Lines Matching defs:scan
1436 * Approximate time to scan a full NUMA task in ms. The task scan period is
1443 /* Portion of address space to scan in MB */
1510 /* For sanity's sake, never scan more PTEs than MAX_SCAN_WINDOW MB/sec. */
1516 unsigned int scan, floor;
1523 scan = sysctl_numa_balancing_scan_period_min / task_nr_scan_windows(p);
1524 return max_t(unsigned int, floor, scan);
1533 /* Scale the maximum scan period with the amount of shared memory. */
1558 /* Scale the maximum scan period with the amount of shared memory. */
1796 * used to record scan time instead of CPU and PID. When tiering mode
1797 * is disabled at run time, the scan time (in cpupid) will be
1835 * For memory tiering mode, when page tables are scanned, the scan
1841 * hint page fault latency = hint page fault time - scan time
2669 * When adapting the scan rate, the period is divided into NUMA_PERIOD_SLOTS
2670 * increments. The more local the fault statistics are, the higher the scan
2671 * period will be for the next scan window. If local/(local+remote) ratio is
2673 * the scan period will decrease. Aim for 70% local accesses.
2679 * Increase the scan period (slow down scanning) if the majority of
2682 * Otherwise, decrease the scan period.
2699 * node is overloaded. In either case, scan slower
2712 * Prepare to scale scan period relative to the current period.
2713 * == NUMA_PERIOD_THRESHOLD scan period stays the same
2714 * < NUMA_PERIOD_THRESHOLD scan period decreases (scan faster)
2715 * >= NUMA_PERIOD_THRESHOLD scan period increases (scan slower)
2922 /* Decay existing window, copy faults since last scan */
3213 * scan rate to slow down when a workload has settled down.
3308 * Enforce maximal scan/migration frequency..
3381 /* Reset happens after 4 times scan delay of scan start */
3388 * delay the scan for new VMAs.
3394 /* Do not scan the VMA if task has not accessed */
3417 * Try to scan sysctl_numa_balancing_size worth of
3421 * PTEs, scan up to virtpages, to skip through those
3440 * would find the !migratable VMA on the next scan but not reset the
3556 * Allow resets if faults have been trapped before one scan
3562 * Avoid scan adjustments if moving to the preferred
7443 * comparing the average scan cost (tracked in sd->avg_scan_cost) against the
7493 /* because !--nr is the condition to stop scan */
7525 * Account for the scan cost of wakeups against the average
10770 * Update the number of CPUs to scan in LLC domain, which could
10790 * sum_util hits 85% or above, the scan stops.
10800 * and the number of CPUs to scan is calculated by: