RE: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks

From: Chao Yu
Date: Fri Aug 22 2014 - 02:57:37 EST

Next message: Geert Uytterhoeven: "Re: [PATCH] backlight: platform_lcd: remove unnecessary #ifdef CONFIG_OF"
Previous message: Andrew Morton: "Re: [patch 4/4] prctl: PR_SET_MM -- Introduce PR_SET_MM_MAP operation, v3"
In reply to: Jaegeuk Kim: "Re: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks"
Next in thread: Jaegeuk Kim: "Re: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi Jaegeuk,

> -----Original Message-----
> From: Jaegeuk Kim [mailto:jaegeuk@xxxxxxxxxx]
> Sent: Friday, August 22, 2014 12:45 AM
> To: Chao Yu
> Cc: linux-fsdevel@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> linux-f2fs-devel@xxxxxxxxxxxxxxxxxxxxx
> Subject: Re: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks
>
> On Wed, Aug 20, 2014 at 10:07:04AM +0800, Chao Yu wrote:
> > > -----Original Message-----
> > > From: Jaegeuk Kim [mailto:jaegeuk@xxxxxxxxxx]
> > > Sent: Wednesday, August 20, 2014 12:58 AM
> > > To: Chao Yu
> > > Cc: linux-fsdevel@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> > > linux-f2fs-devel@xxxxxxxxxxxxxxxxxxxxx
> > > Subject: Re: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks
> > >
> > > On Tue, Aug 19, 2014 at 04:04:11PM +0800, Chao Yu wrote:
> > > > Hi Jaegeuk,
> > > >
> > > > > -----Original Message-----
> > > > > From: Jaegeuk Kim [mailto:jaegeuk@xxxxxxxxxx]
> > > > > Sent: Saturday, August 16, 2014 6:04 AM
> > > > > To: linux-kernel@xxxxxxxxxxxxxxx; linux-fsdevel@xxxxxxxxxxxxxxx;
> > > > > linux-f2fs-devel@xxxxxxxxxxxxxxxxxxxxx
> > > > > Cc: Jaegeuk Kim
> > > > > Subject: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks
> > > > >
> > > > > The init_inode_metadata calls truncate_blocks when error is occurred.
> > > > > The callers holds f2fs_lock_op, so we should not call it again in
> > > > > truncate_blocks.
> > > >
> > > > Nice catch! Your solution is a good way to fix this issue.
> > > >
> > > > Previously, in create inode path, I found there are some redundant codes between
> > > > init_inode_metadata and evict_inode, including:
> > > > truncate_inode_pages(&inode->i_data, 0);
> > > > truncate_blocks(inode, 0);
> > > > remove_dirty_dir_inode(inode);
> > > > remove_inode_page(inode);
> > > >
> > > > So I think there is another way to fix this issue by removing error path handling
> > > > codes in init_inode_metadata, not making the inode bad to left garbage clean work in
> > > > evict_inode. In this way we can avoid adding additional argument for all different
> > > > callers.
> > >
> > > Well, possible.
> > > But we need to take a closer look at the race condition on the inode cache.
> > > What can happen if this bad inode is reassigned to the other thread?
> >
> > I don't get it. As I know, in evict(), we call ->evict_inode before
> > remove_inode_hash(), so before all clean work was done we will not reassign
> > this hashed uncleaned inode to other thread.
> >
> > Am I missing anything?
>
> What I meant was it may happen between init_inode_metadata and iput.

Actually, can happen between unlock_new_inode and iput in error path of
create/symlink/mkdir/mknod/tmpfile. Scenario is like this:

->f2fs_mkdir
->f2fs_add_link
->__f2fs_add_link
->init_inode_metadata failed here
->gc_thread_func
->f2fs_gc
->do_garbage_collect
->gc_data_segment
->f2fs_iget
->iget_locked
->wait_on_inode
->unlock_new_inode
->move_data_page
->make_bad_inode
->iput

No problem now, but I'd like to remove unlock_new_inode from error path of inode
creating procedure as we'd better wakeup waiter in end of ->iput instead of
wakeup waiter in unlock_new_inode before invoking make_bad_inod.

How do you think?

Anyway, drop this proposal if you do not like it, although I think it will be
ok theoretically after we fix above issue.

