On Thu, Oct 01, 2020 at 09:20:15PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
>
> When xfs_defer_capture extracts the deferred ops and transaction state
> from a transaction, it should record the remaining block reservations so
> that when we continue the dfops chain, we can reserve the same number of
> blocks to use. We capture the reservations for both data and realtime
> volumes.
>
> This adds the requirement that every log intent item recovery function
> must be careful to reserve enough blocks to handle both itself and all
> defer ops that it can queue. On the other hand, this enables us to do
> away with the handwaving block estimation nonsense that was going on in
> xlog_finish_defer_ops.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> ---
> v4.2: don't fiddle with transaction internals, and save the rt
> reservation too
> ---
Reviewed-by: Brian Foster <bfoster@xxxxxxxxxx>
> fs/xfs/libxfs/xfs_defer.c | 4 ++++
> fs/xfs/libxfs/xfs_defer.h | 4 ++++
> fs/xfs/xfs_log_recover.c | 21 +++------------------
> 3 files changed, 11 insertions(+), 18 deletions(-)
>
> diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c
> index 85c371d29e8d..10aeae7353ab 100644
> --- a/fs/xfs/libxfs/xfs_defer.c
> +++ b/fs/xfs/libxfs/xfs_defer.c
> @@ -575,6 +575,10 @@ xfs_defer_ops_capture(
> dfc->dfc_tpflags = tp->t_flags & XFS_TRANS_LOWMODE;
> tp->t_flags &= ~XFS_TRANS_LOWMODE;
>
> + /* Capture the remaining block reservations along with the dfops. */
> + dfc->dfc_blkres = tp->t_blk_res - tp->t_blk_res_used;
> + dfc->dfc_rtxres = tp->t_rtx_res - tp->t_rtx_res_used;
> +
> return dfc;
> }
>
> diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
> index 3af82ebc1249..5c0e59b69ffa 100644
> --- a/fs/xfs/libxfs/xfs_defer.h
> +++ b/fs/xfs/libxfs/xfs_defer.h
> @@ -75,6 +75,10 @@ struct xfs_defer_capture {
> /* Deferred ops state saved from the transaction. */
> struct list_head dfc_dfops;
> unsigned int dfc_tpflags;
> +
> + /* Block reservations for the data and rt devices. */
> + unsigned int dfc_blkres;
> + unsigned int dfc_rtxres;
> };
>
> /*
> diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
> index 7804906d145b..1be5208e2a2f 100644
> --- a/fs/xfs/xfs_log_recover.c
> +++ b/fs/xfs/xfs_log_recover.c
> @@ -2439,27 +2439,12 @@ xlog_finish_defer_ops(
> {
> struct xfs_defer_capture *dfc, *next;
> struct xfs_trans *tp;
> - int64_t freeblks;
> - uint64_t resblks;
> int error = 0;
>
> list_for_each_entry_safe(dfc, next, capture_list, dfc_list) {
> - /*
> - * We're finishing the defer_ops that accumulated as a result
> - * of recovering unfinished intent items during log recovery.
> - * We reserve an itruncate transaction because it is the
> - * largest permanent transaction type. Since we're the only
> - * user of the fs right now, take 93% (15/16) of the available
> - * free blocks. Use weird math to avoid a 64-bit division.
> - */
> - freeblks = percpu_counter_sum(&mp->m_fdblocks);
> - if (freeblks <= 0)
> - return -ENOSPC;
> -
> - resblks = min_t(uint64_t, UINT_MAX, freeblks);
> - resblks = (resblks * 15) >> 4;
> - error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, resblks,
> - 0, XFS_TRANS_RESERVE, &tp);
> + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate,
> + dfc->dfc_blkres, dfc->dfc_rtxres,
> + XFS_TRANS_RESERVE, &tp);
> if (error)
> return error;
>
>
