On Tue, May 13, 2025 at 09:32:26AM -0700, Nicole Findlay wrote: > On 2025-05-13 00:56, Stuart Henderson wrote: > > On 2025-05-12, Nicole Findlay <nicolefind...@airmail.cc> wrote: > > > I am trying to install OpenBSD 7.7 amd64 install77.iso in a VM, but > > > it panics while installing the sets. This is the error message: > > > "wdc_atapi_start: not ready, st = 50" > > > > > > The VM then reboots within approximately 30 seconds to a minute. > > > > > > These web pages I found have screenshots that are like what I am > > > experiencing, including the full error message: > > > <https://www.wezm.net/v2/posts/2023/openbsd-db-atapi-start-not-ready/> > > > <https://www.reddit.com/r/openbsd/comments/12jzg2y/comment/jhhk1gx/> > > > > > > They suggest that configuring the VM's virtual CD drive as SATA, > > > rather than IDE, is a way to avoid it. > > > > > > Unfortunately for me, the bespoke and strange QEMU-based (I think) > > > hosting infrastructure I am stuck using does not give me such > > > control or insight, so I cannot try that. There are several existing > > > pre-7.7 OpenBSD VMs hosted by this infrastructure, although I do not > > > know their provenance. This infrastructure's bespoke management web > > > app also makes extracting log messages impossible for me so far. > > > > > > Is this panic a bug, or expected behavior? (This is my first > > > experience with OpenBSD, so I am ignorant.) > > > > Probable that it's a bug, but not sure whether it's in OpenBSD or your > > VM environment. > > After further digging on my local computer, completely separate from > the aforementioned remote infrastructure, I think it may be an OpenBSD > bug.
I've just committed some changes that fix this, so upcoming snapshots should work better in this sort of setup. The information you provided here lead fairly directly to the problem, a lack of working delay function in the kernel on machines with no ACPI HPET. Out of curiosity, how did you end up testing with hpet=off?
