I apologize for the too broad subject line.

I use LUKS to encrypt the physical volume of my LVM setup on this laptop. I always use the generic kernel with an initrd.

On this laptop, I had updated my Slackware64-current system on November 6, 2021 with a new kernel and an initrd based on it. That initrd worked (and is working now, as a matter of fact).

I recently upgraded this system and build a new initrd based on the 5.15.3 kernel on November 18, 2021, with that kernel and an initrd based on it. That initrd does not work; futzing about in the resulting command line tells me that the nvme device defined as the luks device doesn't exist at all when that bit of the init code executes.

That was a bit of a surprise and I wonder if anyone else has seen this problem.

Maybe not the same issue but: I installed Slackware-current with a kernel 5.15.3 yesterday on an external SSD drive connected through USB. At first I couldn't boot, as /dev/sdc5 didn't exist at time of mounting it on /mnt to switch root.

Today rebuilding an initrd with geninitrd I didn't have this issue with the generic kernel, but using the huge one leads to a kernel panic, even with a big rootdelay in the command line.

2 members found this post helpful.

@didier, thank you for your response.

I used geninitrd after your post, but the end result was the same for me as my original posting. I don't know if the French have a "Thank God that we've had a good enough harvest that we might live until next year" festival, but if you do, I wish you a happy one of those.

2 members found this post helpful.

Could this be related to why LILO stopped working on my SSD/NVME with recent kernels?

1 members found this post helpful.

i'm using LILO and it's still working on my machine, but it's normal SSD, not NVME

1 members found this post helpful.

Sunday's 5.15.4 kernel fixed the problem.

2 members found this post helpful.

Nice One, Richard Cranium