Mc / mcedit throwing a bus fault

[zdebel]

I’ve updated recently, the kernel switched from 6.10 to 6.12. Everything works just fine (torrents, KDE, various arr services and docker containers), except for mc / mcedit, they throw a bus fault and cause a NVME error in dmesg. Here’s a log:

[Sun Mar  2 15:39:07 2025] JFS: nTxBlock = 8192, nTxLock = 65536
[Sun Mar  2 15:39:07 2025] raid6: skipped pq benchmark and selected avx2x4
[Sun Mar  2 15:39:07 2025] raid6: using avx2x2 recovery algorithm
[Sun Mar  2 15:39:07 2025] xor: automatically using best checksumming function   avx       
[Sun Mar  2 15:39:07 2025] Btrfs loaded, zoned=yes, fsverity=yes
[Sun Mar  2 15:49:40 2025] nvme0n1: Read(0x2) @ LBA 856027192, 512 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE 
[Sun Mar  2 15:49:40 2025] critical medium error, dev nvme0n1, sector 856027192 op 0x0:(READ) flags 0x80700 phys_seg 7 prio class 0
[Sun Mar  2 15:49:40 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:49:40 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:49:40 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:49:40 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:49:40 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:49:40 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:52:29 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:52:29 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:52:29 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:52:29 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:52:29 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:52:29 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0

and now I’ll start mc once again:

[zdebel@zdebel-um350 ~]$ mc
Błąd szyny (zrzut pamięci) <- (it means bus fault)
[zdebel@zdebel-um350 ~]$ sudo dmesg -T | tail
[Sun Mar  2 15:57:53 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:57:53 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:57:54 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:57:54 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:58:07 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:58:07 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:58:07 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:58:07 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:58:07 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:58:07 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0

The system is working stable, everything as expected, EXCEPT for mc and mcedit which I’m addicted to. Maybe someone will have an idea.

[andreas85]

If you read the error messages carefully, the problem is not with mc or mcedit itself but with errors in the file system or on the data storage.

And it seems that you are running btrfs in the not recommended Raid 6 mode

@zdebel

[Sun Mar  2 15:39:07 2025] JFS: nTxBlock = 8192, nTxLock = 65536
[Sun Mar  2 15:39:07 2025] raid6: skipped pq benchmark and selected avx2x4
[Sun Mar  2 15:39:07 2025] raid6: using avx2x2 recovery algorithm
[Sun Mar  2 15:39:07 2025] xor: automatically using best checksumming function   avx       
[Sun Mar  2 15:39:07 2025] Btrfs loaded, zoned=yes, fsverity=yes
[Sun Mar  2 15:49:40 2025] nvme0n1: Read(0x2) @ LBA 856027192, 512 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE 
[Sun Mar  2 15:49:40 2025] critical medium error, dev nvme0n1, sector 856027192 op 0x0:(READ) flags 0x80700 phys_seg 7 prio class 0

https://btrfs.readthedocs.io/en/latest/mkfs.btrfs.html#mkfs-section-profiles

Note 4:

Since kernel 5.5 it’s possible to use RAID1C3 as replacement for RAID6, higher space cost but reliable.

---

Welcome to Manjaro!

1. Please read the information behind this link. It will help you to post necessary information.

[HowTo] Request support

Difficulty: ★☆☆☆☆ Strong of its many members, the Manjaro support forum can provide you help whenever you have an issue with your Manjaro installation. But in order to work efficiently, we shall also ask you to follow three major points. Provide context Simply signaling an issue is rarely enough to understand how it occurred. It is thus important to provide details on how it happened: Detail prior actions leading to the issue. List solutions and guides you already tried, with links whene…

2. Please press the 3 dots … below your post and then press the

• If you give us information about your system, we can see what we’re talking about and make better suggestions.
• You can do this by using inxi in a terminal or in console.

sudo inxi --verbosity=8 --filter 

• Personally identifiable information such as serial numbers and MAC addresses are filtered out by this command
• Presenting the information in this way allows everyone to be familiar with the format and quickly find the items they need without missing anything.

3. Copy the output from inxi (including the command) and paste it into your post.

• To make it more readable, add 3 backticks ``` on an extra line before and after the pasted text.

[HowTo] Post command output and file content as preformatted text

Difficulty: ★☆☆☆☆ Formatting differences Discourse – this forum’s engine – uses Markdown to format its posts, and removes “unnecessary” whitespace. Terminal commands often produce output formatted so that they are better readable by the user, taking advantage of the monospaced font used. Programming code often rely on indentation to make it more readable. Using preformatting in Markdown Markdown includes a pattern to provide preformatting when necessary. You only need to put the part to form…

4. Or we can try to play hide and seek
   Have you already for your problem in the forum ? (Wisdom lies in asking → listening → reading )

---

[linux-aarhus]

mc cannot be blamed for your disk errors.

