Hi, i am doing some filesystems testing (aka fuzzing), got this issue on vanilla 3.8.4: Mar 27 12:23:47 linux kernel: [ 373.018889] WARNING: at fs/jbd/journal.c:470 __log_start_commit+0x94/0xb0() Mar 27 12:23:47 linux kernel: [ 373.018891] Hardware name: PowerEdge T420 Mar 27 12:23:47 linux kernel: [ 373.018894] jbd: bad log_start_commit: 2801795176 2801795176 0 0 Mar 27 12:23:47 linux kernel: [ 373.018895] Modules linked in: ipmi_devintf ipmi_si ipmi_msghandler nls_iso8859_1 coretemp kvm_intel kvm ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw aes_x86_64 sb_edac xts edac_core gf128mul lp joydev mei gpio_ich microcode dcdbas lpc_ich parport shpchp acpi_power_meter wmi mac_hid raid10 raid456 async_pq async_xor xor async_memcpy async_raid6_recov hid_generic ahci libahci usbhid tg3 ptp hid raid6_pq pps_core async_tx raid1 raid0 multipath linear Mar 27 12:23:47 linux kernel: [ 373.018935] Pid: 1924, comm: umount Tainted: G W 3.8.4 #1 Mar 27 12:23:47 linux kernel: [ 373.018937] Call Trace: Mar 27 12:23:47 linux kernel: [ 373.018943] [<ffffffff810589ef>] warn_slowpath_common+0x7f/0xc0 Mar 27 12:23:47 linux kernel: [ 373.018947] [<ffffffff81058ae6>] warn_slowpath_fmt+0x46/0x50 Mar 27 12:23:47 linux kernel: [ 373.018950] [<ffffffff8127c044>] __log_start_commit+0x94/0xb0 Mar 27 12:23:47 linux kernel: [ 373.018955] [<ffffffff816b5a0e>] ? _raw_spin_lock+0xe/0x20 Mar 27 12:23:47 linux kernel: [ 373.018959] [<ffffffff8127c496>] log_start_commit+0x36/0x60 Mar 27 12:23:47 linux kernel: [ 373.018964] [<ffffffff8120e53f>] ext3_evict_inode+0x26f/0x2c0 Mar 27 12:23:47 linux kernel: [ 373.018968] [<ffffffff811a5e36>] evict+0xb6/0x1b0 Mar 27 12:23:47 linux kernel: [ 373.018971] [<ffffffff811a6649>] iput+0x109/0x190 Mar 27 12:23:47 linux kernel: [ 373.018975] [<ffffffff8127bef7>] journal_destroy+0x1a7/0x220 Mar 27 12:23:47 linux kernel: [ 373.018978] [<ffffffff8107e080>] ? finish_wait+0x80/0x80 Mar 27 12:23:47 linux kernel: [ 373.018983] [<ffffffff8121c317>] ext3_put_super+0x47/0x2a0 Mar 27 12:23:47 linux kernel: [ 373.018988] [<ffffffff8118dd91>] generic_shutdown_super+0x61/0xf0 Mar 27 12:23:47 linux kernel: [ 373.018992] [<ffffffff8118de50>] kill_block_super+0x30/0x80 Mar 27 12:23:47 linux kernel: [ 373.018995] [<ffffffff8118e247>] deactivate_locked_super+0x57/0x80 Mar 27 12:23:47 linux kernel: [ 373.018999] [<ffffffff8118ee0e>] deactivate_super+0x4e/0x70 Mar 27 12:23:47 linux kernel: [ 373.019003] [<ffffffff811aa881>] mntput_no_expire+0x101/0x160 Mar 27 12:23:47 linux kernel: [ 373.019007] [<ffffffff811ab7dc>] sys_umount+0x9c/0x3c0 Mar 27 12:23:47 linux kernel: [ 373.019012] [<ffffffff8106ed92>] ? __set_current_blocked+0x52/0x70 Mar 27 12:23:47 linux kernel: [ 373.019017] [<ffffffff816be399>] system_call_fastpath+0x16/0x1b Mar 27 12:23:47 linux kernel: [ 373.019019] ---[ end trace 7a9e3bafae5e7653 ]--- how to reproduce: *you need* to be root then issue the cmds below mkdir /media/test gunzip ext3.poc.img.gz mount -t ext3 -o loop ext3.poc.img /media/test/ unmount /media/test the unmount would never return, check the logs to see the backtrace. best regards, David Maciejak Fortinet's FortiGuard Labs
Created attachment 96361 [details] ext3 poc file
Thanks for report. The fix for this problem is already in my tree (commit e643692138cfa33528f054b071ba2583509bb217).
And I now verified that the fix really works.
so quick, thx !