On 06/13/2013 03:10 PM, Eric Sandeen wrote:
On 6/13/13 8:01 AM, Torbjørn wrote:
Hi,
I have a 8 drive md-raid6 + dm-crypt with xfs on top.
When trying to mount using 3.10-rc5 (ubuntu mainline ppa) I get the following kernel bug:
[ 1017.056091] SGI XFS with ACLs, security attributes, realtime, large block/inode numbers, no debug enabled
[ 1017.057607] XFS (dm-11): Mounting Filesystem
[ 1017.195409] ------------[ cut here ]------------
[ 1017.195881] Kernel BUG at ffffffff81485fb2 [verbose debug info unavailable]
Hm that's not so helpful :( So we don't have thread info or
line number information.
[ 1017.196603] invalid opcode: 0000 [#1] SMP
[ 1017.197050] Modules linked in: xfs vhost_net macvtap macvlan ip6table_filter ip6_tables ebtable_nat ebtables ipt_MASQUERADE ipt_REJECT xt_CHECKSUM sch_prio bridge stp llc xt_state iptable_filter dm_crypt xt_CLASSIFY xt_tcpudp xt_DSCP iptable_mangle iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack ip_tables x_tables intel_powerclamp kvm_intel kvm psmouse serio_raw microcode ppdev lpc_ich mac_hid parport_pc w83627ehf hwmon_vid coretemp nfsd lp nfs_acl auth_rpcgss nfs parport fscache lockd sunrpc btrfs zlib_deflate libcrc32c raid1 raid0 multipath linear raid456 async_pq async_xor xor async_memcpy async_raid6_recov raid6_pq async_tx raid10 hid_generic usbhid hid ast ttm crc32_pclmul drm_kms_helper ghash_clmulni_intel drm aesni_intel ablk_helper cryptd lrw gf128mul glue_helper e1000e mpt2sas i2c_algo_bit ptp sysimgblt sysfillrect pps_core ahci aes_x86_64 syscopyarea scsi_transport_sas libahci raid_class video
[ 1017.206695] CPU: 1 PID: 486 Comm: md0_raid6 Not tainted 3.10.0-031000rc5-generic #201306082135
[ 1017.207603] Hardware name: To be filled by O.E.M. To be filled by O.E.M./P8B-X series, BIOS 2107 05/04/2012
[ 1017.208681] task: ffff88040e509770 ti: ffff88040de2a000 task.ti: ffff88040de2a000
[ 1017.209498] RIP: 0010:[<ffffffff81485fb2>] [<ffffffff81485fb2>] scsi_setup_fs_cmnd.part.32+0x82/0x90
so it crashed in scsi, and nothing in the stack is from xfs.
Barring weird interactions, I think you need to look elsewhere for the bug;
this doesn't look like an xfs problem to me.
Actually,
https://lkml.org/lkml/2013/6/12/440 looks relevant, which references
https://lkml.org/lkml/2013/5/19/75
Guessing this is an md bug.
-Eric
[ 1017.210467] RSP: 0018:ffff88040de2bb68 EFLAGS: 00010046
[ 1017.211021] RAX: 0000000000000000 RBX: ffff8804106d4800 RCX: 0000000000000002
[ 1017.211772] RDX: 0000000000001000 RSI: ffff8803d3b89028 RDI: ffff8804106d4800
[ 1017.212521] RBP: ffff88040de2bb78 R08: ffff8803d3b88f30 R09: ffff9ef774422900
[ 1017.213300] R10: 0000000018422880 R11: 00000000ffffffff R12: ffff8803d3b89028
[ 1017.214054] R13: 0000000000000001 R14: ffff8804106d4800 R15: ffff88041032a800
[ 1017.214802] FS: 0000000000000000(0000) GS:ffff88042fc40000(0000) knlGS:0000000000000000
[ 1017.215691] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1017.216287] CR2: 00007f9eaa5a4000 CR3: 0000000001c0c000 CR4: 00000000001427e0
[ 1017.217069] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 1017.217819] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 1017.218568] Stack:
[ 1017.218769] ffff8804106d4800 ffff8803d3b89028 ffff88040de2bb98 ffffffff81485fef
[ 1017.219577] ffff8803d3b89028 ffff8804106ac100 ffff88040de2bc08 ffffffff81496c0f
[ 1017.220379] ffff88040de2bbd8 ffffffff81328aa5 ffff880400001000 0000000018422880
[ 1017.221176] Call Trace:
[ 1017.221415] [<ffffffff81485fef>] scsi_setup_fs_cmnd+0x2f/0x40
[ 1017.222024] [<ffffffff81496c0f>] sd_prep_fn+0xff/0xb00
[ 1017.222567] [<ffffffff81328aa5>] ? deadline_remove_request.isra.3+0x55/0x90
[ 1017.223336] [<ffffffff81310d0e>] blk_peek_request+0xfe/0x270
[ 1017.223953] [<ffffffff8148588f>] scsi_request_fn+0x4f/0x430
[ 1017.224546] [<ffffffff8130b757>] __blk_run_queue+0x37/0x50
[ 1017.225145] [<ffffffff8130d9fd>] queue_unplugged+0x3d/0xc0
[ 1017.225723] [<ffffffff81311203>] blk_flush_plug_list+0x183/0x210
[ 1017.226360] [<ffffffff813112a8>] blk_finish_plug+0x18/0x50
[ 1017.226943] [<ffffffffa0148497>] raid5d+0x1b7/0x1d0 [raid456]
[ 1017.227548] [<ffffffff8153b66d>] md_thread+0x11d/0x170
[ 1017.228090] [<ffffffff8106c070>] ? add_wait_queue+0x60/0x60
[ 1017.228681] [<ffffffff8153b550>] ? md_rdev_init+0x110/0x110
[ 1017.229274] [<ffffffff8106b8b0>] kthread+0xc0/0xd0
[ 1017.229795] [<ffffffff8106b7f0>] ? flush_kthread_worker+0xb0/0xb0
[ 1017.230468] [<ffffffff816d545c>] ret_from_fork+0x7c/0xb0
[ 1017.231048] [<ffffffff8106b7f0>] ? flush_kthread_worker+0xb0/0xb0
[ 1017.231719] Code: fd ff ff 5b 41 5c 5d c3 48 8b 00 48 85 c0 74 b7 48 8b 40 48 48 85 c0 74 ae ff d0 85 c0 74 a8 eb e2 b8 02 00 00 00 0f 1f 00 eb d8 <0f> 0b 66 66 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 55 48
[ 1017.234403] RIP [<ffffffff81485fb2>] scsi_setup_fs_cmnd.part.32+0x82/0x90
[ 1017.235121] RSP <ffff88040de2bb68>
[ 1017.482522] ---[ end trace fa18c0d8cd90bd2f ]---
3.10-rc4 has the same issue. I have not tried any earlier 3.10 kernels
The system mounts fine using 3.9.5 (also ubuntu ppa)
If I can provide any other info to help, please let me know.
--
Torbjørn
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
Hi,
Thanks for the insight Eric.
I'll compile a kernel with proper debug info, and see if linux-raid can
make any use of it.
--
Torbjørn
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
