Kernel BUG when unmounting dm-cached device

[bvr...@gmail.com]

Hi,

I created a 100G cache backend test volume on LVS, called cachetest. I then created metadata and data devices on two separate partitions of a 200G ssd with the command

dmsetup create cached --table '0 209715200 cache /dev/sde3 /dev/sde2 /dev/mapper/vla-cachetest 512 1 writeback default 0' . Created and ext4 filesystem, mounted the device and tested it for several days, everything seemed to work as expected. When unmounting, I got the following:

[641134.067306] EXT4-fs (dm-7): sb orphan head is 12
[641134.067342] sb_info orphan list:
[641134.067370]   inode dm-7:12 at ffff8800439aac90: mode 100644, nlink 0, next 14
[641134.067425]   inode dm-7:14 at ffff88003cb10458: mode 100644, nlink 0, next 13
[641134.067480]   inode dm-7:13 at ffff8800450bc0b0: mode 100644, nlink 0, next 6291458
[641134.067536]   inode dm-7:6291458 at ffff880058484f50: mode 100644, nlink 0, next 5636098
[641134.067593]   inode dm-7:5636098 at ffff88000ea840b0: mode 100644, nlink 0, next 5767170
[641134.067649]   inode dm-7:5767170 at ffff8800532b0ba8: mode 100644, nlink 0, next 3670030
[641134.067706]   inode dm-7:3670030 at ffff88000eabc800: mode 100644, nlink 0, next 0
[641134.067789] ------------[ cut here ]------------
[641134.067820] kernel BUG at fs/ext4/super.c:804!
[641134.067851] invalid opcode: 0000 [#1] SMP
[641134.067882] Modules linked in: dm_cache_mq dm_cache dm_bio_prison dm_persistent_data dm_bufio crc32c libcrc32c cbc ipt_MASQUERADE xt_nat iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack xt_tcpudp iptable_filter ip_tables x_tables nfsd nfs_acl exportfs auth_rpcgss oid_registry nfs lockd sunrpc 8021q dm_crypt raid456 async_pq async_xor xor async_memcpy async_raid6_recov raid6_pq async_tx loop md_mod arc4 iwldvm mac80211 evdev coretemp acpi_cpufreq mperf iwlwifi serio_raw pcspkr i2c_i801 lpc_ich cfg80211 ehci_pci mfd_core processor i2c_core rfkill button ext4 jbd2 mbcache crc16 dm_mod sd_mod crc_t10dif ahci libahci thermal fan thermal_sys libata scsi_mod r8169 mii uhci_hcd ehci_hcd
[641134.068324] CPU: 0 PID: 13853 Comm: umount Tainted: G        W    3.10.0 #5
[641134.068360] Hardware name: OEM OEM/G43-F71862, BIOS 6.00 PG 12/18/2009
[641134.068396] task: ffff880037fc0b60 ti: ffff880022d72000 task.ti: ffff880022d72000
[641134.068450] RIP: 0010:[<ffffffffa0110b04>]  [<ffffffffa0110b04>] ext4_put_super+0x22e/0x2e8 [ext4]
[641134.068522] RSP: 0018:ffff880022d73e78  EFLAGS: 00010283
[641134.068555] RAX: 0000000000000046 RBX: ffff8800524de000 RCX: 0000000000000007
[641134.068608] RDX: 0000000000000006 RSI: 0000000000000046 RDI: 0000000000000046
[641134.068662] RBP: ffff8800152f0400 R08: 0000000000000000 R09: 0000000000000000
[641134.068716] R10: 0000000000000000 R11: 0000000000000000 R12: ffff8800524de200
[641134.068769] R13: ffff8800524de200 R14: ffff8800524de190 R15: ffff880022d73f38
[641134.068824] FS:  00007f1c12d0a740(0000) GS:ffff880077a00000(0000) knlGS:0000000000000000
[641134.068880] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[641134.068914] CR2: 00007fc309b54e02 CR3: 0000000071bc4000 CR4: 00000000000007f0
[641134.068967] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[641134.069021] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[641134.069074] Stack:
[641134.069098]  ffff880000000000 ffff880022d73e78 ffff8800152f0400 ffff8800152f04a0
[641134.069157]  ffffffffa012dcf0 ffff88003797e120 ffff88003797e100 ffffffff810a8988
[641134.069215]  ffff880074c249c0 0000000000000083 ffff88003797e100 ffffffff810a8a1e
[641134.069274] Call Trace:
[641134.069305]  [<ffffffff810a8988>] ? generic_shutdown_super+0x4d/0xc5
[641134.069342]  [<ffffffff810a8a1e>] ? kill_block_super+0x1e/0x5f
[641134.069378]  [<ffffffff810a8d23>] ? deactivate_locked_super+0x1b/0x46
[641134.069415]  [<ffffffff810bdbac>] ? SyS_umount+0x2d0/0x2f1
[641134.069451]  [<ffffffff812a5296>] ? system_call_fastpath+0x1a/0x1f
[641134.069486] Code: 45 0f b7 44 24 38 49 8b 54 24 78 48 81 c6 20 03 00 00 89 04 24 31 c0 e8 d7 e7 18 e1 4d 8b 24 24 4d 39 ec 0f 84 6e ff ff ff eb b7 <0f> 0b 48 8b bd 20 01 00 00 e8 de bd fb e0 48 8b bb 50 02 00 00
[641134.069663] RIP  [<ffffffffa0110b04>] ext4_put_super+0x22e/0x2e8 [ext4]
[641134.069708]  RSP <ffff880022d73e78>
[641134.070022] ---[ end trace d9e241314b707d3b ]---

it is no longer possible to mount the underlying device (cachetest). Should I test anything else before rebooting and not putting dm-cache into production?

[Mario Salisbury]

Hello,
It appears that you are using the built-in kernel 3.9 dm-cache, which is different from our dm-cache module.  However, when working with the 3.9 dm-cache, I was unable to recreate the error you're experiencing.  The issue also looks more file-system related than directly related to the dm-cache module.  The link below is a similar issue to the one you are faced with.  Hope this helps
http://www.spinics.net/lists/linux-ext4/msg38318.html
- Mario