Re: [BUG] INFO: task hung in nilfs_segctor_thread (fs/nilfs2/segment.c:357, v7.4-rc3)
From: Ryusuke Konishi
Date: Thu Nov 19 2026 - 09:21:39 EST
On Thu, Nov 19, 2026 at 11:07 PM Zheng Zheng wrote:
>
> Hi Ryusuke, Viacheslav,
>
> We found a reproducible NILFS2 segment-constructor hang on v7.4-rc3 with
> syzkaller. The supplied recipe reproduced on three independent fresh VMs; no
> state snapshot is required.
>
> Repository and reproducer:
>
> https://github.com/Zheng-zheng-codes/nilfs2-segctor-thread-hang
> commit 2d94e07741d28d2dccaba174d06d6a5255ce0134
>
> Hung-task report
> ----------------
>
> INFO: task segctord:35823 blocked for more than 183 seconds.
> task:segctord state:D stack:29568 pid:35823
> Call Trace:
> <TASK>
> __schedule
> schedule_preempt_disabled
> rwsem_down_write_slowpath
> __down_write
> down_write
> nilfs_transaction_lock+0x25d/0x410 fs/nilfs2/segment.c:357
> nilfs_segctor_thread_construct fs/nilfs2/segment.c:2601 [inline]
> nilfs_segctor_thread+0x600/0xa60 fs/nilfs2/segment.c:2716
> kthread
> ret_from_fork
> ret_from_fork_asm
> </TASK>
>
> The blocked thread is waiting to acquire `nilfs->ns_segctor_sem`
> for writing in `nilfs_transaction_lock()`:
>
> for (;;) {
> down_write(&nilfs->ns_segctor_sem);
> if (!test_bit(NILFS_SC_PRIOR_FLUSH, &sci->sc_flags))
> break;
>
> nilfs_segctor_do_immediate_flush(sci);
> up_write(&nilfs->ns_segctor_sem);
> cond_resched();
> }
>
> The lock dump also associates `&nilfs->ns_segctor_sem` with a task in
> `nilfs_ioctl_clean_segments()` holding `sb_writers`, while `segctord` is
> blocked in `down_write()` on that semaphore. The serial log repeatedly reports:
>
> NILFS (loop6): error -28 cleaning segments
>
> i.e. `-ENOSPC` from the clean-segments path. The reason why the
> clean-segments ioctl holds the semaphore
> for an extended period is not yet clear. We have not established
> a circular lock dependency or identified the root cause.
> We would appreciate your analysis.
>
> Reproduction
> ------------
>
> git clone https://github.com/Zheng-zheng-codes/nilfs2-segctor-thread-hang
> cd nilfs2-segctor-thread-hang
>
> # boot v7.4-rc3 with the supplied kernel config
> # using a Debian trixie x86_64 guest
>
> # in the guest, as root:
> syz-execprog -procs=8 -threaded -repeat=0 \
> -executor ./syz-executor reproducer/repro.kernsight.prog
>
> `-repeat=0` enables indefinite repetition. The recipe has 227 entries: 226
> replay programs followed by the testcase.
>
> Verification:
>
> - Main KernSight pipeline: attempt 2, `confirmed=true`,
> `status=verified`, 226 replay programs, 2026-11-19T11:10:28Z
> - Pipeline verification scope: `snapshot_restore`
> - Independent external fresh-VM verification: **3/3**
> - run 1 hit after 422 s
> - run 2 hit after 257 s
> - run 3 hit after 320 s
> - External serial logs are in `external-verify/serial-{1,2,3}.log`
> - Runs were counted only when a `nilfs_transaction_lock` frame appeared next
> to `segctord` or `nilfs_segctor_thread`; checker:
> `external-verify/verify-logs.py`
> - C reproducer: not generated; the syzkaller recipe is the primary reproducer
>
> Kernel/guest details:
>
> kernel: v7.4-rc3, bzImage sha256
> 0c16409f8aec508ac44a63e631c2bd352536401f3fd0856df1d839fb953a1882
> guest: Debian trixie x86_64, 6 GiB RAM, 4 vCPU, QEMU/KVM
> boot: oops=panic panic_on_warn=1 hung_task_panic=0
> config: CONFIG_KASAN=y, CONFIG_PROVE_LOCKING=y, CONFIG_NILFS2_FS=y
>
> Thanks,
> Zheng
Hi Zheng Zheng,
Thank you for the feedback.
I have filed this issue at:
https://github.com/nilfs-dev/nilfs2/issues/182
The issue you reported earlier was related to GC error handling, and
Wang Jianjian has just proposed a patch for it, which I was about to
look into.
Since this one also appears to be a GC error handling issue, I will
look into both together.
Thanks,
Ryusuke Konishi