRCMonkey: My partitions named: https://termbin.com/0ykq [00:04] which is why I was confused === Buggys- is now known as Buggys [00:05] https://people.canonical.com/~ubuntu-archive/madison.cgi?package=gcc&s=hirsute says 10.3.0 [00:05] Oh c00 [00:05] @ Bashing-om [00:08] I named my C:\ "Op Systems" & D:\ "Data" [00:09] IRCMonkey: My booted system's disk usage: https://termbin.com/jc42 . As you can see the operating system does not need much room. [00:10] Brilliant ! [00:10] IRCMonkey: Ouch - in linux a space is seeen as a file delimiter - you will have a learning curve to deal with any spaces. [00:11] Do I select "Third-party" drivers during install for it to automatically detect & install Intel HD Graphics\ WiFi or will I have to provide the driver with the disk for it ? [00:12] yup [00:12] IRCMonkey: you need neither [00:12] IRCMonkey: Neither - Intell drivers are in the kernel :) [00:12] File names cannot have spaces ?? [00:13] Oh good [00:13] If you dont need them they wontbe installed [00:13] never have spaces in file names!!!! [00:13] OMG ! [00:13] But I want my screen resolution to be 1366x768 [00:14] What do I do with all my files in my USB sticks with spaces ? [00:14] I installed Ubuntu 20.4 hoping for a miracle and no... Same issue on first boot. I'm gonna give up for today, I'll try again tomorrow. [00:14] Thanks everyone for your help. [00:15] B4 you install any linux...run a live version to see if everything works...then install... [00:15] The live install works perfectly fine [00:15] jacke: that doesn't always give you a good idea of how well it's going to run [00:15] this is a video card issue [00:16] IRCMonkey: Spaces: not what I implied - just that a space in a fike name takes extra care to cope with. Op Systems straight up will be seen as 2 files: "Op" and "Systems" also linux is case sensitive - op and Op are seen as 2 different files. [00:16] then you can make the install work just as well [00:16] I think.. [00:16] Guest39: :-/ sorry we couldn't help. one more thing that may be the diffference between live and install may be microcode updates. [00:16] Ohhh O.K. [00:16] No I don't plan on naming my drives or anything. [00:16] tomreyn: and hwe kernel and no running from squashfs and video drivers, etc [00:16] But filenames can contain spaces amirite ? [00:17] jacke Probably yeah but I've been trying for the last ~6 hours with the help of people here and nothing worked so far so I'm done for the day :) [00:17] leftyfb: 21.04 has no hwe kernel by default [00:17] tomreyn: " I installed Ubuntu 20.4 hoping for a miracle and no" [00:17] Guest39: similar to "nomodeset" you can use "dis_ucode_ldr" to override the microcode loading [00:17] leftyfb: yes, but it was all about 21.04 before this. [00:18] ok, yeah, just joining now [00:18] IRCMonkey: Yes file names "can" contain spaces but in use these spaces will have to be escaped so the system sees them as one file. A pain sometimes to keep track of. [00:18] tomreyn No worries, really appreciate your inputs. Alright last test and I'm done for today ^^ [00:19] Guest39: you're welcome [00:20] Oh O.K. [00:20] Guest39: yet another suggestion: this computer is old enough that it was probably shipped (and primarily tested) with classic bios booting by default. you could try switching to that (but need to reinstall then). [00:23] The first install was with the BIOS option enable in BIOS. But I guess I also need to change the way I create my booting drive right? [00:24] if you make the bios boot in 'legacy' bios mode (rather than UEFI), then do a standard installation, you'll be set [00:24] but i'm not sure i got the question right [00:26] Right now I'm using ventoy for my bootable drive using GPT partition. Do I need to create a more "old school" installation drive in addition to enabling Legacy BIOS? [00:28] ventoy ... never heard of that before [00:28] i don't know for sure whether ventoy will be bale to boot in both uefi and bios mode, and whether it will be able to boot the ubuntu