blktests loop/007 failed

[yizhanglinux]

Just found this failure when I ran full blktests on the latest Fedora distro.

# ./check loop/007
loop/007 (update loop device capacity with filesystem)       [failed]
    runtime  0.047s  ...  0.048s
    --- tests/loop/007.out	2026-03-18 12:41:50.196296917 -0400
    +++ /root/blktests/results/nodev/loop/007.out.bad	2026-03-18 12:43:25.082783568 -0400
    @@ -1,2 +1,5 @@
     Running loop/007
    +mount: /root/blktests/results/tmpdir.loop.007.fym/mnt: fsconfig() failed: Structure needs cleaning.
    +       dmesg(1) may have more information after failed mount system call.
    +umount: /root/blktests/results/tmpdir.loop.007.fym/mnt: not mounted.
     Test complete

# cat   /root/blktests/results/nodev/loop/007.out.bad
Running loop/007
mount: /root/blktests/results/tmpdir.loop.007.fym/mnt: fsconfig() failed: Structure needs cleaning.
       dmesg(1) may have more information after failed mount system call.
umount: /root/blktests/results/tmpdir.loop.007.fym/mnt: not mounted.
Test complete

# dmesg
[115105.866402] run blktests loop/007 at 2026-03-18 12:43:25
[115105.881017] loop0: detected capacity change from 0 to 2097152
[115105.904291] EXT4-fs (loop0): orphan file too big: 1048576
[115105.910447] EXT4-fs (loop0): mount failed

# uname -r
7.0.0-0.rc3.260313g0257f64bdac7.33.fc45.x86_64

[yizhanglinux]

Just found CKI also reported this failure recently.

[yizhanglinux]

Seems it's one regression which introduced from e2fsprogs-1.47.4, I've filed one issue for it.

https://lore.kernel.org/linux-ext4/CAHj4cs8o5HExeOX2YCYrWH86iz4xwL52YtsJNV2YOyDPbyEffg@mail.gmail.com/