From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Mason Subject: Re: [2.6.39-rc1] extent reference leaking... Date: Wed, 30 Mar 2011 07:24:10 -0400 Message-ID: <1301484231-sup-8860@think> References: Content-Type: text/plain; charset=UTF-8 Cc: Linux BTRFS To: Daniel J Blueman Return-path: In-reply-to: List-ID: Excerpts from Daniel J Blueman's message of 2011-03-30 06:37:57 -0400: > When running the Linux Test Project against a BTRFS RAID 1 array, > after some time I see BTRFS trying to free an extent that still has > state [1]. > > Let me know if anyone is interested in a more specific reproducer and > I'll take a look. Very interested. I didn't see this one in testing and it isn't good at all. -chris > > Daniel > > --- [1] > > WARNING: at fs/btrfs/extent_io.c:3371 free_extent_buffer+0x31/0x40() > Hardware name: X8STi > Modules linked in: tun microcode loop raid10 raid456 async_memcpy > async_pq async_xor xor async_raid6_recov raid6_pq async_tx raid1 raid0 > multipath linear md_mod > Pid: 14202, comm: ftest08 Tainted: G W 2.6.39-rc1-350cd #1 > Call Trace: > [] warn_slowpath_common+0x7a/0xb0 > [] warn_slowpath_null+0x15/0x20 > [] free_extent_buffer+0x31/0x40 > [] btrfs_lock_root_node+0x20/0x50 > [] btrfs_search_slot+0x400/0x790 > [] ? btrfs_alloc_path+0x15/0x30 > [] ? btrfs_alloc_path+0x15/0x30 > [] btrfs_lookup_csums_range+0x83/0x4b0 > [] ? unmap_extent_buffer+0xe/0x40 > [] ? btrfs_file_extent_compression+0xe4/0xf0 > [] copy_items+0x315/0x3f0 > [] btrfs_log_inode+0x3ea/0x520 > [] ? start_log_trans+0x6b/0x150 > [] btrfs_log_inode_parent+0x193/0x2e0 > [] ? dget_parent+0x94/0x100 > [] ? dget_parent+0x17/0x100 > [] btrfs_log_dentry_safe+0x44/0x70 > [] btrfs_sync_file+0xeb/0x190 > [] vfs_fsync_range+0x73/0x90 > [] ? fget_raw+0x260/0x260 > [] vfs_fsync+0x17/0x20 > [] do_fsync+0x35/0x60 > [] sys_fsync+0xb/0x10 > [] system_call_fastpath+0x16/0x1b > ---[ end trace a7919e7f17c0a728 ]---