[v2] fs: do not update freeing inode io_list

  After commit cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode
already has I_DIRTY_INODE") writeiback_single_inode can push inode with
I_DIRTY_TIME set to b_dirty_time list. In case of freeing inode with
I_DIRTY_TIME set this can happened after deletion of inode io_list at
evict. Stack trace is following.

evict
fat_evict_inode
fat_truncate_blocks
fat_flush_inodes
writeback_inode
sync_inode_metadata(inode, sync=0)
writeback_single_inode(inode, wbc) <- wbc->sync_mode == WB_SYNC_NONE

This will lead to use after free in flusher thread.

Similar issue can be triggered if writeback_single_inode in the
stack trace update inode->io_list. Add explicit check to avoid it.

Fixes: cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode already has I_DIRTY_INODE")
Reported-by: syzbot+6ba92bd00d5093f7e371@syzkaller.appspotmail.com
Signed-off-by: Svyatoslav Feldsherov <feldsherov@google.com>
---
 V1 -> V2: 
 - address review comments
 - skip inode_cgwb_move_to_attached for freeing inode 

 fs/fs-writeback.c | 30 +++++++++++++++++++-----------
 1 file changed, 19 insertions(+), 11 deletions(-)
  

Hi--

Please see a small nit below.

On 11/14/22 13:21, Svyatoslav Feldsherov wrote:
> After commit cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode
> already has I_DIRTY_INODE") writeiback_single_inode can push inode with
> I_DIRTY_TIME set to b_dirty_time list. In case of freeing inode with
> I_DIRTY_TIME set this can happened after deletion of inode io_list at
> evict. Stack trace is following.
> 
> evict
> fat_evict_inode
> fat_truncate_blocks
> fat_flush_inodes
> writeback_inode
> sync_inode_metadata(inode, sync=0)
> writeback_single_inode(inode, wbc) <- wbc->sync_mode == WB_SYNC_NONE
> 
> This will lead to use after free in flusher thread.
> 
> Similar issue can be triggered if writeback_single_inode in the
> stack trace update inode->io_list. Add explicit check to avoid it.
> 
> Fixes: cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode already has I_DIRTY_INODE")
> Reported-by: syzbot+6ba92bd00d5093f7e371@syzkaller.appspotmail.com
> Signed-off-by: Svyatoslav Feldsherov <feldsherov@google.com>
> ---
>  V1 -> V2: 
>  - address review comments
>  - skip inode_cgwb_move_to_attached for freeing inode 
> 
>  fs/fs-writeback.c | 30 +++++++++++++++++++-----------
>  1 file changed, 19 insertions(+), 11 deletions(-)
> 
> diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
> index 443f83382b9b..c4aea096689c 100644
> --- a/fs/fs-writeback.c
> +++ b/fs/fs-writeback.c
> @@ -1712,18 +1712,26 @@ static int writeback_single_inode(struct inode *inode,
>  	wb = inode_to_wb_and_lock_list(inode);
>  	spin_lock(&inode->i_lock);
>  	/*
> -	 * If the inode is now fully clean, then it can be safely removed from
> -	 * its writeback list (if any).  Otherwise the flusher threads are
> -	 * responsible for the writeback lists.
> +	 * If the inode is freeing, it's io_list shoudn't be updated

	                            its

> +	 * as it can be finally deleted at this moment.
>  	 */
> -	if (!(inode->i_state & I_DIRTY_ALL))
> -		inode_cgwb_move_to_attached(inode, wb);
> -	else if (!(inode->i_state & I_SYNC_QUEUED)) {
> -		if ((inode->i_state & I_DIRTY))
> -			redirty_tail_locked(inode, wb);
> -		else if (inode->i_state & I_DIRTY_TIME) {
> -			inode->dirtied_when = jiffies;
> -			inode_io_list_move_locked(inode, wb, &wb->b_dirty_time);
> +	if (!(inode->i_state & I_FREEING)) {
> +		/*
> +		 * If the inode is now fully clean, then it can be safely
> +		 * removed from its writeback list (if any). Otherwise the
> +		 * flusher threads are responsible for the writeback lists.
> +		 */
> +		if (!(inode->i_state & I_DIRTY_ALL))
> +			inode_cgwb_move_to_attached(inode, wb);
> +		else if (!(inode->i_state & I_SYNC_QUEUED)) {
> +			if ((inode->i_state & I_DIRTY))
> +				redirty_tail_locked(inode, wb);
> +			else if (inode->i_state & I_DIRTY_TIME) {
> +				inode->dirtied_when = jiffies;
> +				inode_io_list_move_locked(inode,
> +							  wb,
> +							  &wb->b_dirty_time);
> +			}
>  		}
>  	}
>
  

Thank you for noticing that!
I will send a fixed patch in 8-10 hours if no other comment will arrive.

