Re: [f2fs-dev] [PATCH] f2fs: do checkpoint in kill_sb

[Chao Yu]

On 2018/7/12 15:34, Jaegeuk Kim wrote:
> On 07/10, Chao Yu wrote:
>> On 2018/7/10 4:36, Jaegeuk Kim wrote:
>>> On 07/09, Chao Yu wrote:
>>>> On 2018/7/7 9:20, Jaegeuk Kim wrote:
>>>>> When unmounting f2fs in force mode, we can get it stuck by io_schedule()
>>>>
>>>> force mode means force shutdown?
>>>
>>> Yes.
>>>
>>>>
>>>>> by some pending IOs in meta_inode.
>>>>>
>>>>> io_schedule+0xd/0x30
>>>>> wait_on_page_bit_common+0xc6/0x130
>>>>
>>>> Looks like a deadlock here? Does this mean we forget to submit cached bio in
>>>> somewhere?
>>>
>>> I'm hitting this randomly, when running all the testcases in xfstests.
>>> I can't find the missing flow, and only get this during kill_sb.
>>
>> I guess we need this before returning from ->writepage{,s}:
>>
>> if (is_sbi_flag_set(sbi, SBI_IS_SHUTDOWN))
>> 	f2fs_submit_merged_write()
>>
>> How about trying this instead?
> 
> No, it's not shutdown case unfortunately.

Still I didn't understand why this patch can fix the issue... :(

Thanks,

> 
>>
>> Thanks,
>>
>>>
>>>>
>>>> Thanks,
>>>>
>>>>> __filemap_fdatawait_range+0xbd/0x100
>>>>> filemap_fdatawait_keep_errors+0x15/0x40
>>>>> sync_inodes_sb+0x1cf/0x240
>>>>> sync_filesystem+0x52/0x90
>>>>> generic_shutdown_super+0x1d/0x110
>>>>> ceph_kill_sb+0x28/0x80 [ceph]
>>>>> deactivate_locked_super+0x35/0x60
>>>>> cleanup_mnt+0x36/0x70
>>>>> task_work_run+0x79/0xa0
>>>>> exit_to_usermode_loop+0x62/0x70
>>>>> do_syscall_64+0xdb/0xf0
>>>>> entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>>>> 0xffffffffffffffff
>>>>>
>>>>> Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
>>>>> ---
>>>>>  fs/f2fs/super.c | 16 +++++++++++++---
>>>>>  1 file changed, 13 insertions(+), 3 deletions(-)
>>>>>
>>>>> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
>>>>> index 3995e926ba3a..1dc6809fac38 100644
>>>>> --- a/fs/f2fs/super.c
>>>>> +++ b/fs/f2fs/super.c
>>>>> @@ -3089,9 +3089,19 @@ static struct dentry *f2fs_mount(struct file_system_type *fs_type, int flags,
>>>>>  static void kill_f2fs_super(struct super_block *sb)
>>>>>  {
>>>>>  	if (sb->s_root) {
>>>>> -		set_sbi_flag(F2FS_SB(sb), SBI_IS_CLOSE);
>>>>> -		f2fs_stop_gc_thread(F2FS_SB(sb));
>>>>> -		f2fs_stop_discard_thread(F2FS_SB(sb));
>>>>> +		struct f2fs_sb_info *sbi = F2FS_SB(sb);
>>>>> +
>>>>> +		set_sbi_flag(sbi, SBI_IS_CLOSE);
>>>>> +		f2fs_stop_gc_thread(sbi);
>>>>> +		f2fs_stop_discard_thread(sbi);
>>>>> +
>>>>> +		if (is_sbi_flag_set(sbi, SBI_IS_DIRTY) ||
>>>>> +				!is_set_ckpt_flags(sbi, CP_UMOUNT_FLAG)) {
>>>>> +			struct cp_control cpc = {
>>>>> +				.reason = CP_UMOUNT,
>>>>> +			};
>>>>> +			f2fs_write_checkpoint(sbi, &cpc);
>>>>> +		}
>>>>>  	}
>>>>>  	kill_block_super(sb);
>>>>>  }
>>>>>
>>>
>>> .
>>>
> 
> .
>