On Thu, Feb 16, 2017 at 12:35:23AM +0800, colyli@xxxxxxx wrote: > When I run a parallel reading performan testing on a md raid1 device with > two NVMe SSDs, I observe very bad throughput in supprise: by fio with 64KB > block size, 40 seq read I/O jobs, 128 iodepth, overall throughput is > only 2.7GB/s, this is around 50% of the idea performance number. > > The perf reports locking contention happens at allow_barrier() and > wait_barrier() code, > - 41.41% fio [kernel.kallsyms] [k] _raw_spin_lock_irqsave > - _raw_spin_lock_irqsave > + 89.92% allow_barrier > + 9.34% __wake_up > - 37.30% fio [kernel.kallsyms] [k] _raw_spin_lock_irq > - _raw_spin_lock_irq > - 100.00% wait_barrier > > The reason is, in these I/O barrier related functions, > - raise_barrier() > - lower_barrier() > - wait_barrier() > - allow_barrier() > They always hold conf->resync_lock firstly, even there are only regular > reading I/Os and no resync I/O at all. This is a huge performance penalty. > > The solution is a lockless-like algorithm in I/O barrier code, and only > holding conf->resync_lock when it has to. > > The original idea is from Hannes Reinecke, and Neil Brown provides > comments to improve it. I continue to work on it, and make the patch into > current form. > > In the new simpler raid1 I/O barrier implementation, there are two > wait barrier functions, > - wait_barrier() > Which calls _wait_barrier(), is used for regular write I/O. If there is > resync I/O happening on the same I/O barrier bucket, or the whole > array is frozen, task will wait until no barrier on same barrier bucket, > or the whold array is unfreezed. > - wait_read_barrier() > Since regular read I/O won't interfere with resync I/O (read_balance() > will make sure only uptodate data will be read out), it is unnecessary > to wait for barrier in regular read I/Os, waiting in only necessary > when the whole array is frozen. > > The operations on conf->nr_pending[idx], conf->nr_waiting[idx], conf-> > barrier[idx] are very carefully designed in raise_barrier(), > lower_barrier(), _wait_barrier() and wait_read_barrier(), in order to > avoid unnecessary spin locks in these functions. Once conf-> > nr_pengding[idx] is increased, a resync I/O with same barrier bucket index > has to wait in raise_barrier(). Then in _wait_barrier() if no barrier > raised in same barrier bucket index and array is not frozen, the regular > I/O doesn't need to hold conf->resync_lock, it can just increase > conf->nr_pending[idx], and return to its caller. wait_read_barrier() is > very similar to _wait_barrier(), the only difference is it only waits when > array is frozen. For heavy parallel reading I/Os, the lockless I/O barrier > code almostly gets rid of all spin lock cost. > > This patch significantly improves raid1 reading peroformance. From my > testing, a raid1 device built by two NVMe SSD, runs fio with 64KB > blocksize, 40 seq read I/O jobs, 128 iodepth, overall throughput > increases from 2.7GB/s to 4.6GB/s (+70%). > > Changelog > V3: > - Add smp_mb__after_atomic() as Shaohua and Neil suggested. > - Change conf->nr_queued[] from atomic_t to int. I missed this part. In the code, the nr_queued sometimes is protected by device_lock, sometimes (raid1d) no protection at all. Can you explain this? Thanks, Shaohua -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
