This patchset does things a bit differently from the previous attempts
to find the journal head based on Andreas' suggestions.
It uses a pool of bios to maintain a readahead queue of sorts that allows
us to process the completed bios in sequential order to locate the jhead.
I've done a little bit of testing and it seems to be holding up so far.
I plan to do more testing.
I haven't done a performance analysis vs the old method yet, so I don't
know how well this does. There might be some optimizations we can do w.r.t
repeated allocations and such.
Abhi Das (3):
gfs2: add more timing info to the journal recovery process
gfs2: add a helper function to get_log_header that can be used
elsewhere
gfs2: introduce bio_pool to readahead journal to find jhead
fs/gfs2/bmap.c | 8 +-
fs/gfs2/incore.h | 3 +
fs/gfs2/lops.c | 359 +++++++++++++++++++++++++++++++++++++++++++++++++++
fs/gfs2/lops.h | 1 +
fs/gfs2/ops_fstype.c | 2 +
fs/gfs2/recovery.c | 171 ++++++------------------
fs/gfs2/recovery.h | 2 +
fs/gfs2/sys.c | 27 ++--
8 files changed, 433 insertions(+), 140 deletions(-)
--
2.4.11
