I was using Crucial P3 Plus PCIe 4.0 NVMe M.2 as NAS for backup and file sharing. However, the access to the SSD no longer work after upgrade to 25.12.4.
Tried couple of things:
Reinstall Samba4
Check the volume and everything fine, now cannot mount the disk to /mnt/
Device: BPI-R4
OpenWrt version: 25.12.4
Issue: After upgrade to 25.12.4, LUKS on NVMe (ct2000p3pssd8, /dev/nvme0n1p1/p2)
hangs when running any cryptsetup command (even isLuks).
Here is the lsmod output for nvme:
Checked that the nvme is loaded.
Any information you forgot to discuss or maybe assumed we'd guess (e.g., explaining a Samba issue to accompany the configuration post, or it it non-functional because of the encrypted partition)?
(In other words - are we to assume your mounts are on the still encrypted partition, hence related?)
I bought the 2TB SSD with following paritions, while using 25.10.6 version and everything work.
Below is the parition information with LUKS encryption.
259 3 1953514584 nvme0n1
259 4 1048576000 nvme0n1p1 -> For Timemachine backup, mount to /mnt/timemachine
259 5 904937560 nvme0n1p2 -> For NAS data files, and mount to /mnt/data
After upgrade to 25.12.4, I found that all mount file were no data, and 2 keys files were disappeared. then I restore the key files back to the same folder, and using command chmod 600 to change the permission.
The result still not good, and then I check the samba and the service is enabled.
So, I check the partition is good or not using.
cryptsetup isLuks /dev/nvme0n1p1
cryptsetup isLuks /dev/nvme0n1p1
the terminal hangs without return.
I also checked the nvme version and the kmod-nvme was installed.
kmod-nvme-6.12.87-r1 aarch64_cortex-a53 {feeds/base/kernel/linux} (GPL-2.0-only) [installed]
libnvme-1.15-r1 aarch64_cortex-a53 {feeds/packages/feeds/packages/libs/libnvme} (GPL-2.0-or-later)
nvme-cli-2.15-r1 aarch64_cortex-a53 {feeds/packages/feeds/packages/utils/nvme-cli} (GPL-2.0-or-later)
I am trying to detach the SSD and attache to my macbook and test.
I don't know whether still have anything I can check. please comment and I am willing to try.
Yes, I can open another terminal and use ssh command.
below is the return from logread -f command, but seems not related (please correct me if I am wrong)
Fri Jun 5 20:53:45 2026 daemon.err tailscaled[3533]: 2026/06/05 12:53:45 magicsock: derp-20 does not know about peer [xxxx], removing route
Fri Jun 5 20:53:45 2026 daemon.err tailscaled[3533]: 2026/06/05 12:53:45 [RATELIMIT] format("magicsock: derp-%d does not know about peer %s, removing route")
For testing the SSD from my mac, need some time, will reply later.