Thanks,
Yu

>
> Thanks,
>
> >
> > Regards,
> > Yu
> >
> > >
> > > >
> > > > How do you think?
> > > >
> > > > Thanks,
> > > > Yu
> > > >
> > > > >
> > > > > Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
> > > > > ---
> > > > > fs/f2fs/data.c | 2 +-
> > > > > fs/f2fs/dir.c | 2 +-
> > > > > fs/f2fs/f2fs.h | 2 +-
> > > > > fs/f2fs/file.c | 13 ++++++++-----
> > > > > fs/f2fs/inline.c | 2 +-
> > > > > 5 files changed, 12 insertions(+), 9 deletions(-)
> > > > >
> > > > > diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
> > > > > index 68834e2..14cc3e8 100644
> > > > > --- a/fs/f2fs/data.c
> > > > > +++ b/fs/f2fs/data.c
> > > > > @@ -935,7 +935,7 @@ static void f2fs_write_failed(struct address_space *mapping, loff_t
> > > to)
> > > > >
> > > > > if (to > inode->i_size) {
> > > > > truncate_pagecache(inode, inode->i_size);
> > > > > - truncate_blocks(inode, inode->i_size);
> > > > > + truncate_blocks(inode, inode->i_size, true);
> > > > > }
> > > > > }
> > > > >
> > > > > diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c
> > > > > index a69bbfa..155fb05 100644
> > > > > --- a/fs/f2fs/dir.c
> > > > > +++ b/fs/f2fs/dir.c
> > > > > @@ -391,7 +391,7 @@ put_error:
> > > > > error:
> > > > > /* once the failed inode becomes a bad inode, i_mode is S_IFREG */
> > > > > truncate_inode_pages(&inode->i_data, 0);
> > > > > - truncate_blocks(inode, 0);
> > > > > + truncate_blocks(inode, 0, false);
> > > > > remove_dirty_dir_inode(inode);
> > > > > remove_inode_page(inode);
> > > > > return ERR_PTR(err);
> > > > > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > > > > index 2723b2d..7f976c1 100644
> > > > > --- a/fs/f2fs/f2fs.h
> > > > > +++ b/fs/f2fs/f2fs.h
> > > > > @@ -1122,7 +1122,7 @@ static inline void f2fs_stop_checkpoint(struct f2fs_sb_info *sbi)
> > > > > */
> > > > > int f2fs_sync_file(struct file *, loff_t, loff_t, int);
> > > > > void truncate_data_blocks(struct dnode_of_data *);
> > > > > -int truncate_blocks(struct inode *, u64);
> > > > > +int truncate_blocks(struct inode *, u64, bool);
> > > > > void f2fs_truncate(struct inode *);
> > > > > int f2fs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
> > > > > int f2fs_setattr(struct dentry *, struct iattr *);
> > > > > diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> > > > > index ecbdf6a..a8e97f8 100644
> > > > > --- a/fs/f2fs/file.c
> > > > > +++ b/fs/f2fs/file.c
> > > > > @@ -422,7 +422,7 @@ out:
> > > > > f2fs_put_page(page, 1);
> > > > > }
> > > > >
> > > > > -int truncate_blocks(struct inode *inode, u64 from)
> > > > > +int truncate_blocks(struct inode *inode, u64 from, bool lock)
> > > > > {
> > > > > struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
> > > > > unsigned int blocksize = inode->i_sb->s_blocksize;
> > > > > @@ -438,14 +438,16 @@ int truncate_blocks(struct inode *inode, u64 from)
> > > > > free_from = (pgoff_t)
> > > > > ((from + blocksize - 1) >> (sbi->log_blocksize));
> > > > >
> > > > > - f2fs_lock_op(sbi);
> > > > > + if (lock)
> > > > > + f2fs_lock_op(sbi);
> > > > >
> > > > > set_new_dnode(&dn, inode, NULL, NULL, 0);
> > > > > err = get_dnode_of_data(&dn, free_from, LOOKUP_NODE);
> > > > > if (err) {
> > > > > if (err == -ENOENT)
> > > > > goto free_next;
> > > > > - f2fs_unlock_op(sbi);
> > > > > + if (lock)
> > > > > + f2fs_unlock_op(sbi);
> > > > > trace_f2fs_truncate_blocks_exit(inode, err);
> > > > > return err;
> > > > > }
> > > > > @@ -463,7 +465,8 @@ int truncate_blocks(struct inode *inode, u64 from)
> > > > > f2fs_put_dnode(&dn);
> > > > > free_next:
> > > > > err = truncate_inode_blocks(inode, free_from);
> > > > > - f2fs_unlock_op(sbi);
> > > > > + if (lock)
> > > > > + f2fs_unlock_op(sbi);
> > > > > done:
> > > > > /* lastly zero out the first data page */
> > > > > truncate_partial_data_page(inode, from);
> > > > > @@ -480,7 +483,7 @@ void f2fs_truncate(struct inode *inode)
> > > > >
> > > > > trace_f2fs_truncate(inode);
> > > > >
> > > > > - if (!truncate_blocks(inode, i_size_read(inode))) {
> > > > > + if (!truncate_blocks(inode, i_size_read(inode), true)) {
> > > > > inode->i_mtime = inode->i_ctime = CURRENT_TIME;
> > > > > mark_inode_dirty(inode);
> > > > > }
> > > > > diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c
> > > > > index 520758b..4d1f39f 100644
> > > > > --- a/fs/f2fs/inline.c
> > > > > +++ b/fs/f2fs/inline.c
> > > > > @@ -247,7 +247,7 @@ process_inline:
> > > > > update_inode(inode, ipage);
> > > > > f2fs_put_page(ipage, 1);
> > > > > } else if (ri && (ri->i_inline & F2FS_INLINE_DATA)) {
> > > > > - truncate_blocks(inode, 0);
> > > > > + truncate_blocks(inode, 0, false);
> > > > > set_inode_flag(F2FS_I(inode), FI_INLINE_DATA);
> > > > > goto process_inline;
> > > > > }
> > > > > --
> > > > > 1.8.5.2 (Apple Git-48)
> > > > >
> > > > >
> > > > > ------------------------------------------------------------------------------
> > > > > _______________________________________________
> > > > > Linux-f2fs-devel mailing list
> > > > > Linux-f2fs-devel@xxxxxxxxxxxxxxxxxxxxx
> > > > > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
> > > >
> > > >
> > > > ------------------------------------------------------------------------------
> > > > _______________________________________________
> > > > Linux-f2fs-devel mailing list
> > > > Linux-f2fs-devel@xxxxxxxxxxxxxxxxxxxxx
> > > > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Geert Uytterhoeven: "Re: [PATCH] backlight: platform_lcd: remove unnecessary #ifdef CONFIG_OF"
Previous message: Andrew Morton: "Re: [patch 4/4] prctl: PR_SET_MM -- Introduce PR_SET_MM_MAP operation, v3"
In reply to: Jaegeuk Kim: "Re: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks"
Next in thread: Jaegeuk Kim: "Re: [f2fs-dev] [PATCH 3/4] f2fs: avoid double lock in truncate_blocks"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]