At Thu, 29 May 2025 16:30:39 -0400, Chuck Zmudzinski <frchu...@gmail.com> wrote: Subject: Re: Xen boot strangeness (Was: Re: [SOLVED] Re: Xen 4.18.5_20250521nb0 not ELF binary (Was: Re: EFI and Xen)) > > Ok, let's see what happens. I use this boot command line: > > menu=Boot normally with Xen:dev hd2d:;load /netbsd-XEN3_DOM0.gz -c > console=xencons bootdev=wd1;multiboot /xen.gz dom0_mem=2G dom0_max_vcpus=4 > com1=9600,8n1,0x40c0,16,1:0.0 console=com1 cet=no-ibt pv-l1tf=false > > Only difference from the one that works is that I do a s/com2/com1/g > > Still, DOM0 sees the serial port that Xen is using as com2, not com0, so > with this configuration it still crashes unless I disable com* using userconf. > > So it behaves exactly the same as when I told Xen the console is com2 instead > of com1.
Ah, that's unfortunate.
I started looking to see how the normal ISA COM ports are "hidden" from
the guest dom0 kernel, but I can't seem to find the right code.
I do find calls to pci_hide_device() for the debug device drivers, but
not for the console, and of course those are only for PCI devices.
I need to do more searching to see how that works, and maybe then I'll
see why it isn't hiding the PCI serial device used by the console.
> I boot GENERIC without Xen with this:
>
> menu=Boot normally:dev hd2d:;rndseed /var/db/entropy-file;boot
Ah ha, I see! That makes sense.
> Very simple and no searching for dk* devices needed, apparently. It
> can find root at boot because I set 'dev hd2d:' which is the fourth
> GPT partition on what Xen also sees as dk12.
because.... More magic based on old PC conventions!
> Note I also set dev hd2d: when booting Xen, but this information does not
> get passed to NetBSD dom0, so with Xen, we need to go through the search
> of all the dk* devices.
Indeed.
--
Greg A. Woods <gwo...@acm.org>
Kelowna, BC +1 250 762-7675 RoboHack <wo...@robohack.ca>
Planix, Inc. <wo...@planix.com> Avoncote Farms <wo...@avoncote.ca>
pgplMlHIztG78.pgp
Description: OpenPGP Digital Signature
