62 lines
2.5 KiB
Plaintext
62 lines
2.5 KiB
Plaintext
#
|
|
# $Id: lpage.help,v 1.3 1999/05/11 19:26:03 kenmcd Exp $ #
|
|
#
|
|
|
|
@ mem.lpage.coalesce.scans large page coalescion attempts
|
|
Cumulative number of scans made by the coalescing daemon over the
|
|
entire set of pages. The daemon periodically scans all pages in memory
|
|
to see if they can be coalesced, and is also active when a requested
|
|
page size is unavailable.
|
|
|
|
@ mem.lpage.coalesce.success successful large page merges
|
|
Cumulative number of successful large page merges by the coalescing
|
|
daemon.
|
|
|
|
@ mem.lpage.faults count of large page faults
|
|
Cumulative number of large page faults that were successfully
|
|
satisfied.
|
|
|
|
@ mem.lpage.allocs count of vfault large page allocations
|
|
As part of a large page fault, vfault() may request allocation of a
|
|
large page. This metric denotes the number of such requests that were
|
|
satisfied.
|
|
|
|
@ mem.lpage.downgrade count of large page downgrades
|
|
Cumulative number of page allocation failures where the large page
|
|
requested cannot be immediately provided, so the even and odd address
|
|
of the larger page size is downgraded to base page size, and this
|
|
smaller page is allocated.
|
|
|
|
@ mem.lpage.page_splits count of large page splits
|
|
Cumulative number of splits of large pages to satisfy requests for
|
|
smaller sized pages.
|
|
|
|
@ mem.lpage.basesize minimum page size
|
|
The base (smallest) page size supported by the kernel, in Kilobytes.
|
|
|
|
@ mem.lpage.maxsize maximum page size
|
|
The maximum page size supported by the kernel, in Kilobytes.
|
|
|
|
@ mem.lpage.maxenabled maximum enabled page size
|
|
The maximum enabled page size, in Kilobytes.
|
|
|
|
@ mem.lpage.enabled enabled large page sizes
|
|
Set of large page sizes which are currently enabled, in Kilobytes.
|
|
Range of values is 16, 64, 256, 1024, 4096, and 16384.
|
|
|
|
#@ mem.lpage.reserved large pages reserved at boot time
|
|
#Since it is difficult to to coalesce very large pages at run time due
|
|
#to memory fragmentation, tunable parameters exist to allow large pages
|
|
#to be reserved at boot time. This metric provides the number of
|
|
#reserved pages for each page large page size.
|
|
#
|
|
#@ mem.lpage.coalesce.limits upper limit of free pages
|
|
#The kernel tries to keep a percentage of total free memory in the
|
|
#system at a certain page size. This metric is the percentage of total
|
|
#memory which has been specified as the current upper limit for free
|
|
#pages at each page size. These limits can be tuned dynamically.
|
|
#
|
|
#Since very large pages (>=1MB) are more difficult to coalesce at run
|
|
#time it is recommended that these tunables be set at boot time on busy
|
|
#systems. A higher setting will result in higher coalescing activity.
|