installer in either mode. [00:28] bale -> able [00:29] I've got tons of usb drive I can just create another one the old school way [00:29] Secure Boot was supported from Ventoy 1.0.07, but the solution is not perfect enough. So from ventoy 1.0.09, an option for secure boot is added in Ventoy2Disk.exe/Ventoy2Disk.sh and default is disabled. So by default, you need to disabled secure boot in BIOS before boot Ventoy in UEFI mode. [00:29] hmmm, questionable tool it is.. [00:29] https://www.ventoy.net/en/doc_secure.html [00:30] Guest39: Tje mode that you boot the installer in is the mode that will be used to install; Bios or UEFI. [00:30] use mkusb, rufus .. [00:30] GPT for UEFI boot in bios [00:30] turn off secure boot [00:31] secureboot is not a problem, fastboot is. [00:31] I've used rufus before. Ventoy is more convenient, I can just drop a collection of ISOs on a key and install whatever I need [00:31] secureboot is disabled. Fastboot is enable, I'm not sure how that could cause the issue I'm seeing [00:31] secureboot can be aproblemif he's booting an unsigned kernel [00:31] Guest39, interesting, but you are here for hours now.. [00:32] omg .. fastboot enabled.. [00:33] take some time to read the UEFI manual [00:33] !uefi [00:33] UEFI is a specification that defines a software interface between an operating system and platform firmware. It is meant as a replacement for the BIOS. For information on how to set up and install Ubuntu and its derivatives on UEFI machines please read https://help.ubuntu.com/community/UEFI [00:35] It just says to disable it without any explanation why. I just disabled it and no change. [00:36] we don't know either about that tool, it is not in our recommended list [00:37] https://ubuntu.com/tutorials/create-a-usb-stick-on-windows [00:38] tomreyn dis_ucode_ldr didn't help :( [00:39] Guest39: a pity. :-/ have a good night, more luck tomorrow! [00:39] oerheks I mean at this point I'm willing to try anything but I don't see how just changing the tool I use to setup my install drive going to change anything. It works perfectly fine on live image, issue appears after install. [00:40] tomreyn Thanks again, see you [00:40] O.K. that'll be all thanks Bashing-om, tomreyn & jushur for answering my woes. [00:45] Guest39, good, try rufus and let us know. [00:48] this is my first time using an irc [00:49] * jacke slaps myself with a trout.. [00:49] * jacke slaps himself with a trout.. [00:50] /etc/apt/trusted.gpg [00:50] me_: Hi! [00:51] me_: dont paste things that can be used against you here! [00:51] ...even just a path [00:51] ok [00:56] you have the right to remain pasteless [01:01] You have the right to remain pasteless. Anything you paste can be used against you in court. You have the right to talk to a linux user for advice before we ask you any questions. You have the right to have a linux user with you during questioning. If you cannot afford a inux user, one will be appointed for you before any questioning if you wish. If you decide to answer questions now without a linux user present, you have the right to stop answering alt any [01:01] time. [01:08] jacke: please take it to #ubuntu-offtopic. This is a support channel === blackest_mamba_ is now known as blackest_mamba [01:22] oerheks Just used the ubuntu recommended way to create a install stick using Startup Disk Creator on Ubuntu (don't have access to a windows machine right now) and same issue.. [01:26] I have several unity-scope packages preventing me from upgrading gir1.2-unity. I can safely remove them, right? [01:27] Guest39, did you validate your ISO? https://tutorials.ubuntu.com/tutorial/tutorial-how-to-verify-ubuntu#0 [01:32] maxh: If ' apt show ' shows as "optional"; then yes can safely be removed. [01:34] guiverc just did all good [01:36] I'd assume bad write to media then.. I find about 5-8% of writes are bad (usb-flash-media is made to cost and not quality), so I'd try booting it on other box; if it doesn't complete media validation there - likely a bad