btrfs problem on 6.7.4

Since updating to 6.7.4 I've been getting a stack trace like this every 7-10 days. I can log in but the prompt usually hangs and I have to power cycle the box. Stack trace has some minor differences but it always starts with do_fsync and ends with warn_slowpath_common.

Jul 12 23:08:26 warehouse6 kernel: Call Trace:
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff883d3e98>] dump_stack+0x4d/0x65
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff8806b27c>] warn_slowpath_common+0x7c/0xb0
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff8806b365>] warn_slowpath_null+0x15/0x20
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff882d40ef>] btrfs_free_extent+0x11f/0x130
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88309f3e>] ? btrfs_get_token_64+0xee/0x110
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88304f18>] __btrfs_drop_extents+0xaa8/0xed0
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff8833255b>] btrfs_log_changed_extents+0x2ab/0x610
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88331045>] ? fill_inode_item.isra.19+0x215/0x260
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88313306>] ? free_extent_buffer+0x46/0x80
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88333165>] btrfs_log_inode+0x8a5/0xf00
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88310a3d>] ? extent_write_cache_pages.isra.43.constprop.58+0x21d/0x3b0
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88333a56>] btrfs_log_inode_parent+0x216/0x960
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff882ec51d>] ? wait_current_trans.isra.21+0x1d/0xf0
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff8812e185>] ? kmem_cache_alloc+0x135/0x150
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff882ebe10>] ? join_transaction.isra.15+0x20/0x400
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff8833519d>] btrfs_log_dentry_safe+0x5d/0x80
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff883074a6>] btrfs_sync_file+0x206/0x340
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88165226>] vfs_fsync_range+0x36/0xa0
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff881652e8>] do_fsync+0x38/0x60
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88078061>] ? SyS_rt_sigprocmask+0x81/0xa0
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff8816553b>] SyS_fsync+0xb/0x10
Jul 12 23:08:26 warehouse6 kernel: [<ffffffff88a50617>] entry_SYSCALL_64_fastpath+0x12/0x6a

I seem to have a few bursts of these some time before the crash, but I'm not sure if they are related because I see a few of them at times when there is no problem:

Jul 12 22:42:44 warehouse6 connmand[3074]: eth0 {newlink} ignoring slave, index 2 address 00:1F:33:EA:BA:20
Jul 12 22:45:46 warehouse6 connmand[3074]: eth0 {newlink} ignoring slave, index 2 address 00:1F:33:EA:BA:20

I'm not sure what these latter entries mean. Any ideas?

steve