@zdebel

[...]
[Sun Mar  2 15:39:07 2025] Btrfs loaded, zoned=yes, fsverity=yes
[Sun Mar  2 15:49:40 2025] nvme0n1: Read(0x2) @ LBA 856027192, 512 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE 
[Sun Mar  2 15:49:40 2025] critical medium error, dev nvme0n1, sector 856027192 op 0x0:(READ) flags 0x80700 phys_seg 7 prio class 0
[...]

[Molski]

It’s the one part of btrfs that scares a good chunk of people away thinking btrfs is not stable. But unfortunately, RAID 5 (or with double parity, aka RAID 6), has an outlying bug on unclean shutdowns; it’s rare, but that makes it unusable for most cases.

But like @andreas85 said, there are lots of safe options with any combination(s) of mirrors and stripes.

Just boot a live image and test to be 100% certain you get those errors. You will most likely get the exact same errors, but for a chance at full recovery of data, why not.

There are also ways to see which data is affected if you want that as well. (But it does require a tool, or converting LBA numbers to inodes to file names.)

[zdebel]

Oh boy, this will be a long one

[zdebel@zdebel-um350 ~]$ cat /etc/fstab 
# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a device; this may
# be used with UUID= as a more robust way to name devices that works even if
# disks are added and removed. See fstab(5).
#
# <file system>             <mount point>  <type>  <options>  <dump>  <pass>
UUID=AB20-1D79                            /boot/efi      vfat    umask=0077 0 2
UUID=06991fb8-c25b-4564-a30d-aa8b09f0ec25 /              xfs     defaults,noatime 0 1
tmpfs                                     /tmp           tmpfs   defaults,noatime,mode=1777 0 0

and follow this:

[zdebel@zdebel-um350 ~]$ date 
Sun Mar  2 20:35:51 CET 2025
[zdebel@zdebel-um350 ~]$ sudo dmesg -T | tail -n5 
[Sun Mar  2 15:58:07 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:58:07 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:58:07 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:58:07 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 17:32:39 2025] pcieport 0000:00:08.1: PME: Spurious native interrupt!
[zdebel@zdebel-um350 ~]$ echo "How do you like dem applez?" > applez
[zdebel@zdebel-um350 ~]$ sudo dmesg -T | tail -n5
[Sun Mar  2 15:58:07 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:58:07 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 15:58:07 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 15:58:07 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 17:32:39 2025] pcieport 0000:00:08.1: PME: Spurious native interrupt!
[zdebel@zdebel-um350 ~]$ mcedit applez
Bus error (core dumped)
[zdebel@zdebel-um350 ~]$ sudo dmesg -T | tail -n5
[Sun Mar  2 20:36:19 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 20:36:19 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 20:36:19 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[Sun Mar  2 20:36:19 2025] nvme0n1: Read(0x2) @ LBA 856027296, 128 blocks, Unrecovered Read Error (sct 0x2 / sc 0x81) MORE DNR 
[Sun Mar  2 20:36:19 2025] critical medium error, dev nvme0n1, sector 856027296 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[zdebel@zdebel-um350 ~]$ cat applez
How do you like dem applez?
[zdebel@zdebel-um350 ~]$ thoughts?

[linux-aarhus]

@zdebel

Oh boy, this will be a long one

Use the following to provide insight in your blockdevice structure

lsblk -f

@zdebel

UUID=06991fb8-c25b-4564-a30d-aa8b09f0ec25 /              xfs     defaults,noatime 0 1

@zdebel

zdebel@zdebel-um350

From your hostname I am thinking minisforum UM350 - with a single nvme disk.

Since you are using xfs for the root fs - I think it is safe to assume the journal entries regarding btrfs and raid6 is completely unrelated to nvme0n1.

The reason mcedit is puking on you is likely becuase the editor’s program files is located where the nvme has read errors.

You may be able to work around it by renaming mc binaries then resyncing the package thus placing the binaries at other sectors of the disk.

[zdebel]

Good idea man, but still the same issue I also tried copying mc to my user dir, but executing it causes that error. Nothing else causes these errors, and a lot of stuff is happening on my server.

lsblk output here, but like I said I only have a single XFS partition.

[zdebel@zdebel-um350 bin]$ lsblk -f
NAME        FSTYPE FSVER LABEL UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
nvme0n1
├─nvme0n1p1 vfat   FAT32       AB20-1D79                             299,1M     0% /boot/efi
└─nvme0n1p2 xfs                06991fb8-c25b-4564-a30d-aa8b09f0ec25  117,1G    75% /

All of the problems started when I upgraded to 6.12. I can’t find an easy way to downgrade to 6.10.

[andreas85]

@zdebel

Good idea man, but still the same issue I also tried copying mc to my user dir, but executing it causes that error.

Either mc or mcedit (which is used by mc) or one of the libraries required by mc or mcedit is corrupted as a file.

• You could ask your file system whether it finds any(!!!) corrupted files (with btrfs it is scrub, with xfs ???)
• You could examine the coredumps to see if the name of a library appears (in the call path)
• You could manually examine which libraries are used by mc (recursively)

[zdebel]

Ok I’ve used dd and tried reading from the LBA address dmesg shows, indeed it is NVME errors. Thank you all for your help.