Kernel 6.8.5: Cannot mount even a blank ext4 image

Support Kernel
1

After upgrading to 6.8.5, mount command always fails for me, even if with a new blank ext4 image.

Reproduce:

$ dd if=/dev/zero of=a.img bs=1M count=100
$ mkfs.ext4 a.img
$ mkdir mp
$ sudo mount a.img mp
mount: /path/to/mp: 
 /dev/loop0 already mounted or mount point busy.
       dmesg(1) may have more information after failed mount system call.
$ echo $?
32

dmesg is: (only this line added after mount)

[  622.633122] loop0: detected capacity change from 0 to 204800

In kernel 6.6.26-1 LTS, mount won’t fail.

2

Did you reboot after the kernel change?

Do you have snap installed? It will mount to these loop points.

(I can run the above commands without issues on 6.9.0-rc5. Doesn’t help you, though, right?)

3

As hinted to in your other thread:
I can do this no problem with kernel66 or kernel515
But with kernel68 it does end in this error for me, too.
… something with that kernel version ain’t right - I’d avoid it

3 Likes
4

Agreed on that one. Doesn’t want to play nicely on my T440.

5

Just for the troubleshooting - I cannot reproduce - with the given commands

 $ uname -a
Linux tiger 6.8.7-1-MANJARO #1 SMP PREEMPT_DYNAMIC Wed Apr 17 13:26:06 UTC 2024 x86_64 GNU/Linux

 $ lsblk
NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINTS
loop0         7:0    0   100M  0 loop /home/fh/mp
[...]
6

I rebooted and haven’t installed snap.

It would help to know that you can switch to other version to avoid this problem (though I knew this before :wink: thanks to @Nachlese ).

I just see existing bug reports of kernel 6.8.5 are about ntfs mount error, so I report it, a simple ext4 mount error. Hopefully it will help others.

(In fact, I met this problem by mounting an erofs image originally. And I found that it’s a mount error, neither an unsupported filesystem error nor a broken image file).

7

I tested it also on my system with kernel 6.8.5-1 with same behavior.

@linux-aarhus uses 6.8.7-1 which is in “testing”-branch now … so maybe you still need to wait for a little bit. :grinning:

8

Linking to the OP’s previous thread, for possible context.

9

Isn’t the underlaying problem the same as here?

So loop device is reported as free, but some phantom device exists?
Check if /dev/loop0 file exists with

ls -lah /dev/loop0

And then check what losetup reports, so:

losetup -f

So if /dev/loop0 exists, but losetup -f reports /dev/loop0 as free, than the problem is the same as with snapd.

10

but that is quite another matter - you topic title says ext4?

ntfs is a a foreign filesystem - there is not gurarantee that is doable - especially if the ntfs filesystem is not clealy unmounted in windows - hibernation or windows fast startup - don’t rely on ntfs with Linux - warning - icebergs ahead …

11

The output is different:

$ ls -lah /dev/loop0
ls: cannot access '/dev/loop0': No such file or directory
$ sudo ls -lah /dev/loop0
ls: cannot access '/dev/loop0': No such file or directory
$ losetup -f
losetup: cannot find an unused loop device: Permission denied
$ sudo losetup -f
/dev/loop0
$ sudo mount a.img mp # the first mount since booted
mount: /path/to/mp: /dev/loop0 already mounted or mount point busy.
       dmesg(1) may have more information after failed mount system call.
$ ls -lah /dev/loop0
brw-rw---- 1 root disk 7, 0 Apr 23 09:31 /dev/loop0
$ losetup -f
/dev/loop0