diff options
author | Petri Savolainen <petri.savolainen@linaro.org> | 2018-09-12 17:20:10 +0300 |
---|---|---|
committer | Maxim Uvarov <maxim.uvarov@linaro.org> | 2018-09-28 10:09:06 +0300 |
commit | d6a88b032af4ad2720da360190153fc020d5d3f4 (patch) | |
tree | 1eb41cb68c42ea2a261dcbee1bb65f58ae765f84 /config | |
parent | f0611da0be1f4002c5f49933204451eafe4f0219 (diff) |
linux-gen: sched: per priority burst size configuration
Change burst size configuration to be per priority instead of
only two levels of control (high/low priority). Also maximum
burst size is configurable, so that application may request
a large burst of events without a worry that a large burst of
low priority events is received.
Signed-off-by: Petri Savolainen <petri.savolainen@linaro.org>
Reviewed-by: Bill Fischofer <bill.fischofer@linaro.org>
Signed-off-by: Maxim Uvarov <maxim.uvarov@linaro.org>
Diffstat (limited to 'config')
-rw-r--r-- | config/odp-linux-generic.conf | 25 |
1 files changed, 14 insertions, 11 deletions
diff --git a/config/odp-linux-generic.conf b/config/odp-linux-generic.conf index bddc92dd4..4fe02cf29 100644 --- a/config/odp-linux-generic.conf +++ b/config/odp-linux-generic.conf @@ -16,7 +16,7 @@ # Mandatory fields odp_implementation = "linux-generic" -config_file_version = "0.0.1" +config_file_version = "0.1.0" # Shared memory options shm: { @@ -73,14 +73,17 @@ sched_basic: { # value is the number of threads using the scheduler. prio_spread = 4 - # Default burst sizes for high and low priority queues. The default - # and higher priority levels are considered as high. Scheduler - # rounds up number of requested events up to these values. In general, - # larger requests are not round down. So, larger bursts than these may - # received when requested. A large burst size improves throughput, - # but decreases application responsiveness to high priority events - # due to head of line blocking cause by a burst of low priority - # events. - burst_size_hi = 32 - burst_size_low = 16 + # Burst size configuration per priority. The first array element + # represents the highest queue priority. The scheduler tries to get + # burst_size_default[prio] events from a queue and stashes those that + # cannot be passed to the application immediately. More events than the + # default burst size may be returned from application request, but no + # more than burst_size_max[prio]. + # + # Large burst sizes improve throughput, but decrease application + # responsiveness to higher priority events due to head of line blocking + # caused by a burst of lower priority events. + burst_size_default = [ 32, 32, 32, 32, 32, 16, 8, 4] + burst_size_max = [255, 255, 255, 255, 255, 16, 16, 8] + } |