Timer and scheduler parameters

Parameter	Dflt	Min	Max 64/256/1024	Auto
FC_AFFINITY_ON	1	0	1
FCMAXUPRI	20	10	30
FP_AFFINITY_ON	1	0	1
INITCLASS	TS
MAXCLSYSPRI	99	40	500
NCALL	80	30	1000	KVM, Max only
TS_AFFINITY_ON	1	0	1
TS_SLEEPMAX	60	10	6000
TS_SLEEPWAIT	4	1	500
TSMAXUPRI	20	10	30

FC_AFFINITY_ON

FP_AFFINITY_ON

The setting indicates whether a lightweight process (LWP) of the scheduler class (fixed class or fixed priority class, respectively) should be scheduled on the same processor that it ran on last, when the cache is still warm. A setting of ``1'' indicates that the LWP should be run on the same processor when the cache is still warm. See ``Processor cache affinity parameters'' for how it is determined whether the cache is warm.

FCMAXUPRI

Defines the range for the fixed class scheduler priorities. This is equivalent to TSMAXUPRI, which is used for the time sharing class.

INITCLASS

INITCLASS is the scheduler class assigned to the init process. This scheduler class is inherited by all descendants of init, which normally include all user login shells. INITCLASS is a string tunable and by default, it is TS; that is, all login shells are time-sharing processes in the default configuration.

MAXCLSYSPRI

Maximum global priority used by the SYS scheduling class for scheduling kernel processes. When the kernel starts system processes, it assigns their priorities using MAXCLSYSPRI as a reference point.

Changing this parameter changes the range of priorities used to schedule kernel processes and can have a significant effect on the performance of the system. In general, there is no need to change this parameter unless you add new scheduling classes or reconfigure the priorities of other currently configured classes.

The value of MAXCLSYSPRI must be ``40'' or greater, because the kernel assumes it has at least that great a range of priorities below MAXCLSYSPRI. If you request a MAXCLSYSPRI below ``40'', the system might panic during initialization.

The most important system processes get global priorities at or near MAXCLSYSPRI; the least important system processes get global priorities at or near (MAXCLSYSPRI - 40). The default value of MAXCLSYSPRI is ``99'', which gives all system processes higher priorities than all user processes.

NCALL

Specifies the size of the private reserve of callout table entries. The callout reserve is used by device drivers and by the kernel to perform periodic tasks, to detect timeouts, and so forth. Each callout table entry represents a timed event scheduled for later execution by the kernel clock handler. Entries are generally allocated dynamically. The private reserve exists as a fallback to be used when dynamic allocation is impossible. This value must be greater than ``30'' and is typically in the range of ``50'' to ``250'', although the maximum value is autotuned.

TS_AFFINITY_ON

The setting indicates whether a lightweight process (LWP) of the time sharing scheduler class should be scheduled on the same processor that it ran on last, when the cache is still warm. A setting of ``1'' indicates that the LWP should be run on the same processor when the cache is still warm. See ``Processor cache affinity parameters'' for how it is determined whether the cache is warm.

TS_SLEEPMAX

When a time sharing LWP is given a new priority and time quantum, as a result of exceeding TS_SLEEPWAIT accumulated sleep time, its accumulated sleep time is reduced to a value that is no greater than half of TS_SLEEPMAX. Alter this tunable parameter with caution since it will change the time-slicing behavior of the time sharing scheduler.

TS_SLEEPWAIT

A time sharing LWP that has accumulated a total of TS_SLEEPWAIT clock ticks of sleep time is given a new time quantum and a new (typically higher) priority. Alter this tunable parameter with caution since it will change the time-slicing behavior of the time sharing scheduler.

TSMAXUPRI

Defines the range for time sharing scheduler priorities. The range within which users can adjust the user priority of a time sharing process is -TSMAXUPRI to +TSMAXUPRI. Configuring higher values gives users more control over the priority of their processes. Remember that only a process with appropriate privilege can raise priority. The default value of ``20'' provides a degree of control equivalent to what has been available in the past through the nice(S-osr5) interface.

The value of TSMAXUPRI is independent of the configured number of global time-sharing priorities, though we recommend configuring at least 40 time-sharing priorities. In the default configuration, there are 60 time-sharing priorities. The system may use the remaining priorities depending on process behavior.