write. [01:36] Guest39, ^ [01:40] guiverc Thanks for your input but this not my issue at all. I'm created 2 different install drive with 2 different method, the issue is the same. Live image is all good. Installed ubuntu "works" I just have display issue probably because of the Nvidia card. [01:41] In any case I said I would stop troubleshooting for tonight, Just wanted to confirm that my way to create bootable drive had no issue.. [01:41] I didn't see your prior posts, your message I did see mentioned "used the ubuntu recommended way to create a install stick using Startup Disk Creator on Ubuntu... same issue" so my comment was based on that. [01:43] got it my bad if I sounded snapy [01:43] not a problem Guest39 === jacke is now known as Guest4766 === M4he is now known as mahe [02:03] Hello! I've used "sudo dpkg --verify" to check if everything is fine and I got this output: https://paste.ubuntu.com/p/XTcNP2FNSC/ - what does this mean? those files were changed? [02:03] as far as I know, dpkg --verify verifies the installed apps and other stuff right? === cacti is now known as ice === jacke is now known as Guest4478 === Guest4478 is now known as jacke [02:58] I just discovered focuswriter, and it looks great, but I want a dead simple kid friendly write words over scanned background page for a kid with disgraphia. Focuswriter doesn't seem to do any sort of background image on the page (just around it) over [02:59] Any suggestions? === jacke is now known as Guest688 === bahamat_ is now known as bahamat === Abrax is now known as |_ === |_ is now known as |__ === |__ is now known as ||_ === ||_ is now known as |__ === Abrax is now known as ___ === ___ is now known as group === group is now known as ___ === ___ is now known as __ [04:47] Should I get Ubuntu Long-term-support or the latest one ? [04:47] Itching to try out the latest version […] [04:48] the users choice IRCMonkey [04:48] feel free to start a topic in #ubuntu-discuss [04:49] Alright, I'll go with the latest version —even my IRC client's beta build. :D [04:50] lotuspsychje: Isn't that on-topic for here though? [04:51] it is on topic imo [04:51] it's a question about Ubuntu [04:52] i never said its offtopic, im suggesting he can discuss it further [04:52] not needed in discuss :) [04:53] IRCMonkey: long term support versus 9 month support, choose wisely [04:53] Alright thanks, I will ! === EriC^^_ is now known as EriC^^ [07:15] I want to open a terminal on startup, run a command in it, and then keep it open [07:15] how can I do that? [07:18] if I put `mate-terminal -x COMMAND` in "startup applications", will it stay open after running that command? [07:20] mate-terminal --tab -e "command1" --tab -e "command2" --tab & [07:20] That one opens 2 commands in two tabs, and a third tab with the prompt [07:20] If you want just one, it's just one -e [07:22] thanks alkisg! [07:23] BRB, testing [07:30] okay, I'm halfway there [07:30] I rebooted, the terminal with the result of the command (less) was opened, but the issue is, when I exit less, the terminal closes, and I'd like it to remain open - is there a way to do this? [07:30] bomzh1018: that would mean that you'd run bash, not a command [07:30] So: mate-terminal -e "bash -c 'command'" [07:31] Wait [07:32] mate-terminal -e "bash -c 'less /etc/adduser.conf && exec bash'" [07:32] This works, but it would be simpler to do it via a bash script, like mate-terminal my-session [07:32] Then my-session would run less and then it would exec bash [07:34] do I need to explicitly add `exec bash` to this new script? [07:34] Yes, to specify that you want to read the user's stdin from that point on [07:36] thanks, trying... [07:41] not quite. Now I got a terminal, but it was empty, with no less output [07:41] what I added to "Startup applications" was the following: [07:41] mate-terminal ~/Estate/record_time.sh [07:41] and record_time.sh is https://paste.ubuntu.com/p/sgTRRVZcpB/ [07:42] use full path [07:43] BRB, trying [07:46] nope, the same, empty terminal [07:46] bomzh1018: you forgot the -e, `mate-terminal -e ...