Re: [PATCHv3 2/3] z3fold: change per-pool spinlock to rwlock

2016-11-01 Thread Vitaly Wool
On Tue, Nov 1, 2016 at 9:16 PM, Dan Streetman wrote: > On Thu, Oct 27, 2016 at 7:12 AM, Vitaly Wool wrote: >> Mapping/unmapping goes with no actual modifications so it makes >> sense to only take a read lock in map/unmap functions. >> >> This change gives up to 10% performance gain and lower late

Re: [PATCHv3 2/3] z3fold: change per-pool spinlock to rwlock

2016-11-01 Thread Dan Streetman
On Thu, Oct 27, 2016 at 7:12 AM, Vitaly Wool wrote: > Mapping/unmapping goes with no actual modifications so it makes > sense to only take a read lock in map/unmap functions. > > This change gives up to 10% performance gain and lower latencies > in fio sequential read/write tests, e.g. before: > >

[PATCHv3 2/3] z3fold: change per-pool spinlock to rwlock

2016-10-27 Thread Vitaly Wool
Mapping/unmapping goes with no actual modifications so it makes sense to only take a read lock in map/unmap functions. This change gives up to 10% performance gain and lower latencies in fio sequential read/write tests, e.g. before: Run status group 0 (all jobs): WRITE: io=2700.0GB, aggrb=3234.