Re: 13.2-RELEASE-p11 dom0 fails to boot unlike -p10

On 6/15/24 11:24, Marcin Cieslak wrote:
> Hello,
> 
> After upgrading 13.2-RELEASE-p10 amd64  machine to 13.2-RELEASE-p11,
> the dom0 stops booting, last lines being:
> 
> pcib0: <ACPI Host-PCI bridge> port 0xcf8-0xcff on acpi0
> pci0: <ACPI PCI bus> on pcib0
> 
> (see screenshot)
> 
> After the pci0 the following would normally follow:
> 
> pcib1: <ACPI PCI-PCI bridge> irq 16 at device 1.0 on pci0
> pci1: <ACPI PCI bus> on pcib1
> vgapci0: <VGA-compatible display> port 0xf000-0xf03f mem 
> 0xf7800000-0xf7bfffff,0xe0000000-0xefffffff irq 16 at device 2.0 on pci0
> vgapci0: Boot video device
> 
> 
> GENERIC kernel.
> Xen version used:
> 
> # pkg -c /mnt info -x xen
> edk2-xen-x64-g202308_4
> xen-kernel-4.18.2.20240411
> xen-tools-4.18.2.20240411_1
> 
> xen 4.18.0 also didn't boot.
> 
> I realized that reverting 8f1f4e60ceb9b8e5eddd54cf1fde62944f56eaa4
> fixes the issue.
> 
> I am a bit puzzled, since I don't even load if_wg, but
> checking out f5ac4e174fdd3497749e351c27aafb34171c5730
> and reverting 8f1f4e60 helped.
> 

There's 0% chance of that change directly causing this, especially if 
you don't even load the module or build it into your kernel.  Even if 
you did, this is too early for tunnels to be constructed and pass 
traffic which would certainly be a prerequisite.

> Reverting f3195cc08ccc99365ec00900a3c1abc59ceefc9c alone didn't fix
> it (I tried this first).
> 
> What's going on..?
> 

I would rather suspect that for some reason or another your kernel is 
linking in a different order between the two builds and you're getting 
really unlucky (e.g., maybe there's a pair of sysinits that, when 
flipped, are causing this).  CC'ing cperciva@ because he's done more 
recent work with sysinit reordering (though I don't think anything that 
landed in 13.x) in case this particular case rings a bell.

> Marcin

Thanks,

Kyle Evans