Re: [PATCH V2 3/3] block: model freeze & enter queue as lock for supporting lockdep

[Ming Lei <ming.lei@redhat.com>]

On Mon, Jan 13, 2025 at 02:39:22PM +0000, Chris Bainbridge wrote:
> Hi,
> 
> With latest mainline 6.13-rc6, I have been getting intermittent lock
> warnings when using a btrfs filesystem. The warnings bisect to this
> commit:
> 
> f1be1788a32e8fa63416ad4518bbd1a85a825c9d is the first bad commit
> commit f1be1788a32e8fa63416ad4518bbd1a85a825c9d
> Author: Ming Lei <ming.lei@redhat.com>
> Date:   Fri Oct 25 08:37:20 2024 +0800
> 
>     block: model freeze & enter queue as lock for supporting lockdep
> 
> 
> On my system, these lockdep warnings are reproducible just by doing some
> large fs operation, like copying the whole linux kernel git repo to the
> btrfs filesystem.
> 
> The lockdep warning is:
> 
> [  437.745808] ======================================================
> [  437.745810] WARNING: possible circular locking dependency detected
> [  437.745811] 6.13.0-rc6-00037-gac70f027bab6 #112 Not tainted
> [  437.745813] ------------------------------------------------------
> [  437.745814] kswapd0/141 is trying to acquire lock:
> [  437.745815] ffff925c11095e90 (&delayed_node->mutex){+.+.}-{4:4}, at: __btrfs_release_delayed_node.part.0+0x3f/0x280 [btrfs]
> [  437.745862]
>                but task is already holding lock:
> [  437.745863] ffffffffb9cc8c80 (fs_reclaim){+.+.}-{0:0}, at: balance_pgdat+0x578/0xa80
> [  437.745869]
>                which lock already depends on the new lock.
> 
> [  437.745870]
>                the existing dependency chain (in reverse order) is:
> [  437.745871]
>                -> #3 (fs_reclaim){+.+.}-{0:0}:
> [  437.745873]        fs_reclaim_acquire+0xbd/0xf0
> [  437.745877]        __kmalloc_node_noprof+0xa1/0x4f0
> [  437.745880]        __kvmalloc_node_noprof+0x24/0x100
> [  437.745881]        sbitmap_init_node+0x98/0x240
> [  437.745885]        scsi_realloc_sdev_budget_map+0xdd/0x1d0
> [  437.745889]        scsi_add_lun+0x458/0x760
> [  437.745891]        scsi_probe_and_add_lun+0x15e/0x480
> [  437.745892]        __scsi_scan_target+0xfb/0x230
> [  437.745893]        scsi_scan_channel+0x65/0xc0
> [  437.745894]        scsi_scan_host_selected+0xfb/0x160
> [  437.745896]        do_scsi_scan_host+0x9d/0xb0
> [  437.745897]        do_scan_async+0x1c/0x1a0
> [  437.745898]        async_run_entry_fn+0x2d/0x120
> [  437.745901]        process_one_work+0x210/0x730
> [  437.745903]        worker_thread+0x193/0x350
> [  437.745905]        kthread+0xf3/0x120
> [  437.745906]        ret_from_fork+0x40/0x70
> [  437.745910]        ret_from_fork_asm+0x11/0x20
> [  437.745912]
>                -> #2 (&q->q_usage_counter(io)#10){++++}-{0:0}:

Hello,

This one has been solved in for-6.14/block:

https://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git/log/?h=for-6.14/block

Thanks,
Ming