On Mon, Nov 14, 2022 at 11:25 PM Randy Dunlap <rdunlap@infradead.org> wrote:
>
> Hi--
>
> Please see a small nit below.
>
> On 11/14/22 13:21, Svyatoslav Feldsherov wrote:
> > After commit cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode
> > already has I_DIRTY_INODE") writeiback_single_inode can push inode with
> > I_DIRTY_TIME set to b_dirty_time list. In case of freeing inode with
> > I_DIRTY_TIME set this can happened after deletion of inode io_list at
> > evict. Stack trace is following.
> >
> > evict
> > fat_evict_inode
> > fat_truncate_blocks
> > fat_flush_inodes
> > writeback_inode
> > sync_inode_metadata(inode, sync=0)
> > writeback_single_inode(inode, wbc) <- wbc->sync_mode == WB_SYNC_NONE
> >
> > This will lead to use after free in flusher thread.
> >
> > Similar issue can be triggered if writeback_single_inode in the
> > stack trace update inode->io_list. Add explicit check to avoid it.
> >
> > Fixes: cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode already has I_DIRTY_INODE")
> > Reported-by: syzbot+6ba92bd00d5093f7e371@syzkaller.appspotmail.com
> > Signed-off-by: Svyatoslav Feldsherov <feldsherov@google.com>
> > ---
> >  V1 -> V2:
> >  - address review comments
> >  - skip inode_cgwb_move_to_attached for freeing inode
> >
> >  fs/fs-writeback.c | 30 +++++++++++++++++++-----------
> >  1 file changed, 19 insertions(+), 11 deletions(-)
> >
> > diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
> > index 443f83382b9b..c4aea096689c 100644
> > --- a/fs/fs-writeback.c
> > +++ b/fs/fs-writeback.c
> > @@ -1712,18 +1712,26 @@ static int writeback_single_inode(struct inode *inode,
> >       wb = inode_to_wb_and_lock_list(inode);
> >       spin_lock(&inode->i_lock);
> >       /*
> > -      * If the inode is now fully clean, then it can be safely removed from
> > -      * its writeback list (if any).  Otherwise the flusher threads are
> > -      * responsible for the writeback lists.
> > +      * If the inode is freeing, it's io_list shoudn't be updated
>
>                                     its
>
> > +      * as it can be finally deleted at this moment.
> >        */
> > -     if (!(inode->i_state & I_DIRTY_ALL))
> > -             inode_cgwb_move_to_attached(inode, wb);
> > -     else if (!(inode->i_state & I_SYNC_QUEUED)) {
> > -             if ((inode->i_state & I_DIRTY))
> > -                     redirty_tail_locked(inode, wb);
> > -             else if (inode->i_state & I_DIRTY_TIME) {
> > -                     inode->dirtied_when = jiffies;
> > -                     inode_io_list_move_locked(inode, wb, &wb->b_dirty_time);
> > +     if (!(inode->i_state & I_FREEING)) {
> > +             /*
> > +              * If the inode is now fully clean, then it can be safely
> > +              * removed from its writeback list (if any). Otherwise the
> > +              * flusher threads are responsible for the writeback lists.
> > +              */
> > +             if (!(inode->i_state & I_DIRTY_ALL))
> > +                     inode_cgwb_move_to_attached(inode, wb);
> > +             else if (!(inode->i_state & I_SYNC_QUEUED)) {
> > +                     if ((inode->i_state & I_DIRTY))
> > +                             redirty_tail_locked(inode, wb);
> > +                     else if (inode->i_state & I_DIRTY_TIME) {
> > +                             inode->dirtied_when = jiffies;
> > +                             inode_io_list_move_locked(inode,
> > +                                                       wb,
> > +                                                       &wb->b_dirty_time);
> > +                     }
> >               }
> >       }
> >
>
> --
> ~Randy