` [07:46] Also did you chmod +x the shell script? [07:46] You don't need to logout to test [07:46] Just run the command in a terminal [07:48] 1) got it [07:48] 2) yes, I set a+x [07:48] 3) works now in terminal, need to reboot to see if it works in real situation, BRB [07:49] bomzh1018: sudo systemctl restart lightdm [07:50] A faster test than reboot [07:51] YES! Now it works exactly as I want it to. Thanks a lot, guys! [07:51] have fun! [08:55] sarnold, oerheks, remember me? https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/1848892 [08:55] Launchpad bug 1848892 in grub2 (Debian) "'error: Unknown TPM error.' after upgrading to grub 2.04" [Unknown, Confirmed] [08:56] yeah, rmmod tpm works, the system now boots with recent/updated grub2. funny thing is, i have two identical boxes, neither have a tpm and only one is affected by this issue... [08:58] also, not an issue on grub2 2.02-2ubuntu8.21 (but on every following update) [09:02] shibboleth, file a bugreport? [09:02] !bug [09:02] If you find a bug in Ubuntu or any of its official !flavors, please report it using the command « ubuntu-bug » - See https://help.ubuntu.com/community/ReportingBugs for other ways to report bugs. [09:02] it's the same as they existing one [09:02] also, same as https://bugzilla.redhat.com/show_bug.cgi?id=1833148 [09:02] bugzilla.redhat.com bug 1833148 in Fedora "DNF update which includes grub2-efi leads to a system with unbootable kernel entries including the recovery kernels" [High, Closed: Duplicate] [09:02] https://bugzilla.redhat.com/show_bug.cgi?id=1779611 [09:02] bugzilla.redhat.com bug 1779611 in Fedora "System fails to boot after upgrade to grub2-efi-x64-1:2.02-103.fc31.x86_64" [High, Closed: Errata] [09:04] or confirm that one. comment #3 said 2.04-1ubuntu12.1 fixed it. [09:05] yeah, it did not [09:05] at least not on 1804, the thread is by a user on 1904 [09:35] Hi, In my log (20.04) I see messages about gltlab-runsvdir.service being unable to start as there is no runsvdir-start file [09:35] actually I do not recall installing anything gitlab related. How do I get rid of this? [09:37] the only thing I have in /opt/gitlab/embedded/bin is a symlink prometheus linking to itself; not even sure what that is (or how I got /opt/gitlab in the first place) [09:38] eFfeM: by installing something from outside the Ubuntu archives [09:39] TJ-: hmm, yes, but what (and how to find out what it is) [09:39] eFfeM: only you can know that! [09:39] actually this system is updated over and over again from 12.04 or so [09:40] so it carries a lot of history :-) [09:40] /opt/ is for optional non-distro packages [09:40] What's the output of `dpkg -S /opt` ? [09:40] eFfeM: maybe the timestamps on those directories will give you a clue? [09:40] progs in /opt/ probably never updated [09:41] TJ-: output is [09:41] signal-desktop, skypeforlinux, wine-stable-amd64, wine-stable-i386:i386, wine-stable, youtube-to-mp3, google-chrome-stable, net.downloadhelper.coapp, gitlab-ee, jd-gui, google-earth-stable: /opt [09:41] guess gitlab-ee is the bad guy, timestamp is from 2018 [09:55] and here are a bunch of others: https://bugs.launchpad.net/snapd/+bug/1878541 [09:55] Launchpad bug 1878541 in grub2 (Ubuntu Groovy) "Grub fails to load kernel from squashfs if mem < 1500mb" [High, Fix Released] [09:56] https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/1848892 [09:56] Launchpad bug 1848892 in grub2 (Debian) "'error: Unknown TPM error.' after upgrading to grub 2.04" [Unknown, Confirmed] [09:56] also affects debian. confirmed and unassigned since last year [09:59] shibboleth: what architecture is the UEFI on your system? The only significant recent change that is in grub-2.04 is to add support for 32-bit UEFI [09:59] x86, haswell [09:59] shibboleth: that is, enable tpm module on 32-bit firmwares [10:00] https://lists.gnu.org/archive/html/grub-devel/2019-10/msg00104.html [10:01] Mathieu Trudel-Lapierre wouldn't happen