lmb / vimto

Run Go tests on a custom kernel
MIT License
18 stars 2 forks source link

vimto won't boot on kernel 6.12-rc1 #19

Open ti-mo opened 1 week ago

ti-mo commented 1 week ago

https://github.com/cilium/ebpf/actions/runs/11162212449/job/31026497906?pr=1574

On 6.12-rc1, the kernel doesn't init properly:

[    1.436835] Run /home/runner/go/bin/vimto as init process
[    6.935466] request_module: modprobe binfmt-0000 cannot be processed, kmod busy with 50 threads for more than 5 seconds now
[    6.935781] kworker/u8:49 (143) used greatest stack depth: 13872 bytes left
[    6.943482] kworker/u8:38 (121) used greatest stack depth: 13696 bytes left
[    6.950141] kworker/u8:5 (55) used greatest stack depth: 13616 bytes left
[    6.952581] kworker/u8:3 (51) used greatest stack depth: 13600 bytes left
[    6.954567] kworker/u8:2 (49) used greatest stack depth: 13520 bytes left
[    6.957106] Kernel panic - not syncing: Requested init /home/runner/go/bin/vimto failed (error -8).
[    6.959449] CPU: 1 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.12.0-rc1 #1
[    6.961277] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
[    6.963493] Call Trace:
[    6.964197]  <TASK>
[    6.964778]  panic+0x32b/0x390
[    6.965403]  ? kernel_execve+0x60/0x150
[    6.966270]  ? __pfx_kernel_init+0x10/0x10
[    6.967180]  kernel_init+0xd2/0x1c0
[    6.967965]  ret_from_fork+0x34/0x50
[    6.968750]  ? __pfx_kernel_init+0x10/0x10
[    6.969666]  ret_from_fork_asm+0x1a/0x30
[    6.970558]  </TASK>
[    6.971576] Kernel Offset: disabled

It's not obvious to me what the issue could be, will have to spelunk the kernel source. Disabling the mainline CI job for now.

lmb commented 5 days ago

Weird! Maybe this is an rc1 regression?

dylandreimerink commented 4 days ago

It is. https://github.com/torvalds/linux/commit/ee4cdf7ba857a894ad1650d6ab77669cbbfa329e broke v6.12-RC1. It will be fixed in RC2 https://lore.kernel.org/all/cbaf141ba6c0e2e209717d02746584072844841a.1727722269.git.osandov@fb.com/

So I think we can close this issue, as its not a vimto bug but a p9 bug