Changwei Ge
2017-Apr-10 06:52 UTC
[Ocfs2-devel] [Bug Report] crash in the path of direct IO
Hi, We encountered a crash issue days ago. The call trace follows as below:>From the call trace, we can see that a direct read request caused thiscrash issue, which triggered a BUG_ON check point. With the help of debugfs.ocfs2 tool, I can see that clusters owned by the target file are extremely scattered . Most of them are not continuous. Does anyone have ideas about this issue? [10628085.052236] RIP: 0010:[<ffffffff81240ba0>] [<ffffffff81240ba0>] do_blockdev_direct_IO+0x1e10/0x1eb0 [10628085.052248] RSP: 0018:ffff88018038fa38 EFLAGS: 00010246 [10628085.052252] RAX: 0000000000000000 RBX: ffff8800ae579680 RCX: 0000000000000009 [10628085.052258] RDX: 0000000000000000 RSI: 000000000fb25900 RDI: 000000007d92c800 [10628085.052265] RBP: ffff88018038fc68 R08: ffff8805e042a298 R09: 0000000000000001 [10628085.052271] R10: 0000000000000003 R11: 0000000000000000 R12: 0000000000000000 [10628085.052277] R13: 0000000000000a00 R14: ffff8800ae5796a8 R15: ffffea0014865280 [10628085.052284] FS: 00007f7f517fa700(0000) GS:ffff8806276c0000(0000) knlGS:0000000000000000 [10628085.052290] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [10628085.052294] CR2: 000000001fbc0000 CR3: 0000000bdb8f7000 CR4: 00000000001426e0 [10628085.052300] Stack: [10628085.052304] ffff88018038fc60 ffff880626c18d00 0000000000117c00 ffff8800ae5796e8 [10628085.052311] ffff88018038fb90 ffff8800ae5796b4 0000000000000200 0000000000131000 [10628085.052318] 000000007d92c5fb ffff88052194a000 0000000117c00000 fffffffffffffe00 [10628085.052326] Call Trace: [10628085.052371] [<ffffffffc06ea3c0>] ? ocfs2_zero_new_buffers+0x130/0x130 [ocfs2] [10628085.052379] [<ffffffff81240c83>] __blockdev_direct_IO+0x43/0x50 [10628085.052395] [<ffffffffc06e9af0>] ? ocfs2_zero_tail+0x50/0x50 [ocfs2] [10628085.052411] [<ffffffffc06e933f>] ocfs2_direct_IO+0x8f/0xa0 [ocfs2] [10628085.052426] [<ffffffffc06e9af0>] ? ocfs2_zero_tail+0x50/0x50 [ocfs2] [10628085.052432] [<ffffffff8118ba1a>] generic_file_read_iter+0x56a/0x610 [10628085.052449] [<ffffffffc070004b>] ? __ocfs2_cluster_unlock.isra.36+0xab/0xf0 [ocfs2] [10628085.052469] [<ffffffffc070a2da>] ocfs2_file_read_iter+0xca/0x410 [ocfs2] [10628085.052476] [<ffffffff810ffe44>] ? do_futex+0x304/0xbc0 [10628085.052481] [<ffffffff8120263b>] __vfs_read+0xab/0xf0 [10628085.052487] [<ffffffff812039d5>] vfs_read+0x95/0x130 [10628085.052491] [<ffffffff81203c62>] SyS_pread64+0x92/0xa0 [10628085.052498] [<ffffffff817f81b2>] system_call_fastpath+0x16/0x75 Br. Thanks. Changwei.
Hi Changwei, Did you get this kernel crash file? Could you paste longer syslog before crash (e.g. 100 ~ 200 lines) via crash tool? e.g. crash vmlinux-4.4.21-69-default.gz vmcore crash> log Thanks Gang>>> > Hi, > > We encountered a crash issue days ago. > > The call trace follows as below: > > From the call trace, we can see that a direct read request caused this > crash issue, which triggered a BUG_ON check point. > > With the help of debugfs.ocfs2 tool, I can see that clusters owned by > the target file are extremely scattered . Most of them are not > > continuous. > > Does anyone have ideas about this issue? > > [10628085.052236] RIP: 0010:[<ffffffff81240ba0>] [<ffffffff81240ba0>] > do_blockdev_direct_IO+0x1e10/0x1eb0 > [10628085.052248] RSP: 0018:ffff88018038fa38 EFLAGS: 00010246 > [10628085.052252] RAX: 0000000000000000 RBX: ffff8800ae579680 RCX: > 0000000000000009 > [10628085.052258] RDX: 0000000000000000 RSI: 000000000fb25900 RDI: > 000000007d92c800 > [10628085.052265] RBP: ffff88018038fc68 R08: ffff8805e042a298 R09: > 0000000000000001 > [10628085.052271] R10: 0000000000000003 R11: 0000000000000000 R12: > 0000000000000000 > [10628085.052277] R13: 0000000000000a00 R14: ffff8800ae5796a8 R15: > ffffea0014865280 > [10628085.052284] FS: 00007f7f517fa700(0000) GS:ffff8806276c0000(0000) > knlGS:0000000000000000 > [10628085.052290] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > [10628085.052294] CR2: 000000001fbc0000 CR3: 0000000bdb8f7000 CR4: > 00000000001426e0 > [10628085.052300] Stack: > [10628085.052304] ffff88018038fc60 ffff880626c18d00 0000000000117c00 > ffff8800ae5796e8 > [10628085.052311] ffff88018038fb90 ffff8800ae5796b4 0000000000000200 > 0000000000131000 > [10628085.052318] 000000007d92c5fb ffff88052194a000 0000000117c00000 > fffffffffffffe00 > [10628085.052326] Call Trace: > [10628085.052371] [<ffffffffc06ea3c0>] ? > ocfs2_zero_new_buffers+0x130/0x130 [ocfs2] > [10628085.052379] [<ffffffff81240c83>] __blockdev_direct_IO+0x43/0x50 > [10628085.052395] [<ffffffffc06e9af0>] ? ocfs2_zero_tail+0x50/0x50 [ocfs2] > [10628085.052411] [<ffffffffc06e933f>] ocfs2_direct_IO+0x8f/0xa0 [ocfs2] > [10628085.052426] [<ffffffffc06e9af0>] ? ocfs2_zero_tail+0x50/0x50 [ocfs2] > [10628085.052432] [<ffffffff8118ba1a>] generic_file_read_iter+0x56a/0x610 > [10628085.052449] [<ffffffffc070004b>] ? > __ocfs2_cluster_unlock.isra.36+0xab/0xf0 [ocfs2] > [10628085.052469] [<ffffffffc070a2da>] ocfs2_file_read_iter+0xca/0x410 > [ocfs2] > [10628085.052476] [<ffffffff810ffe44>] ? do_futex+0x304/0xbc0 > [10628085.052481] [<ffffffff8120263b>] __vfs_read+0xab/0xf0 > [10628085.052487] [<ffffffff812039d5>] vfs_read+0x95/0x130 > [10628085.052491] [<ffffffff81203c62>] SyS_pread64+0x92/0xa0 > [10628085.052498] [<ffffffff817f81b2>] system_call_fastpath+0x16/0x75 > > > Br. > > Thanks. > > Changwei.