to work at canonical/ubuntu? [10:01] shibboleth: I'm aware of that [10:01] since i have two identical boxes and only one is affected it is likely i might be able to help narrow down the exact cause [10:04] People here running succesfully docker swarm on 21.04 ? I'm having swarm connection issues and I would like to know it's it's distro related [10:06] shibboleth: at what point does the boot fail? before the linux kernel is loaded into memory, or after? [10:07] bios post->grub menu->black screen+cpu freeze after selecting/timing out [10:09] when menu -> c -> set debug=all -> esc -> select it freezes during.... disk/efi/efidisk.c reading 0x40 sector from sector... something [10:10] due to OOM if i've understood the threads [10:11] shibboleth: have you tested using one of the Advanced sub-menu options where they write "Loading kernel..." "Loading initrd..." so you know which stage it reached? [10:11] shibboleth: I do a lot of hacking on GRUB so may be able to narrow this down [10:11] the boxes are identical. call the alice and bob. bob always fails to boot with grub > grub2 2.02-2ubuntu8.2. i can take the drive from alice, also fails in bob. drive from bob works fine in alice [10:12] TJ-, please xplain the adv submenu stuff? [10:14] i was hit by this issue in after an apt-get upgrade eraly/mid May. only thing that worked was to replace the .efi binaries in /boot/efi/... with the ones from grub2 2.02-2ubuntu8.2. issue persist on all of the... four grub2 updates released since [10:14] shibboleth: if you can access the GRUB menu (tapping Esc key as GRUB starts) or set the config to show it by default (/etc/default/grub "GRUB_TIMEOUT_STYLE="menu" " [10:15] shibboleth: then the Advanced sub-menu lists Recovery boot options, and they report what stage the file loading has got to. [10:15] yeah, one sec [10:15] Therefore, if using one of those you don't see "Loading Linux..." we know it is happening before or during that stage. By adding your own "echo 'Reached stage X' " in front of each line in a menuentry block, you can identify exactly where GRUB is getting to [10:16] ok, in advanced. do in just select the recovery option or do i add any flags? [10:16] shibboleth: initially I'd add a line just before the 'linux ...' line saying "echo 'About to load Linux...' " then Ctrl+X to boot with that, and see what you get [10:17] there already is an echo "Loading $linuxkernelversion" [10:18] and a sep one for init [10:18] rd [10:18] shibboleth: oh yeah, of course! [10:18] shibboleth: ok, so just boot with that entry (I was thinking of the opposite of what I do when there are initrd.img problems and I add a line /after/ the intrd ... line to ensure it was loaded into memory [10:19] ok, loading linux $version, loading initial ramdisk -> freeze [10:20] shibboleth: great... so now try again with my initrd process - add a line after "initrd ..." with "echo 'Loaded initrd, handing over to kernel now...' " [10:23] doesn't echo what i added after initrd [10:23] shibboleth: good - so we know it fails during loading of the initrd [10:24] well, yeah. since the kernel is signed it would be bizarre otherwise :) [10:24] shibboleth: can you tell me how large the initrd.img file is "ls -l /boot/initrd.img*" [10:28] shibboleth: did you see the message written by this line? grub_error (GRUB_ERR_OUT_OF_MEMORY, "cannot allocate pages"); [10:29] no, and now it froze doing ls -al (hd0,gpt2)/* [10:30] shibboleth: so EFI didn't run out of memory. Secondly, can you put GRUB into debug mode and tell if/what this line reports as values " grub_dprintf ("linux", "[addr=0x%lx, size=0x%lx]\n", " - this is just before it calls the EFI loader for the initrd.img [10:32] ok, froze a second time doing ls -al (hd0,gpt2)/* [10:32] ten, the third time i did rmmod tpm, ls -al (hd0,gpt2)/* [10:32] now it lists and doesn't freeze [10:33] initrd: 59985151 [10:34] unlikely to be the initrd itself since i actually get a cpu stall doing ls -al from grub shell without rrmod ttpm [10:35] shibboleth: