5.8 install hangs

[Blinker]

Hi, me again

After the kernel messages have scrolled past nothing happens. The last line reads

root on rd0a swap on rd0b dump on rd0b

and then just sits there. 5.7 on the same USB stick then drops into the installer but 5.8 doesn't. Installation fails the same way when booting off CD1 ,CD2, miniroot58.fs and the latest install58.fs from snapshots. The kernel can't configure Intel H57 LPC and Intel 3400 SMBus but 5.7 is the same and that boots OK. Machine is a Core i5 with 4Gb RAM and 250GB of spinning rust with one NTFS partition.

[jggimi]

Originally Posted by Blinker

Hi, me again

Welcome back!

After the kernel messages have scrolled past nothing happens.

Just like four years ago, you have a boot problem. But it's a different problem this time.

The last line reads

root on rd0a swap on rd0b dump on rd0b

and then just sits there.

I don't think a verbose boot will help, because this is the last kernel message expected during boot. After issuing that message, the kernel should pass control to init(8). You could always try a verbose boot, because like chicken soup, it wouldn't hurt. But I doubt it will help.

If you were to report this problem to the misc@ or bugs@ mailing list, someone would ask if you could capture the preceding dmesg in its entirety, which is only going to be possible if you have a serial port, a null-modem cable, and a second computer. Your working, 5.7 dmesg would probably be even more useful to developers to understand what's unique about your hardware.

If I recall correctly, the ddb(4) debugger is not included in these RAMDISK kernels, due to space. You might be able to boot a 5.8 (or -current) GENERIC kernel, and see if it fails the same way. If so, a backtrace from ddb(4) may provide insights to developers. Though if you do this, I'd recommend booting the kernel in single-user mode (boot -s), to avoid causing any data issues with userland daemons that might otherwise be started.

The crash(8) man page provides guidance on using both ddb(4) and gdb(1), and may be helpful.

[Blinker]

I got hold of a PCIe serial card but unfortunately it uses the MosChip MCS9922 chip which doesn't seem to be supported. Looking at other cards on the net they say in the blurb that the UART is a 16550 but looking closer at the pics it is again the cheap and cheerful MosChip that is on the board.

[jggimi]

Sorry about that.

As I'd mentioned above, a 5.7 GENERIC dmesg might be helpful in a bug report.