[Top] [All Lists]

Re: [PATCH v4 0/6] speculative preallocation quota throttling

To: Brian Foster <bfoster@xxxxxxxxxx>
Subject: Re: [PATCH v4 0/6] speculative preallocation quota throttling
From: Mark Tinguely <tinguely@xxxxxxx>
Date: Fri, 22 Feb 2013 14:09:04 -0600
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <1361373019-30891-1-git-send-email-bfoster@xxxxxxxxxx>
References: <1361373019-30891-1-git-send-email-bfoster@xxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:9.0) Gecko/20120122 Thunderbird/9.0
On 02/20/13 09:10, Brian Foster wrote:
Hi all,

Here is v4 of the speculative preallocation quota throttling set. Patch 3 of v3
is dropped due to a redundancy and a few changes are made to the threshold
management code (patch 4/6). The remaining patches are unchanged from v3.


- Drop patch 3/7 from v3 (xfs: cap prealloc size to free space before shift).
- Several updates to patch 4/6:
   - Rename xfs_dquot_init_prealloc() to xfs_dquot_set_prealloc_limits().
   - Unroll the prealloc threshold loop and remove the increment def.
   - Fix up some comments.
- Rebased on top of updated speculative preallocation algorithm.
- Fix up xfs_iomap_prealloc_size() rounding (patch 2).
- Add pre-calculated fields to xfs_dquot to support throttling.
- Move to logarithmic (shift) throttler and finer tuned trigger/throttle logic.

Brian Foster (6):
   xfs: reorganize xfs_iomap_prealloc_size to remove indentation
   xfs: push rounddown_pow_of_two() to after prealloc throttle
   xfs: pass xfs_dquot to xfs_qm_adjust_dqlimits() instead of
   xfs: xfs_dquot prealloc throttling watermarks and low free space
   xfs: add quota-driven speculative preallocation throttling
   xfs: xfs_iomap_prealloc_size() tracepoint

  fs/xfs/xfs_dquot.c       |   44 ++++++++++++-
  fs/xfs/xfs_dquot.h       |   14 ++++-
  fs/xfs/xfs_iomap.c       |  161 +++++++++++++++++++++++++++++++++++++---------
  fs/xfs/xfs_qm.c          |    2 +-
  fs/xfs/xfs_qm_syscalls.c |    1 +
  fs/xfs/xfs_trace.h       |   24 +++++++
  fs/xfs/xfs_trans_dquot.c |    2 +-
  7 files changed, 212 insertions(+), 36 deletions(-)

The rest of the series looks good to me. Consider it Review-by: me.

The tracing in patch 6 says how much, but not why, BUT I think the why
would require multiple tracepoints, so I okay with it.

        -- since I am being a grouch ... ---
I wish I had caught in the review for commit a1e16c26 that xfs_iomap_eof_prealloc_initial_size() is returning an int rather than xfs_fsblock_t for consistency.


<Prev in Thread] Current Thread [Next in Thread>