IRC channel logs

<Zopolis4>oh also i still have no idea how to expand a partition in fdisk the manual says nothing about it, and all the guides just tell me to make a new partition

<damo22>mark its start position, then delete the old one and create a new one with the same start, and larger end

<damo22>it doesnt touch the data, it just creates an entry in the table

<Zopolis4>oh ok cool thx

<Zopolis4>it tells me partition 2 contains an ext2 signature in an ugly red and asks me if i want to remove the signature

<Zopolis4>should i remove it?

<damo22>no

<Zopolis4>ok yeah thx everything seems good now

<Zopolis4>mk looking like it fails to boot on account of: https://www.irccloud.com/pastebin/ELIBjDMH/

<damo22>change it to sd0

<Zopolis4>wdym

<damo22>in grub

<damo22>root=part:2:device:sd0

<damo22>its when you swap ide/sata in qemu

<damo22>the disk numbering gets out of whack

<Zopolis4>i didnt use qemu tho

<Zopolis4>anyways where is the grub config? /boot/grub?

<Zopolis4>or /etc/grub.d/10_hurd

<damo22>youre better off entering grub on the running system, edit it live before it boots, then when youve got it working, just run sudo update-grub inside the machine

<Zopolis4>how do i run a system before it boots

<damo22>it boots grub before the system

<Zopolis4>is there like a key i press to stop it or something

<damo22>shift?

<damo22>you probably need GRUB_TERMINAL=console in /etc/default/grub

<damo22>esp if youre running coreboot

<Zopolis4>nvm i forgot about the way grub works

<Zopolis4>anyways as it stands its set root='hd-49,msdos2'

<Zopolis4>should i change that?

<damo22>not that one

<damo22>in the multiboot command

<Zopolis4>ah yeah

<Zopolis4>yeah that worked

<Zopolis4>sort of

<Zopolis4>now its failing because the swap is not working

<Zopolis4> https://www.irccloud.com/pastebin/BHo4Mac8/

<Zopolis4>oh wait

<Zopolis4>its the same issue as before isint it

<Zopolis4>nvm i cant find another hd0 within the grub config anywhere

<damo22>ive got APs running but when i try to reload the descriptor tables they crash

<damo22>youpi: my smp kernel gets to an interesting point, the APs are sitting in a tight loop just before slave_main() and the BSP almost boots the system but gets stuck on an interrupt or something

<damo22>its really difficult to figure out why it crashes when i let the APs go into slave_main

<damo22>i cant seem to get into kdb

<damo22>perhaps i can trigger a trap?

<youpi>I don't know if kdb can work in smp

<youpi>perhaps better use gdb outside ?

<damo22>BSP: Completed SMP init

<damo22>{cpu0} ../i386/i386/mp_desc.c:461: start_other_cpus: Assertion `cpu == 0' failed

<damo22>.Debugger invoked: assertion failure

<damo22>Kernel Breakpoint trap, eip 0xc1025f54

<damo22>Stopped at Debugger+0x13: int $3

<damo22>Debugger(c116159d,c115fa3d,f4845f58,c1025faa,0)+0x13

<damo22>Assert(c115fa3d,c115fa26,1cd,c1155c6c,c115fa6b)+0x69

<damo22>start_other_cpus(f4849ea0,c102ad10,0,c1000010,10)+0xdc

<damo22>start_kernel_threads(c10315a0,0,0,c1055095,0)+0xe1

<damo22>thread_continue(0,f4846ed8,c10315a0,f4845fe4,0)+0x2e

<damo22>Thread_continue()

<damo22>i made it assert from the AP as well and it crashed

<damo22>but asserting from BSP works

<youpi>possibly the traps are not properly set up on the AP

<damo22>yea

<Pellescours>damo22: isn’t there missing a sizeof here http://git.zammit.org/gnumach-sv.git/tree/i386/i386/mp_desc.c?h=feat-smp#n166 ? I think it should be "sizeof(_mp_desc)"

<Pellescours>forget what I said

<civodul>hi!

<civodul>i noticed that the Hurd would fail to boot when run with QEMU -enable-kvm on an AMD EPYC CPU

<civodul>(the kvm-amd Linux kernel module)

<civodul>everything's fine if i remove -enable-kvm

<civodul>(or run the same thing on an Intel CPU)

<civodul>does that ring a bell?

<Pellescours>try to not use cpu=host to see if that works?

<civodul>that doesn't make any difference, but then again it's an x86_64 CPU

<civodul>i tried --cpu pentium3, to no avail

<Pellescours>do you have an error message? does it reboot without no message? It can be a kvm bug on AMD cpu, if that was a AMD specific feature that make it incompatible with the current gnumach then the -cpu pentium 3 would have fix it

<Pellescours>"fix", I mean make it running

<civodul>no error message: the startup process just hangs after starting ext2fs and before starting exec

<civodul>in kdb, i can see that only ext2fs is running, with a bunch of threads

<civodul>some stuck in guix-build-glibc-cross-i586-pc-gnu-2.33.drv-0/build/mach/mach_msg_trap.S:2

<Pellescours>damo22: I’m looking the intel documentation, and for start AP, after having enable lapic spurious_vector, they setup the Local Vector Table (LVT3), do you know what is it?

<Pellescours>It seems to be lvt_error in lapic

<Pellescours>Note: this lvt_error should also be configured in BSP in theory

IRC channel logs

2022-10-05.log