OK, just checking, since I hit a size-related bug recently [10:35] without rmmod tpm it freezes after listing half the files/folders in /boot [10:36] cuts at x86_64-efi [10:36] shibboleth: ok, you need to add "debug=linux" to the grub environment. You should be able to do that by, at the grub menu, pressing 'c' to get to the command-line, then typing "set debug=linux" then press 'Esc' to return to the menu, then go to the Advanced sub-menu and boot one of those entries again [10:36] shibboleth: If I've got the correct you should get debug messages only related to loading Linux [10:37] shibboleth: and bonus points for taking a photo of the resulting messages and uploading it for me! [10:37] yeah, nothing. freezes after loading linux $version, loading initrd [10:38] i tried adding debug=chain,secureboot,lb [10:38] and i got secureboot 0, setupmode 0 then freeze [10:38] hmmm, then I think you will need to add the line "debug=linux" to the menuentry block itself in the Advanced menu. (press 'e' to edit the block as before, put that as the first line, then press Ctrl+X) [10:39] shibboleth: do you have GRUB set to boot in graphical or text mode? text mode might help here [10:39] text [10:40] good [10:40] is it possible to pipe the output from debug=all/debug=linux to a file? [10:40] then i could just pull the drive and upload [10:40] add the 'linux' debug condition and try again [10:40] shibboleth: no [10:41] shibboleth: 'all' slows things down and is generally not a good thing - changes behaviour such as trying to catch race conditions [10:42] shibboleth: you're seeing "setupmode 0" so that shows debug=secureboot is working at least [10:43] i'm downing a sandwich, brb [10:43] I'm chasing flying ants! === jpe__ is now known as sharperguy [10:48] ok, i'm back. trying menu -> c -> set debug=linux -> esc ->boot menuitem first [10:51] loader/i386/linux.c:669 trying linuxefi, loader/i386/linux.c:682 handing off to linuxefi -> freeze [10:52] set debug=linux doesn't produce any extra output when booting the recovery option, but produces the above for the "default" option [10:54] TJ-, but, eh... seriously. the grub binary completely *borks* doing a simple ls -al without rmmod tpm? what in the name of a lot that is holy? [10:55] it's not the file system of /boot. i've done fsck and it works totally fine in alice [10:55] shibboleth: makes sense... GRUB asks EFI to read the file-system... which will contain files not signed [10:56] and TPM tries to 'measure' (verify) file(s) and fails [10:57] well, these are consumer asus mobos with no tpm [10:57] shibboleth: doesn't surprise me. Code originally added by mathew garret :) [10:58] dunno what that means :) [10:58] well, the name doesn't ring abell [11:00] shibboleth: long history :) [11:00] shibboleth: anyhow, try adding "debug=tpm..." to your list of debug conditions; let's see what the module /thinks/ it has found. I've just spotted a potential bug in handling of grub_tpm_handle_find [11:01] and what between 2.02-2ubuntu8.21 and 2.02-2ubuntu8.23 can have caused this? also, mind verifying that 8.23 hit main repos early/mid may? [11:01] ok, trying debug=tpm [11:03] menu -> c -> set debug=tpm -> esc ->boot menuitem -> black screen/freeze [11:03] no output [11:03] set debug=linux,tpm? [11:03] shibboleth: mix in all the other you were using, just in case - it'd be good to know the entire debug= isn't being ignored [11:04] shibboleth: and always use the Advanced sub-menu options so we should at least get those "Loading ..." messages [11:06] yeah, using any of the advanced options i only get loading linuxversion, loading initrd even if i added set debug=foo [11:06] set debug=chain,linux,tpm,secureboot and booting "default" i get trying linuxefi, secureboot0, setupmode0, handing off to linuxefi ->freeze [11:09] shibboleth: you see "efi" just after "handing off to linux" ? [11:09] loader/i386/linux.c:669 trying linuxefi, loader/i386/linux.c:682 handing off to linuxefi -> freeze [11:09] or is "efi" part of your