Re: [PATCH 5.10 CANDIDATE] xfs: drop submit side trans alloc for append ioends

["Darrick J. Wong" <djwong@kernel.org>]

On Wed, Apr 19, 2023 at 07:32:59AM +0300, Amir Goldstein wrote:
> From: Brian Foster <bfoster@redhat.com>
> 
> commit 7cd3099f4925d7c15887d1940ebd65acd66100f5 upstream.
> 
> Per-inode ioend completion batching has a log reservation deadlock
> vector between preallocated append transactions and transactions
> that are acquired at completion time for other purposes (i.e.,
> unwritten extent conversion or COW fork remaps). For example, if the
> ioend completion workqueue task executes on a batch of ioends that
> are sorted such that an append ioend sits at the tail, it's possible
> for the outstanding append transaction reservation to block
> allocation of transactions required to process preceding ioends in
> the list.
> 
> Append ioend completion is historically the common path for on-disk
> inode size updates. While file extending writes may have completed
> sometime earlier, the on-disk inode size is only updated after
> successful writeback completion. These transactions are preallocated
> serially from writeback context to mitigate concurrency and
> associated log reservation pressure across completions processed by
> multi-threaded workqueue tasks.
> 
> However, now that delalloc blocks unconditionally map to unwritten
> extents at physical block allocation time, size updates via append
> ioends are relatively rare. This means that inode size updates most
> commonly occur as part of the preexisting completion time
> transaction to convert unwritten extents. As a result, there is no
> longer a strong need to preallocate size update transactions.
> 
> Remove the preallocation of inode size update transactions to avoid
> the ioend completion processing log reservation deadlock. Instead,
> continue to send all potential size extending ioends to workqueue
> context for completion and allocate the transaction from that
> context. This ensures that no outstanding log reservation is owned
> by the ioend completion worker task when it begins to process
> ioends.
> 
> Signed-off-by: Brian Foster <bfoster@redhat.com>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> Reported-by: Christian Theune <ct@flyingcircus.io>
> Link: https://lore.kernel.org/linux-xfs/CAOQ4uxjj2UqA0h4Y31NbmpHksMhVrXfXjLG4Tnz3zq_UR-3gSA@mail.gmail.com/
> Signed-off-by: Amir Goldstein <amir73il@gmail.com>
> ---
> 
> Darrick,
> 
> As I wrote in $LINK, this fix from v5.13 was missed from my backports
> due to a tool error.
> 
> It's also in good timeing because Chandan has just finished the v5.12
> backports.
> 
> I ran it though 10 kdevops cycles.
> 
> Please approve.

Looks good to me,
Acked-by: Darrick J. Wong <djwong@kernel.org>

--D

> 
> Thanks,
> Amir.
> 
>  fs/xfs/xfs_aops.c | 45 +++------------------------------------------
>  1 file changed, 3 insertions(+), 42 deletions(-)
> 
> diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
> index 953de843d9c3..e341d6531e68 100644
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -39,33 +39,6 @@ static inline bool xfs_ioend_is_append(struct iomap_ioend *ioend)
>  		XFS_I(ioend->io_inode)->i_d.di_size;
>  }
>  
> -STATIC int
> -xfs_setfilesize_trans_alloc(
> -	struct iomap_ioend	*ioend)
> -{
> -	struct xfs_mount	*mp = XFS_I(ioend->io_inode)->i_mount;
> -	struct xfs_trans	*tp;
> -	int			error;
> -
> -	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_fsyncts, 0, 0, 0, &tp);
> -	if (error)
> -		return error;
> -
> -	ioend->io_private = tp;
> -
> -	/*
> -	 * We may pass freeze protection with a transaction.  So tell lockdep
> -	 * we released it.
> -	 */
> -	__sb_writers_release(ioend->io_inode->i_sb, SB_FREEZE_FS);
> -	/*
> -	 * We hand off the transaction to the completion thread now, so
> -	 * clear the flag here.
> -	 */
> -	xfs_trans_clear_context(tp);
> -	return 0;
> -}
> -
>  /*
>   * Update on-disk file size now that data has been written to disk.
>   */
> @@ -191,12 +164,10 @@ xfs_end_ioend(
>  		error = xfs_reflink_end_cow(ip, offset, size);
>  	else if (ioend->io_type == IOMAP_UNWRITTEN)
>  		error = xfs_iomap_write_unwritten(ip, offset, size, false);
> -	else
> -		ASSERT(!xfs_ioend_is_append(ioend) || ioend->io_private);
>  
> +	if (!error && xfs_ioend_is_append(ioend))
> +		error = xfs_setfilesize(ip, ioend->io_offset, ioend->io_size);
>  done:
> -	if (ioend->io_private)
> -		error = xfs_setfilesize_ioend(ioend, error);
>  	iomap_finish_ioends(ioend, error);
>  	memalloc_nofs_restore(nofs_flag);
>  }
> @@ -246,7 +217,7 @@ xfs_end_io(
>  
>  static inline bool xfs_ioend_needs_workqueue(struct iomap_ioend *ioend)
>  {
> -	return ioend->io_private ||
> +	return xfs_ioend_is_append(ioend) ||
>  		ioend->io_type == IOMAP_UNWRITTEN ||
>  		(ioend->io_flags & IOMAP_F_SHARED);
>  }
> @@ -259,8 +230,6 @@ xfs_end_bio(
>  	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
>  	unsigned long		flags;
>  
> -	ASSERT(xfs_ioend_needs_workqueue(ioend));
> -
>  	spin_lock_irqsave(&ip->i_ioend_lock, flags);
>  	if (list_empty(&ip->i_ioend_list))
>  		WARN_ON_ONCE(!queue_work(ip->i_mount->m_unwritten_workqueue,
> @@ -510,14 +479,6 @@ xfs_prepare_ioend(
>  				ioend->io_offset, ioend->io_size);
>  	}
>  
> -	/* Reserve log space if we might write beyond the on-disk inode size. */
> -	if (!status &&
> -	    ((ioend->io_flags & IOMAP_F_SHARED) ||
> -	     ioend->io_type != IOMAP_UNWRITTEN) &&
> -	    xfs_ioend_is_append(ioend) &&
> -	    !ioend->io_private)
> -		status = xfs_setfilesize_trans_alloc(ioend);
> -
>  	memalloc_nofs_restore(nofs_flag);
>  
>  	if (xfs_ioend_needs_workqueue(ioend))
> -- 
> 2.34.1
>