diff options
| author | Eivind Sarto <[email protected]> | 2014-06-10 00:06:19 +0000 |
|---|---|---|
| committer | NeilBrown <[email protected]> | 2014-06-10 01:02:01 +0000 |
| commit | 053f5b6525ae32da397e6c47721961f800d2c924 (patch) | |
| tree | 2846e90e25ae2b5f998bffd2d22119695d139f23 /tools/perf/scripts/python/syscall-counts-by-pid.py | |
| parent | md/raid5: deadlock between retry_aligned_read with barrier io (diff) | |
| download | kernel-053f5b6525ae32da397e6c47721961f800d2c924.tar.gz kernel-053f5b6525ae32da397e6c47721961f800d2c924.zip | |
raid5: speedup sync_request processing
The raid5 sync_request() processing calls handle_stripe() within the context of
the resync-thread. The resync-thread issues the first set of read requests
and this adds execution latency and slows down the scheduling of the next
sync_request().
The current rebuild/resync speed of raid5 is not much faster than what
rotational HDDs can sustain.
Testing the following patch on a 6-drive array, I can increase the rebuild
speed from 100 MB/s to 175 MB/s.
The sync_request() now just sets STRIPE_HANDLE and releases the stripe. This
creates some more parallelism between the resync-thread and raid5 kernel daemon.
Signed-off-by: Eivind Sarto <[email protected]>
Signed-off-by: NeilBrown <[email protected]>
Diffstat (limited to 'tools/perf/scripts/python/syscall-counts-by-pid.py')
0 files changed, 0 insertions, 0 deletions