Latest working kernel version: (not known) Earliest failing kernel version: 2.6.22 Distribution: Ubuntu Linux 7.10 server 64-bit (Gutsy) Hardware Environment: x86_64 Software Environment: Problem Description: Steps to reproduce:
How to reproduce this: create an XFS filesystem, mount it, run bonnie++ on the filesystem and in another shell erase the disk contents with dd if=/dev/zero of=... Result: [ 148.004980] Filesystem "sdb": XFS internal error xfs_btree_check_sblock at line 334 of file /build/buildd/linux-source-2.6.22-2.6.22/fs/xfs/xfs_btree.c. Caller 0xffffffff8828c01e [ 148.006446] [ 148.006447] Call Trace: [ 148.006485] [<ffffffff8827a49e>] :xfs:xfs_btree_check_sblock+0x6e/0xf0 [ 148.006505] [<ffffffff8828c01e>] :xfs:xfs_inobt_lookup+0x17e/0x350 [ 148.006523] [<ffffffff88279f9b>] :xfs:xfs_btree_init_cursor+0x4b/0x200 [ 148.006542] [<ffffffff8828b5bb>] :xfs:xfs_dialloc+0x53b/0x940 [ 148.006569] [<ffffffff80317164>] generic_make_request+0x154/0x310 [ 148.006580] [<ffffffff88140680>] :ext3:ext3_get_block+0x0/0x120 [ 148.006589] [<ffffffff88140680>] :ext3:ext3_get_block+0x0/0x120 [ 148.006601] [<ffffffff8022d26d>] find_busiest_group+0x1bd/0x800 [ 148.006620] [<ffffffff88293527>] :xfs:xfs_ialloc+0x67/0x580 [ 148.006640] [<ffffffff882a8b6b>] :xfs:xfs_dir_ialloc+0xab/0x380 [ 148.006666] [<ffffffff804377f2>] __down_write_nested+0x12/0xb0 [ 148.006684] [<ffffffff882a5eb6>] :xfs:xfs_trans_reserve+0xd6/0x230 [ 148.006703] [<ffffffff882af0b5>] :xfs:xfs_create+0x3b5/0x6f0 [ 148.006718] [<ffffffff882663ed>] :xfs:xfs_attr_get+0xcd/0x100 [ 148.006738] [<ffffffff882bb0ff>] :xfs:xfs_vn_mknod+0x26f/0x3b0 [ 148.006752] [<ffffffff802a239e>] vfs_create+0x11e/0x170 [ 148.006756] [<ffffffff802a6045>] open_namei+0x675/0x6d0 [ 148.006759] [<ffffffff802ad35e>] d_instantiate+0x4e/0x90 [ 148.006771] [<ffffffff802978bc>] do_filp_open+0x1c/0x50 [ 148.006778] [<ffffffff8029794a>] do_sys_open+0x5a/0x100 [ 148.006786] [<ffffffff80209e8e>] system_call+0x7e/0x83 [ 148.006790]
that's not an oops, that is xfs properly shutting down in the face of (your intentional) filesystem corruption. What is the bug here? -Eric
Sorry, this is a misunderstanding of my side: I made the assumption that this call stack was triggered by a kernel oops. The above is indeed the result of an xfs_stack_trace() call and not an oops.
That's ok; people see stack traces and get afraid ;) Can you close it NOTABUG then? Thanks, -Eric
OK, done.