On Nov 05, 2019 / 20:03, Chao Yu wrote:
> On 2019/10/28 14:58, Shin'ichiro Kawasaki wrote:
> > On sudden f2fs shutdown, write pointers of zoned block devices can go
> > further but f2fs meta data keeps current segments at positions before the
> > write operations. After remounting the f2fs, this inconsistency causes
> > write operations not at write pointers and "Unaligned write command"
> > error is reported.
> > 
> > To avoid the error, compare current segments with write pointers of open
> > zones the current segments point to, during mount operation. If the write
> > pointer position is not aligned with the current segment position, assign
> > a new zone to the current segments. Also check the newly assigned zone
> > has write pointer at zone start. If not, make mount fail and ask users to
> > run fsck.
> > 
> > Perform the consistency check twice. Once during fsync recovery. Not to
> > lose the fsync data, do the check after fsync data gets restored and
> > before checkpoint commit which flushes data at current segment positions.
> > The second check is done at end of f2fs_fill_super() to make sure the
> > write pointer consistency regardless of fsync data recovery execution.
> > 
> > Signed-off-by: Shin'ichiro Kawasaki <[email protected]>
> > ---
> >  fs/f2fs/f2fs.h     |   1 +
> >  fs/f2fs/recovery.c |   6 +++
> >  fs/f2fs/segment.c  | 127 +++++++++++++++++++++++++++++++++++++++++++++
> >  fs/f2fs/super.c    |   8 +++
> >  4 files changed, 142 insertions(+)
> > 
> > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > index 4024790028aa..0216282c5b80 100644
> > --- a/fs/f2fs/f2fs.h
> > +++ b/fs/f2fs/f2fs.h
> > @@ -3136,6 +3136,7 @@ void f2fs_write_node_summaries(struct f2fs_sb_info 
> > *sbi, block_t start_blk);
> >  int f2fs_lookup_journal_in_cursum(struct f2fs_journal *journal, int type,
> >                     unsigned int val, int alloc);
> >  void f2fs_flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control 
> > *cpc);
> > +int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi, bool 
> > check_only);
> >  int f2fs_build_segment_manager(struct f2fs_sb_info *sbi);
> >  void f2fs_destroy_segment_manager(struct f2fs_sb_info *sbi);
> >  int __init f2fs_create_segment_manager_caches(void);
> > diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
> > index 783773e4560d..c75d1cbae4d1 100644
> > --- a/fs/f2fs/recovery.c
> > +++ b/fs/f2fs/recovery.c
> > @@ -795,6 +795,12 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, 
> > bool check_only)
> >     if (need_writecp) {
> >             set_sbi_flag(sbi, SBI_IS_RECOVERED);
> >  
> > +           /* recover zoned block devices' write pointer consistency */
> > +           if (!err && f2fs_sb_has_blkzoned(sbi)) {
> > +                   err = f2fs_fix_curseg_write_pointer(sbi, false);
> 
> Can we check and reset current segment under SBI_POR_DOING's protection? since
> once SBI_POR_DOING flag is cleared, kworker is able to flush dirty data/node,
> which may trigger unaligned write command if write pointer is inconsistent.

Yes, will move that part before the SBI_POR_DOING flag clear. Thanks.

--
Best Regards,
Shin'ichiro Kawasaki

_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Reply via email to