Lines Matching refs:and

18 (for sequential workloads) and service trees (for random workloads) before
19 queue is expired and CFQ selects next queue to dispatch from.
24 seeks and see improved throughput.
27 level and one should see an overall improved throughput on faster storage
29 side is that isolation provided from WRITES also goes down and notion of
32 So depending on storage and workload, it might be useful to set slice_idle=0.
33 In general I think for SATA/SAS disks and software RAID of SATA/SAS disks
37 throughput and acceptable latencies.
46 direction and consider them as being the "next" if they are within this
73 in higher end storage due to idle on sequential queue and allow dispatch
75 slice_idle=0 and group_idle=8, so that idling does not happen on individual
76 queues in the group but happens overall on the group and thus still keeps the
79 multiple queues in the group at the same time and achieve higher throughput
106 throughput and read latency.
149 CFQ supports blkio cgroup and has "blkio." prefixed files in each
150 blkio cgroup directory. It is weight-based and there are four knobs
151 for configuration - weight[_device] and leaf_weight[_device].
161 composed of five cgroups - root, A, B, AA and AB - with the following
173 and AB have no child and thus its tasks have no children cgroup to
186 If all cgroups have active IOs and competing with each other, disk
196 A has children and further distributes its 57% among the children and
207 process gets bigger time slice and lower priority process gets smaller time
208 slice. Measuring time becomes harder if storage is fast and supports NCQ and
214 single queue and also allow dispatch from multiple cfq queue at the same time.
218 If one sets slice_idle=0 and if storage supports NCQ, CFQ internally switches
219 to IOPS mode and starts providing fairness in terms of number of requests
233 did not move the disk head and kept on dispatching sequential IO from
236 CFQ has following service trees and various queues are put on these trees.
247 we will idle only once last queue has dispatched the IO and there is
253 CFQ has some optimizations for SSDs and if it detects a non-rotational
255 flight at a time), then it cuts down on idling of individual queues and
256 all the queues move to sync-noidle tree and only tree idle remains. This
268     For example, if there are 10 sequential readers doing IO and they get
271     do not idle, and after a couple of milli seconds a another REQ_NOIDLE
273     and notice how a workload can lose its disk share and suffer due to
277     context of fsync, and later some journaling data is written. Journaling
285     provides isolation from multiple sequential readers and at the same
289 A2. I would think whenever one is doing synchronous write and not expecting
291     to specify REQ_NOIDLE on writes and that probably should work well for