I appreciate your help immensely, and apologize for my incomplete post. The machine is an HP Pavilion Notebook, 15-au123cl. The previous OS on it was Windows 10, under which it was working without any problems. The default seems to be to boot EFI, but now that I've installed NetBSD, there is no way to confirm what it was doing previously. Physically, the machine runs well and is in excellent shape. I have not tried to install an earlier version of NetBSD on it. I am attempting to upgrade my "main server" from a turn-of-the-century IBM Thinkpad running NetBSD 7.0.2 (i386). The dmesg is attached as "dmesghpau123cl.txt.gz". While a longtime user, overall I'm totally clueless. Sorry. Please direct me if I should send anything else that might help pin the problem down. Thanks for working with me. Henry 2021年9月13日(月) 22:43 Greg Troxel <gdt%lexort.com@localhost>: > > > Henry <nbsd4ever%gmail.com@localhost> writes: > > > I've been trying to upgrade my main server to 9.x since early this > > year, first with 9.0_current, then 9.1_current and last week, > > NetBSD-9.2_STABLE. All three of these distributions have a glitch > > with /sbin/shutdown. I reported the problem with 9.1 on May 20, in a > > post "NetBSD 9.1, /sbin/shutdown not working". I was hoping it would > > be fixed in 9.2_STABLE, but no such luck. > > Reading so far, it seems likely that the bug is in the computer's BIOS, > vs netbsd. You didn't explain if this works under other OSes, and you > didn't give the model of your server. > > > Since I cannot power down the machine remotely, this glitch is fatal, > > and I cannot use NetBSD until I can pin down the cause and correct it. > > I'm desperate. > > Perhaps an esp8266 with a power relay? (sort of kidding, not really) > > > To review what happens: > > Issue the command '/sbin/shutdown -p now' and the system proceeds to > > shutdown normally, but comes to a dead stop without powering down. > > The last four lines left on the screen are (same as with 9.1): > > acpitz0: workqueue busy: updates stopped > > core1temp0: workqueue busy: updates stopped > > core2temp: workqueue busy: updates stopped > > acpibat0: workqueue busy: updates stopped > > The only way to power the machine off is to physically press and hold > > down the power button. > > So it looks like the problem is perhaps something in the kernel. It > seems likely that the actual problem is earlier than what you posted. > > > I really have no idea how to proceed. Would compiling a custom kernel > > perhaps solve the problem? What should I configure differently from > > GENERIC? > > It is possible that there is some hardware, for which the driver has > trouble, and removing that driver might help. But that's guessing. > > You also didn't post the dmesg from boot. > > > There is one major difference from previous installs of NetBSD that I > > have done. This is the first time for me to use an EFI machine and > > use a GUID partition table. Would it make sense to try installing > > with a BIOS MBR partition table to perhaps work around the > > /sbin/shutdown problem? > > You didn't explain if this hardware has ever been able to shut down with > NetBSD, and it won't with 9, or if this hardware never does. > > My guess is that gpt vs mbr is not that important, but legacy boot vs > efi boot might be. > > > After 25 years with NetBSD, I don't want to give up now! Any and all > > help appreciated. TIA! > > The biggest suggestion is for you to really describe what is going on in > far more detail.
Attachment:
dmesghpau123cl.txt.gz
Description: GNU Zip compressed data