On Mon, Nov 10, 2025 at 12:36:16PM -0700, Luke A. Call wrote: > On 2025-11-08 22:30:22+0000, Crystal Kolipe <[email protected]> > wrote: > > On Fri, Nov 07, 2025 at 03:40:27PM -0700, Luke A. Call wrote: > > > Which is what happened after I had to do a hard reset and re-do the > > > above, due to the keyboard freezing (^Q did't help nor ^C nor anything > > > else I tried). But I got back where I was, then doing "ls -lR|more" on > > > the mounted directory and ^Z causes it to lock again...so I will > > > reboot again and re-do fsck -p on it. > > > > Did this lock-up happen just once, or is it happening repeatedly whilst the > > machine is booted in to a 7.8 kernel? > > The first time may have been, and the 2nd time definitely it was when I > pressed ^Z while in "ls -lR|more" in the 7.8 kernel. The 3rd time as I > try to reproduce it similarly I get "--More--Segmentation fault" and it > didn't freeze (but I had not attached a disk with bioctl this time). > So not sure.
This is certainly a bit unexpected. If well tested programs such as ls and more are segfaulting whilst running in a known and fairly uncomplicated enviroment such as the ramdisk kernel then I would usually start to suspect hardware problems, (often ram memory failing), but you said that this machine previously ran 7.7, so that seems less likely. It might be worth running memtest86 or something similar to eliminate this possibility, or even just booting with less memory using something like 'machine mem =512M' at the bootloader configuration prompt for a quicker if less thorough test. > > > Maybe the error/panic that is causing the reboot > > > during upgrade comes in some fast subsequent step after the "Mounting > > > root filesystem" message. > > > But it flashes off too fast to see what comes next. Maybe if I took a > > > video, found some video editing software, and paused it to see the very > > > end. But happy to try more other things... > > > > The next step after the "Mounting root filesystem" message is bringing up > > the network devices. It's plausible that something could be going wrong > > there. > > > > But first, quickly check that the "checking root filesystem" and "mounting > > root filesystem" messages specify the expected device. The format of the > > latter message is: > > > > Mounting root filesystem (mount -o ro /dev/FOO /mnt)... > > > > ... where FOO is the device being mounted. Hopefully it is sd0a in your > > case. > > Yes, they say: > Checking root filesystem (fsck -fp /dev/sd0a)... OK. > Mounting root filesystem (mount -o ro /dev/sd0a /mnt)... OK. Well, at least we know that it is finding the correct root fs. > ...and the next thing I can see in the video is that the screen is > filled with repeated, hard-to-read lines, that seem to say: > panic: netlock rwlock 0xffffffff81955d48: enter write deadlock > ...and it immediately goes dark and reboots. So presumably something in the kernel called NET_LOCK(), likely with another thread waiting on a shared lock. That probably explains why it crashes when the installer reaches the point of bringing up the network interfaces. Seems odd that nobody ran in to this pre-release, assuming that it is actually a kernel bug rather than a hardware issue. It could be a timing issue, or maybe for some reason you are exercising a code path that didn't see much testing. > Curiously now, maybe due to something(???) I did while in the 7.8 install > disk shell, while booting it instead of prompting me for the FDE key, it > says this (though I can still boot from the disk and attach it with bioctl > and it prompts me for the FDE key and data is visible): > > Using drive 0, partition 3. > Loading...... > probing: pc0 mem[636K 2045M a20-on] > disk: hd0+ sr0 > >> Openbsd/amd64 BOOT 3.67 > open(hd0a:/etc/boot.conf): Invalid argument > boot> > cannot open hd0a:/etc/random.seed: Invalid argument > [the next lines are the ones I didn't get before this error of no FDE > prompt etc] > booting hd0a:/bsd: open hd0a:/bsd: Invalid argument > failed(22). will try /bsd > boot> > booting hd0a:/bsd: open hd0a:/bsd: Invalid argument > failed(22). will try /bsd > Turning timeout off. > boot> It's trying to find the kernel on the 'a' partition of your boot drive, which in your case is not an FFS filesystem but a softraid volume. It's easily fixed by running installboot and pointing it to the correct, (I.E. sd0a in your case), partition. But don't try to do that from the ramdisk kernel, (yet), because it requires the correct files to be in /usr/mdec. > > ...and then I type at the boot prompt this and again get: > boot> boot /bsd > cannot open hd0a:/etc/random.seed: Invalid argument > booting hd0a:/bsd: open hd0a:/bsd: Invalid argument > failed(22). will try /bsd > boot> > > > Assuming that is not the issue, and returning to the next step of bringing > > up > > the network devices... From your dmesg the only network card in this > > machine > > seems to be bce0. You could try disabling that device in the bootloader and > > see if that allows the 7.8 ramdisk installer to finish booting, (or at least > > get further). > > I did: > boot> boot -c > cannot open cd0a:/etc/random.seed: No such... > booting cd0a:/7.8/amd64/bsd.rd: ... > entry point at ... > .... > UKC> disable bce0 > UKC> disable adfadfa > UKC> find bce* > 144 bce* at pci* dev -1 function -1 flags 0x0 > UKC> disable bce* > 144 bce* disabled > > ...and continued into the install disk, did the shell commands for > bioctl etc., and continued with the install which again said: > Mounting root filesystem (mount -o ro /dev/sd0a /mnt)... OK. > panic: netlock rwlock 0xffffffff81955d48: enter write deadlock > ...filling the screen with the panic message, and reboots as before. > > The same thing happened with I tried disabling the net card from the > BIOS. OK, so it seems to be network related, but not triggered by the bce device. I just noticed you have a bwi wireless adaptor as well. Maybe try disabling that in the same way. Other than that and testing memory as I mentioned earlier, you'll probably need to wait for someone who actually worked on this code during the 7.8 development cycle to chime in with further advice.
