lkml.org 
[lkml]   [2011]   [Sep]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [BUG] D state process after unplug and umount usb disk
From
On 9/4/11, Alan Stern <stern@rowland.harvard.edu> wrote:
> On Sun, 4 Sep 2011, Lin Ming wrote:
>
>> On Sat, Sep 3, 2011 at 6:31 PM, Dave Young <hidave.darkstar@gmail.com>
>> wrote:
>> > Hi,
>> >
>> > Known issue?
>> >
>> > Reproduce by:
>> > mount /dev/sdb2 /mnt/sdb2 -t ext3
>> > cat /mnt/sdb2/* >/dev/null
>> > unplug usb disk
>> > umount -f /mnt/sdb2
>> >
>> > dmesg as below:
>
> BTW, this may or may not be related:
>
> https://bugzilla.kernel.org/show_bug.cgi?id=25832 (kernel crashes when
> a mounted ext3/4 file system is physically removed)

Thanks for the info, but It seems a different problem. I can not
reproduce it with 3.0 kernel for 30 times test. it can easily
reproduce with recent kernel in 1-5 times test

>
>> > [ 1440.567271] SysRq : Show Blocked State
>> > [ 1440.567278] task PC stack pid father
>> > [ 1440.567312] cat D ffff880037a1e378 4296 2381 2310
>> > 0x00000004
>> > [ 1440.567320] ffff88003cca5838 0000000000000046 ffff88003cca57e8
>> > ffffffff00000000
>> > [ 1440.567332] 00000000000134c0 00000000000134c0 00000000000134c0
>> > ffff880037a1e000
>> > [ 1440.567339] 00000000000134c0 ffff88003cca5fd8 00000000000134c0
>> > 00000000000134c0
>> > [ 1440.567347] Call Trace:
>> > [ 1440.567357] [<ffffffff810c0468>] ? lock_page+0x2a/0x2a
>> > [ 1440.567363] [<ffffffff815e452d>] io_schedule+0x5e/0x79
>> > [ 1440.567368] [<ffffffff810c0471>] sleep_on_page+0x9/0xd
>> > [ 1440.567373] [<ffffffff815e4adf>] __wait_on_bit_lock+0x41/0x8a
>> > [ 1440.567378] [<ffffffff810c0437>] __lock_page+0x61/0x68
>> > [ 1440.567384] [<ffffffff81058739>] ?
>> > autoremove_wake_function+0x34/0x34
>> > [ 1440.567390] [<ffffffff8102f704>] ? should_resched+0x9/0x29
>> > [ 1440.567395] [<ffffffff810c9b0b>] lock_page+0x25/0x29
>> > [ 1440.567400] [<ffffffff810ca1ac>]
>> > truncate_inode_pages_range+0x2a6/0x32d
>> > [ 1440.567406] [<ffffffff810ca240>] truncate_inode_pages+0xd/0xf
>> > [ 1440.567411] [<ffffffff811642cf>] ext3_evict_inode+0xc9/0x1d7
>> > [ 1440.567417] [<ffffffff8111288e>] evict+0xa3/0x15e
>> > [ 1440.567422] [<ffffffff81112b34>] dispose_list+0x3d/0x49
>> > [ 1440.567426] [<ffffffff81113475>] evict_inodes+0xf1/0x100
>> > [ 1440.567431] [<ffffffff810ffd21>] generic_shutdown_super+0x47/0xc7
>> > [ 1440.567436] [<ffffffff810ffdc3>] kill_block_super+0x22/0x65
>> > [ 1440.567441] [<ffffffff811000da>] deactivate_locked_super+0x21/0x52
>> > [ 1440.567445] [<ffffffff811009ab>] deactivate_super+0x35/0x39
>> > [ 1440.567452] [<ffffffff81116395>] mntput_no_expire+0xcb/0xd0
>> > [ 1440.567457] [<ffffffff811163bb>] mntput+0x21/0x23
>> > [ 1440.567461] [<ffffffff810ff8e7>] fput+0x196/0x1a5
>> > [ 1440.567467] [<ffffffff810fc7a1>] filp_close+0x6b/0x76
>> > [ 1440.567472] [<ffffffff8103fa8a>] put_files_struct+0x73/0xd1
>> > [ 1440.567477] [<ffffffff8103fb7b>] exit_files+0x46/0x4f
>> > [ 1440.567481] [<ffffffff8103fe01>] do_exit+0x27d/0x7b3
>> > [ 1440.567487] [<ffffffff8104d97f>] ? get_signal_to_deliver+0x81/0x4ac
>> > [ 1440.567493] [<ffffffff812e5e27>] ? do_raw_spin_lock+0x6b/0x122
>> > [ 1440.567497] [<ffffffff810405c5>] do_group_exit+0x7d/0xa8
>> > [ 1440.567502] [<ffffffff8104dd8b>] get_signal_to_deliver+0x48d/0x4ac
>> > [ 1440.567509] [<ffffffff81001ea2>] do_signal+0x39/0x600
>> > [ 1440.567514] [<ffffffff810fdb37>] ? fsnotify_access+0x5d/0x65
>> > [ 1440.567519] [<ffffffff810024a4>] do_notify_resume+0x27/0x69
>> > [ 1440.567524] [<ffffffff812e158e>] ? trace_hardirqs_on_thunk+0x3a/0x3f
>> > [ 1440.567529] [<ffffffff815ecd63>] int_signal+0x12/0x17
>
>> Where does this exit signal come from?
>> Does USB interrupt handler will send kill signal to all processes
>> accessing it when it's unplugged(by hand)?
>
> The USB stack does not send any signals at all unless a process
> specifically asks for them.
>
> Alan Stern
>
>


--
Regards
Yang RuiRui


\
 
 \ /
  Last update: 2011-09-05 11:07    [W:0.058 / U:0.060 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site