added message [11:09] ahhh, that one [11:09] shibboleth: AHA! "i386" [11:10] ok, lemme try the same on alice and see if it dosn't say i386 [11:11] says i386 on alice as well [11:11] but it proceeds after those two lines [11:11] shibboleth: in my source for 2.02 there is no "trying" [11:12] well, right now bob is on bionic grub2 latest, 2.04 something. otherwise there'd be no point in debuggin, eh? [11:13] how would that work? me: boots fine, you: ok [11:13] none of those strings exist in upstream GRUB [11:14] there even are line numbers, 669 and 682 [11:14] i doubt we're talking gremlins [11:15] aha, Ubuntu patch debian/patches/0076-ubuntu-Make-the-linux-command-in-EFI-grub-always-try.patch [11:16] 118 line patch, too, and invasive for the 'linux' command [11:25] the critical part based on our experiments so far is " initrdefi_cmd = grub_command_find ("initrdefi");" in grub-core/loader/i386/linux.c::grub_cmd_linux() [11:27] shibboleth: right, another debug condition to add: "linuxefi" so we can monitor grub_cmd_initrd() [11:29] ok, that produced some more text, one moment [11:32] set debug=linux,chain,tpm,secureboot,linuxefi: loader/i386/linux.c:669 trying linuxefi, kern/efi/sb.c:46: secureboot: 0, kern/efi/sb.c:54: setupmode: 0, then several lines of loader/i386/efi/linux.c, then loader/i386/linux.c:682 handing off to linuxefi -> freeze [11:32] it's like fifteen lines, anything specific too look for or you need it all? [11:44] TJ-, ping [11:50] shibboleth: sorry, attacking ants again! [11:50] shibboleth: uf the last line i the 'handing off to linuxefi' I presume you're not now seeing the "Loading linux ..." echoed? [11:51] ok, something funky. on alice (the one that doesn't freeze), grub menu -> c -> ls -al (hd0,gpt2)/* also cuts at x86_64-efi but it doesn't freeze, returns to grub cmd. doesn't list the kernel, initramfs [11:51] lemme retry after rmmod tpm [11:52] nah, doesn't list kernel, initrd [11:52] TJ-, see pm [11:53] whereas on bob (the one that does freeze) it freezes after listing x86_64-efi, lists all (including kernel, initrd) after rmmod tpm [11:56] dunno if that matters at all, but... funky? [11:57] no, nothing after loader/i386/linux.c:682 handing off to linuxefi [12:04] i'm not sure that grub's "ls" understands "-al" [12:04] set debug=linux,chain,tpm,secureboot,linuxefi <--- add initrdefi? [12:04] well, it does on bob [12:06] yes it does, i just tested as well [12:06] +initrdefi, same output as without [12:06] shibboleth: no "initrdefi" - that is a command name. "linuxefi" is the condition [12:07] shibboleth: at this stage I'd suggest building and testing without the debian/ubuntu patches since I've a strong feeling the interference that several of the Ubuntu patches is doing into this very area is the cause [12:07] shibboleth: some of those patches are copied from RedHat too [12:08] ok, lemme read up on deps/procedure [12:09] but seriously, why doesn grub-efi on alice list vmlinuz, initrd? [12:09] whereas bob freezes unless rmmod tpm, lists vmlinuz, initrd after rmmod tom [12:10] this is like... imagine looking at a petri dish of bacteria in a microscope and seeing 300 hamsters doing the macarena [12:10] shibboleth: interacting in slightly different ways with the firmware [12:13] well, does ls -al (hd0,foo)/* list vmlinuz, initrd on bionic at your end? [12:15] also, which repo should i clone? [12:15] shibboleth: I don't use bionic, and don't have hardware free to test on [12:16] shibboleth: I'd use the Ubuntu package source but remove the patches (they're applied using quilt so easy to reset the Quilt stack to no patches) [12:18] shibboleth: it's a right mess; 238 patches! [12:18] http://archive.ubuntu.com/ubuntu/pool/main/g/grub2-unsigned/grub2-unsigned_2.04.orig.tar.xz ? [12:19] shibboleth: "sudo apt-get build-dep grub2" and "apt-get source grub2" then "cd grub2-unsigned-2.04" then "export QUILT_PATCHES=debian/patches" ... then "quilt applied" will list all applied patches [12:20] at that point you can test removing all