--
Slava
  

On Mon 14-11-22 21:21:55, Svyatoslav Feldsherov wrote:
> After commit cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode
> already has I_DIRTY_INODE") writeiback_single_inode can push inode with
				^^^ writeback

> I_DIRTY_TIME set to b_dirty_time list. In case of freeing inode with
> I_DIRTY_TIME set this can happened after deletion of inode io_list at
				^^ happen		^^^ deletion of
inode *from i_io_list*

> evict. Stack trace is following.
> 
> evict
> fat_evict_inode
> fat_truncate_blocks
> fat_flush_inodes
> writeback_inode
> sync_inode_metadata(inode, sync=0)
> writeback_single_inode(inode, wbc) <- wbc->sync_mode == WB_SYNC_NONE
> 
> This will lead to use after free in flusher thread.
> 
> Similar issue can be triggered if writeback_single_inode in the
> stack trace update inode->io_list. Add explicit check to avoid it.
			^^ inode->i_io_list
 
> Fixes: cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode already has I_DIRTY_INODE")
> Reported-by: syzbot+6ba92bd00d5093f7e371@syzkaller.appspotmail.com
> Signed-off-by: Svyatoslav Feldsherov <feldsherov@google.com>

Besides these gramatical nits the patch looks good to me. Feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

Thanks!

								Honza

> ---
>  V1 -> V2: 
>  - address review comments
>  - skip inode_cgwb_move_to_attached for freeing inode 
> 
>  fs/fs-writeback.c | 30 +++++++++++++++++++-----------
>  1 file changed, 19 insertions(+), 11 deletions(-)
> 
> diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
> index 443f83382b9b..c4aea096689c 100644
> --- a/fs/fs-writeback.c
> +++ b/fs/fs-writeback.c
> @@ -1712,18 +1712,26 @@ static int writeback_single_inode(struct inode *inode,
>  	wb = inode_to_wb_and_lock_list(inode);
>  	spin_lock(&inode->i_lock);
>  	/*
> -	 * If the inode is now fully clean, then it can be safely removed from
> -	 * its writeback list (if any).  Otherwise the flusher threads are
> -	 * responsible for the writeback lists.
> +	 * If the inode is freeing, it's io_list shoudn't be updated
> +	 * as it can be finally deleted at this moment.
>  	 */
> -	if (!(inode->i_state & I_DIRTY_ALL))
> -		inode_cgwb_move_to_attached(inode, wb);
> -	else if (!(inode->i_state & I_SYNC_QUEUED)) {
> -		if ((inode->i_state & I_DIRTY))
> -			redirty_tail_locked(inode, wb);
> -		else if (inode->i_state & I_DIRTY_TIME) {
> -			inode->dirtied_when = jiffies;
> -			inode_io_list_move_locked(inode, wb, &wb->b_dirty_time);
> +	if (!(inode->i_state & I_FREEING)) {
> +		/*
> +		 * If the inode is now fully clean, then it can be safely
> +		 * removed from its writeback list (if any). Otherwise the
> +		 * flusher threads are responsible for the writeback lists.
> +		 */
> +		if (!(inode->i_state & I_DIRTY_ALL))
> +			inode_cgwb_move_to_attached(inode, wb);
> +		else if (!(inode->i_state & I_SYNC_QUEUED)) {
> +			if ((inode->i_state & I_DIRTY))
> +				redirty_tail_locked(inode, wb);
> +			else if (inode->i_state & I_DIRTY_TIME) {
> +				inode->dirtied_when = jiffies;
> +				inode_io_list_move_locked(inode,
> +							  wb,
> +							  &wb->b_dirty_time);
> +			}
>  		}
>  	}
>  
> -- 
> 2.38.1.431.g37b22c650d-goog
>
  

Thank you for the review.

I've sent v3 with proposed fixes.
Also tried to be more consistent and use i_io_list in comments and
commit message instead of io_list//io list.

