zfs-mount fails with 0.6.5.4-r2 on gentoo

[lkastner]

I upgraded from zfs-0.6.4.2 to zfs-0.6.5.4-r2 recently. The zfs-import service works fine, but zfs-mount fails with the following output in dmesg:

[ 71.451369] SPL: using hostid 0x00000000
[ 92.240904] BUG: unable to handle kernel paging request at 0000004fa0036c7c
[ 92.241055] IP: [] zfs_prop_get_type+0xe/0x20 [zcommon]
[ 92.241163] PGD 0
[ 92.241256] Oops: 0000 [#1] SMP
[ 92.241394] Modules linked in: zfs(PO) zunicode(PO) zcommon(PO) znvpair(PO) spl(O) zavl(PO)
[ 92.241771] CPU: 1 PID: 3742 Comm: dump_list_strat Tainted: P O 4.1.15-gentoo-r1 #3
[ 92.241829] Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./Q1900-ITX, BIOS P1.00 03/28/2014
[ 92.241889] task: ffff88015645f7d0 ti: ffff880083d04000 task.ti: ffff880083d04000
[ 92.241944] RIP: 0010:[] [] zfs_prop_get_type+0xe/0x20 [zcommon]
[ 92.242052] RSP: 0018:ffff880083d07588 EFLAGS: 00010202
[ 92.242104] RAX: 0000004fffffffb0 RBX: ffff880083d07728 RCX: 0000000000000000
[ 92.242159] RDX: 0000000000000018 RSI: 0000000000000000 RDI: 00000000ffffffff
[ 92.242214] RBP: ffff880083d07588 R08: 000000000000006f R09: ffff880084559b80
[ 92.242269] R10: 0000000000000000 R11: ffff88015af17500 R12: 0000000000000010
[ 92.242324] R13: ffff880083d07868 R14: 00000000ffffffff R15: 0000000000000000
[ 92.242380] FS: 0000000000000000(0000) GS:ffff88015fc80000(0000) knlGS:0000000000000000
[ 92.242436] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 92.242489] CR2: 0000004fa0036c7c CR3: 0000000001ce3000 CR4: 00000000001006e0
[ 92.242554] Stack:
[ 92.242611] ffff880083d07858 ffffffffa00cefe0 ffff880083d075a8 ffffffff81963449
[ 92.242857] 0000000000000035 ffff880156fd6800 ffff880083d07868 ffff880084559f00
[ 92.243102] ffff880084559b80 0000000000000246 ffff880156fd6800 ffff8800892a8a00
[ 92.243347] Call Trace:
[ 92.243430] [] dsl_prop_inherit+0x6c0/0xa40 [zfs]
[ 92.243499] [] ? mutex_unlock+0x9/0x10
[ 92.243566] [] ? spl_kmem_alloc+0xa5/0x1a0 [spl]
[ 92.243633] [] ? zprop_string_to_index+0xc0/0x190 [zcommon]
[ 92.243701] [] ? spl_kmem_free+0x25/0x40 [spl]
[ 92.243767] [] ? zprop_iter_common+0xd1/0x130 [zcommon]
[ 92.243835] [] ? zprop_name_to_prop+0x37/0xb0 [zcommon]
[ 92.243903] [] ? strfree+0x9/0x10 [spl]
[ 92.243968] [] ? mutex_unlock+0x9/0x10
[ 92.244050] [] dsl_prop_inherit+0x985/0xa40 [zfs]
[ 92.244126] [] ? zrl_init+0x43/0x50 [zfs]
[ 92.244207] [] ? dsl_dataset_fast_stat+0xad/0x170 [zfs]
[ 92.244291] [] ? dsl_prop_unset_hasrecvd+0x60/0x60 [zfs]
[ 92.244375] [] dsl_prop_get_all_new+0x11/0x2790 [zfs]
[ 92.244456] [] zfs_secpolicy_smb_acl+0xdd/0x5ea0 [zfs]
[ 92.244537] [] dataset_name_hidden+0x1ff/0x370 [zfs]
[ 92.244604] [] ? mutex_unlock+0x9/0x10
[ 92.244687] [] ? unique_insert+0x11a/0x140 [zfs]
[ 92.244768] [] ? dsl_dataset_hold_obj+0x583/0xbf0 [zfs]
[ 92.244848] [] dmu_objset_stats+0x16c/0xbd0 [zfs]
[ 92.244915] [] ? spl_kmem_alloc+0xa5/0x1a0 [spl]
[ 92.244994] [] dmu_objset_stats+0x2d9/0xbd0 [zfs]
[ 92.245073] [] dmu_objset_find_dp+0x150/0x1d0 [zfs]
[ 92.245154] [] ? dataset_name_hidden+0x50/0x370 [zfs]
[ 92.245234] [] dump_list_strategy_one+0x5f/0xd0 [zfs]
[ 92.245315] [] dump_list_strategy_impl+0xac/0x340 [zfs]
[ 92.245382] [] ? put_prev_task_fair+0x24/0x40
[ 92.245462] [] ? dump_list_strategy_impl+0x340/0x340 [zfs]
[ 92.245542] [] dump_list_strategy+0x19/0xa0 [zfs]
[ 92.245610] [] __thread_exit+0x8c/0xa0 [spl]
[ 92.245677] [] ? __thread_exit+0x20/0xa0 [spl]
[ 92.245743] [] kthread+0xc4/0xe0
[ 92.245807] [] ? kthread_create_on_node+0x170/0x170
[ 92.245873] [] ret_from_fork+0x42/0x70
[ 92.245937] [] ? kthread_create_on_node+0x170/0x170
[ 92.246000] Code: 04 83 b8 e0 6c 03 a0 01 0f 95 c0 0f b6 c0 c3 66 0f 1f 44 00 00 31 c0 5d c3 0f 1f 40 00 89 ff 55 48 8d 04 bf 48 89 e5 48 c1 e0 04 <8b> 80 cc 6c 03 a0 5d c3 66 2e 0f 1f 84 00 00 00 00 00 89 ff 55
[ 92.249026] RIP [] zfs_prop_get_type+0xe/0x20 [zcommon]
[ 92.249139] RSP
[ 92.250263] CR2: 0000004fa0036c7c
[ 92.250324] ---[ end trace 99ad4540909a6ff6 ]---

Afterwards it is locked up. Is this the same issue as #4318? Since it worked fine with 0.6.4.2, is there some kernel flag that needs to be included now?

[behlendorf]

Can you check if this is reproducible with a manual build of the ZoL 0.6.5.6 tarballs, not the gentoo version. Adding @ryao to this issue.

[drescherjm]

If you install the zfs-9999, spl-9999, zfs-kmod-9999 you should get the current master without the new stable API patches that are in some of the other ebuilds.

[lkastner]

Thanks. It works perfectly with the 0.6.5.6 tarballs, meaning I cannot reproduce. I will try the 9999 version on the weekend. Is this due to the higher version or the non-ebuild version? I have another machine which works just fine with the 0.6.5.4-r2 ebuilds.

[lkastner]

I will close the issue now, as it works for me. If I manage to reproduce, I can reopen. Thank you for your help.

[drescherjm]

9999 usually means live (what is in the source control - master, trunk ...) ebuild of a repository. In this case I believe you get what is in the github master branch at the time you emerge.