patches with "quilt pop -a" [12:20] shibboleth: no guarantee the package will build cleanly though [12:23] isn't the point not to apply the patches? [12:24] yes, that is what we're doing [12:24] strangely I cannot get the build-dep's installed on 20.04! [12:25] shibboleth: when extracting the Debianised sources the patches are already applied, so we have to use 'quilt pop -a' to unapply them [12:35] git clone -b ubuntu/bionic-updates https://git.launchpad.net/~usd-import-team/ubuntu/+source/grub2-unsigned [12:35] shibboleth: solved that (custom kernels here) . Doing a teset build without patches [12:35] this will clone the sources for current bionic grub2 without patches [12:35] 2.04-1ubuntu44.1.2 (patches unapplied) [12:40] TJ-, hit a snag, pardon? [12:47] shibboleth: you hit a snag? [12:47] haven't built grub before, trying trying to find out how to get the raw .efi binary [13:02] tj: ok [13:02] i've built this: git clone -b ubuntu/bionic-updates https://git.launchpad.net/~usd-import-team/ubuntu/+source/grub2-unsign [13:02] should be what we're looking for, eh? [13:03] shibboleth: doesn't that contain the ubuntu patches? [13:03] no, they're not applied [13:03] https://code.launchpad.net/~usd-import-team/ubuntu/+source/grub2-unsigned/+git/grub2-unsigned [13:04] applied/ubuntu/bionic-updates 2021-06-07 [13:04] 2.04-1ubuntu44.1.2 (patches applied) [13:04] ubuntu/bionic-updates 2021-06-07 [13:04] 2.04-1ubuntu44.1.2 (patches unapplied) [13:04] on my laptop it displays 3 power modes, but my desktop only shows 2. [13:05] i've got the .efi binary, i should replace grubx64.efi on bobs /boot/efi/etc? [13:05] shibboleth: what command did you use to build it though? usually the tooling applies the patches automatically when doing the build [13:06] shibboleth: in theory yes but you'll also need to matching modules installing [13:06] /configure --prefix=/foo/grub2 --with-platform=efi [13:06] unless you built with all modules built-in [13:15] is there a way to get the latest virtualbox on ubuntu without using oracle repositories? [13:15] or the deb [13:15] you could build it from source [13:16] hmm so all those options will be unsupported either way [13:16] except for the repo version [13:16] ubuntu [13:17] canonical would not support any of them, i guess. in ubuntu's repositories, virtualbox will be in universe [13:18] ok [13:18] i'm not sure what oracle offers supprot for [13:18] I don't think they'd care I actually prefer KVM but I am tryingo out genymotion for android and it prefers vvirtualbox [13:19] if you want vitualbox working on ubuntu, in my experience your best option is to use oracles' apt repositories on an Ubuntu LTS release [13:20] I am on 20.04.2 LTS focal [13:25] tomreyn: do you know if there is a way to find out and uninstall packages from a user added repor in apt source and then remove that line? [13:25] O had added a collabora repo which I no longer want but dnt want stray packages from it on my system in case they are still there [13:26] aiena, apt-forktracer can be useful [13:27] and aptitude search '~i(!~ODebian)' [13:27] i got that one from https://www.debian.org/releases/buster/amd64/release-notes/ch-upgrading.en.html [13:27] but it should work on Ubuntu, perhaps replace the name, of course :) [13:30] I'm trying to work with a project that requires PHP 8, but it seems Ubuntu only has 7.5.x [13:30] aiena: apt list --installed | grep ',local\]$' [13:30] I'm using Focal Fossa, should I just grab PHP binaries and over-ride symlinks? [13:31] aiena: run the above after deactivating those 3rd party apt repositories + sudo apt update [13:32] tomreyn: so a local package is one which is not linked to any repo anymore? like a mnually installed .deb etc. [13:32] yes, aiena [13:32] thanks [13:33] more precisely, a package version not provided by any of your currentlöy active (and known, to apt's cache) apt repositories [13:34] will it work is package foo was from another repo but had the same name in the official ubuntu repos too [13:35]