[Bug]: ntfs and ntfs3 refuse to mount one device in 7.0.0-1-cachyos

[vbgunz]

Kernel

7.0.0-1-cachyos

Desktop Environment / Window Manager

KDE Plasma

CPU

AMD Ryzen 9 9900X (24) @ 5.66 GHz

GPU / Driver

AMD Radeon RX 9060 XT [Discrete] / NVIDIA GeForce RTX 5060 Ti [Discrete]

Description

Upgrading to 7.0.0-1-cachyos, one NTFS partition on an NVMe disk would not only refuse to mount, it refused to even be seen by lsblk and KDE's Partition Manager. journalctl would complain "timed out waiting for ...". All other NTFS disks and partitions mounted fine. Changing fstab parameters and drivers (from ntfs3 to ntfs) made no difference. lsblk not only couldn't see the partition, it didn't see the disk itself.

Steps to Reproduce

1. On my particular setup, just booting into kernel 7 would break the mounting of one NTFS and NVMe disk.

Expected Behavior

For the disk and partition to work normally.

Actual Behavior

Disk refused to be seen by some tools almost as if it didn't exist.

Logs / Error Messages

Timed out waiting for device /dev/disk/by-uuid/...

Additional Context

Booting into an older snapshot, the disk comes online without problems. Booting into 6.19.11-zen1-1-zen kernel and that too has no problems. For some reason, the 7.0.0-1-cachyos kernel is having an issue with one specific NTFS NVMe drive. The drive is the SABRENT 1TB Rocket 4 Plus NVMe 4.0 Gen4 PCIe M.2 Internal SSD Extreme Performance Solid State Drive.

NOTE: I have a virtual machine that mounts the problematic disk just fine on either kernel 7 or 6.19.

Feel free to ask me anything. Thank you.

[ptr1337]

Hey,

Thanks for the report. The 7.0 Kernel uses a new NTFS Driver (rework/ntfs plus). Do you maybe see something in your dmesg to be printed when this fails?
Would you mind trying to mount and then run sudo cachyos-bugreport.sh?

There is also "ntfsprogs-plus" in the repository, which might would be worthwile to install. Not sure if it helps

[vbgunz]

@ptr1337 Thank you for pointing out sudo cachyos-bugreport.sh because I believe it helped me to actually solve the issue.

After running the report I opened it and immediately noticed the PCI vendor ID and device ID in the kernels cmdline vfio-pci.ids option. I thought I removed this a long time ago because the disk mounts with a slightly older snapshot and it mounts with a newly installed 6.19.11-zen1-1-zen kernel. I double checked /etc/default/limine config and it was indeed removed. So I ran sudo limine-mkinitcpio && sudo limine-update and rebooted.

After booting into the system, the disk was indeed mounted without issue.

I'm sorry for the bug-report but I'm grateful your mention of cachyos-bugreport.sh because I believe it helped me to fix the issue.