[LU-7164] os_extent should hold a refcount to osc_object

Tue, 24 Nov 2015 14:45:30 +0000
Otherwise, it may end up with kernel crash at osc_extent destroying path where osc_object is already destroyed.
The stack trace is as follows:
<3>[15142.816048] LustreError: 0:0:(ldlm_lockd.c:342:waiting_locks_callback()) ### lock callback timer expired after 101s: evicting client at 0@lo  ns: filter-lustre-OST0001_UUID lock: ffff880084165db8/0x3551b72d22b6ecb0 lrc: 3/0,0 mode: PW/PW res: [0xc:0x0:0x0].0 rrc: 2 type: EXT [0->18446744073709551615] (req 0->4095) flags: 0x60000000000020 nid: 0@lo remote: 0x3551b72d22b6eca9 expref: 5 pid: 8119 timeout: 4298677961 lvb_type: 0
<3>[15142.819557] LustreError: 0:0:(ldlm_lockd.c:342:waiting_locks_callback()) Skipped 1 previous similar message
<1>[15143.358700] BUG: unable to handle kernel paging request at ffff8800afc9ff54
<1>[15143.360803] IP: [<ffffffff812a075c>] _raw_spin_unlock+0xc/0xa0
<4>[15143.361349] PGD 1a2e063 PUD 501067 PMD 680067 PTE 80000000afc9f060
<4>[15143.361813] Oops: 0000 [#1] SMP DEBUG_PAGEALLOC
<4>[15143.361813] last sysfs file: /sys/devices/system/cpu/possible
<4>[15143.361813] CPU 1 
<4>[15143.361813] Modules linked in: lustre ofd osp lod ost mdt mdd mgs osd_ldiskfs ldiskfs lquota lfsck obdecho mgc lov osc mdc lmv fid fld ptlrpc obdclass ksocklnd lnet libcfs exportfs jbd sha512_generic sha256_generic ext4 jbd2 mbcache virtio_balloon virtio_console i2c_piix4 i2c_core virtio_blk virtio_net virtio_pci virtio_ring virtio pata_acpi ata_generic ata_piix dm_mirror dm_region_hash dm_log dm_mod nfs lockd fscache auth_rpcgss nfs_acl sunrpc be2iscsi bnx2i cnic uio cxgb3i libcxgbi ipv6 cxgb3 mdio libiscsi_tcp qla4xxx iscsi_boot_sysfs libiscsi scsi_transport_iscsi [last unloaded: libcfs]
<4>[15143.361813] 
<4>[15143.361813] Pid: 5500, comm: ptlrpcd_00_03 Not tainted 2.6.32-rhe6.7-debug #1 Red Hat KVM
<4>[15143.361813] RIP: 0010:[<ffffffff812a075c>]  [<ffffffff812a075c>] _raw_spin_unlock+0xc/0xa0
<4>[15143.361813] RSP: 0018:ffff880083e67a00  EFLAGS: 00010286
<4>[15143.361813] RAX: 0000000000000000 RBX: ffff8800afc9ff50 RCX: 0000000000000000
<4>[15143.361813] RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff8800afc9ff50
<4>[15143.361813] RBP: ffff880083e67a10 R08: 00000000fffffffb R09: 00000000fffffffe
<4>[15143.361813] R10: 0000000000000000 R11: 0000000000000003 R12: ffff88003c658f38
<4>[15143.361813] R13: ffff88003c658f80 R14: ffff88003c658f78 R15: 0000000000000000
<4>[15143.361813] FS:  0000000000000000(0000) GS:ffff880006240000(0000) knlGS:0000000000000000
<4>[15143.361813] CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
<4>[15143.361813] CR2: ffff8800afc9ff54 CR3: 0000000001a2d000 CR4: 00000000000006e0
<4>[15143.361813] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
<4>[15143.361813] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
<4>[15143.361813] Process ptlrpcd_00_03 (pid: 5500, threadinfo ffff880083e64000, task ffff880084e5a4c0)
<4>[15143.361813] Stack:
<4>[15143.361813]  ffff88003c658f38 ffff8800afc9ff50 ffff880083e67a20 ffffffff81530afe
<4>[15143.361813] <d> ffff880083e67a40 ffffffffa04e325c ffff88003c658f38 0000000000000000
<4>[15143.361813] <d> ffff880083e67b50 ffffffffa04e3805 ffff880000000002 0000000000000000
<4>[15143.361813] Call Trace:
<4>[15143.361813]  [<ffffffff81530afe>] _spin_unlock+0xe/0x10
<4>[15143.361813]  [<ffffffffa04e325c>] osc_extent_remove+0x3c/0x50 [osc]
<4>[15143.361813]  [<ffffffffa04e3805>] osc_extent_finish+0x595/0xae0 [osc]
<4>[15143.361813]  [<ffffffffa04cd007>] ? brw_interpret+0x937/0x2310 [osc]
<4>[15143.361813]  [<ffffffffa04cd063>] brw_interpret+0x993/0x2310 [osc]
<4>[15143.361813]  [<ffffffffa07acbf1>] ? libcfs_debug_msg+0x41/0x50 [libcfs]
<4>[15143.361813]  [<ffffffffa07a91f8>] ? libcfs_log_return+0x28/0x40 [libcfs]
<4>[15143.361813]  [<ffffffffa0f5eb8b>] ? ptlrpc_unregister_bulk+0x35b/0xae0 [ptlrpc]
<4>[15143.361813]  [<ffffffffa0f5758b>] ptlrpc_check_set+0x60b/0x1e70 [ptlrpc]
<4>[15143.361813]  [<ffffffffa0f85993>] ptlrpcd_check+0x3e3/0x630 [ptlrpc]
<4>[15143.361813]  [<ffffffffa0f85efb>] ptlrpcd+0x31b/0x500 [ptlrpc]
<4>[15143.361813]  [<ffffffff81063a80>] ? default_wake_function+0x0/0x20
<4>[15143.361813]  [<ffffffffa0f85be0>] ? ptlrpcd+0x0/0x500 [ptlrpc]
<4>[15143.361813]  [<ffffffff8109f82e>] kthread+0x9e/0xc0
<4>[15143.361813]  [<ffffffff8100c2ca>] child_rip+0xa/0x20
<4>[15143.361813]  [<ffffffff8109f790>] ? kthread+0x0/0xc0
<4>[15143.361813]  [<ffffffff8100c2c0>] ? child_rip+0x0/0x20
<4>[15143.361813] Code: 00 00 e8 06 bd 28 00 44 8b 4b 08 41 83 c8 ff 48 c7 c1 75 85 7f 81 eb 98 0f 1f 80 00 00 00 00 55 48 89 e5 53 48 89 fb 48 83 ec 08 <81> 7f 04 ad 4e ad de 75 44 8b 03 89 c2 c1 fa 10 66 39 c2 74 68
Whamcloud Community JIRA

[LU-7164] os_extent should hold a refcount to osc_object