On Tue, Nov 15, 2022 at 12:55 PM Jan Kara <jack@suse.cz> wrote:
>
> On Mon 14-11-22 21:21:55, Svyatoslav Feldsherov wrote:
> > After commit cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode
> > already has I_DIRTY_INODE") writeiback_single_inode can push inode with
>                                 ^^^ writeback
>
> > I_DIRTY_TIME set to b_dirty_time list. In case of freeing inode with
> > I_DIRTY_TIME set this can happened after deletion of inode io_list at
>                                 ^^ happen               ^^^ deletion of
> inode *from i_io_list*
>
> > evict. Stack trace is following.
> >
> > evict
> > fat_evict_inode
> > fat_truncate_blocks
> > fat_flush_inodes
> > writeback_inode
> > sync_inode_metadata(inode, sync=0)
> > writeback_single_inode(inode, wbc) <- wbc->sync_mode == WB_SYNC_NONE
> >
> > This will lead to use after free in flusher thread.
> >
> > Similar issue can be triggered if writeback_single_inode in the
> > stack trace update inode->io_list. Add explicit check to avoid it.
>                         ^^ inode->i_io_list
>
> > Fixes: cbfecb927f42 ("fs: record I_DIRTY_TIME even if inode already has I_DIRTY_INODE")
> > Reported-by: syzbot+6ba92bd00d5093f7e371@syzkaller.appspotmail.com
> > Signed-off-by: Svyatoslav Feldsherov <feldsherov@google.com>
>
> Besides these gramatical nits the patch looks good to me. Feel free to add:
>
> Reviewed-by: Jan Kara <jack@suse.cz>
>
> Thanks!
>
>                                                                 Honza
>
> > ---
> >  V1 -> V2:
> >  - address review comments
> >  - skip inode_cgwb_move_to_attached for freeing inode
> >
> >  fs/fs-writeback.c | 30 +++++++++++++++++++-----------
> >  1 file changed, 19 insertions(+), 11 deletions(-)
> >
> > diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
> > index 443f83382b9b..c4aea096689c 100644
> > --- a/fs/fs-writeback.c
> > +++ b/fs/fs-writeback.c
> > @@ -1712,18 +1712,26 @@ static int writeback_single_inode(struct inode *inode,
> >       wb = inode_to_wb_and_lock_list(inode);
> >       spin_lock(&inode->i_lock);
> >       /*
> > -      * If the inode is now fully clean, then it can be safely removed from
> > -      * its writeback list (if any).  Otherwise the flusher threads are
> > -      * responsible for the writeback lists.
> > +      * If the inode is freeing, it's io_list shoudn't be updated
> > +      * as it can be finally deleted at this moment.
> >        */
> > -     if (!(inode->i_state & I_DIRTY_ALL))
> > -             inode_cgwb_move_to_attached(inode, wb);
> > -     else if (!(inode->i_state & I_SYNC_QUEUED)) {
> > -             if ((inode->i_state & I_DIRTY))
> > -                     redirty_tail_locked(inode, wb);
> > -             else if (inode->i_state & I_DIRTY_TIME) {
> > -                     inode->dirtied_when = jiffies;
> > -                     inode_io_list_move_locked(inode, wb, &wb->b_dirty_time);
> > +     if (!(inode->i_state & I_FREEING)) {
> > +             /*
> > +              * If the inode is now fully clean, then it can be safely
> > +              * removed from its writeback list (if any). Otherwise the
> > +              * flusher threads are responsible for the writeback lists.
> > +              */
> > +             if (!(inode->i_state & I_DIRTY_ALL))
> > +                     inode_cgwb_move_to_attached(inode, wb);
> > +             else if (!(inode->i_state & I_SYNC_QUEUED)) {
> > +                     if ((inode->i_state & I_DIRTY))
> > +                             redirty_tail_locked(inode, wb);
> > +                     else if (inode->i_state & I_DIRTY_TIME) {
> > +                             inode->dirtied_when = jiffies;
> > +                             inode_io_list_move_locked(inode,
> > +                                                       wb,
> > +                                                       &wb->b_dirty_time);
> > +                     }
> >               }
> >       }
> >
> > --
> > 2.38.1.431.g37b22c650d-goog
> >
> --
> Jan Kara <jack@suse.com>
> SUSE Labs, CR

--
Slava