Hello,

here's an update on the log log space management cleanup.  The
notable additions are:

* gfs2: Get rid of current_tail()

  Replace current_tail() with a new sd_log_flush_tail variable in
  the super block.

* gfs2: No revokes for transactions at the tail of the log

  We no longer write out revokes for transactions at the tail
  of the log that have completed; the log header with its
  updated lh_tail will implicitly take care of revoking those.

* gfs2: Rework the log space allocation logic

  Clean up the entire log space allocation code to make it
  more manageable.

* gfs2: Hand out revokes to transactions one by one

  Based on the previous patch, clean up the revoke allocation
  code.

Thanks,
Andreas

Andreas Gruenbacher (20):
  gfs2: Deobfuscate function jdesc_find_i
  gfs2: Simplify the buf_limit and databuf_limit definitions
  gfs2: Minor gfs2_write_revokes cleanups
  gfs2: Some documentation updates
  gfs2: A minor debugging improvement
  gfs2: Rename gfs2_{write => add_aux}_revokes
  gfs2: Clean up ail2_empty
  gfs2: Get rid of on-stack transactions
  gfs2: Get rid of sd_reserving_log
  gfs2: Move lock flush locking to gfs2_trans_{begin,end}
  gfs2: Don't wait for journal flush in clean_journal
  gfs2: Clean up gfs2_log_reserve
  gfs2: Use a tighter bound in gfs2_trans_begin
  gfs2: Get rid of current_tail()
  gfs2: Move function gfs2_ail_empty_tr
  gfs2: No revokes for transactions at the tail of the log
  gfs2: Remove sd_log_committed_revoke
  gfs2: Remove sd_log_blks_reserved
  gfs2: Rework the log space allocation logic
  gfs2: Hand out revokes to transactions one by one

 fs/gfs2/glops.c      |  29 +--
 fs/gfs2/incore.h     |  13 +-
 fs/gfs2/log.c        | 558 +++++++++++++++++++++++++------------------
 fs/gfs2/log.h        |  20 +-
 fs/gfs2/lops.c       |  10 +-
 fs/gfs2/lops.h       |  17 +-
 fs/gfs2/ops_fstype.c |   9 +-
 fs/gfs2/super.c      |  25 +-
 fs/gfs2/trans.c      |  91 ++++---
 fs/gfs2/trans.h      |   2 +
 10 files changed, 444 insertions(+), 330 deletions(-)

-- 
2.26.2

Reply via email to