linux/fs/proc
Hugh Dickins 365e9c87a9 [PATCH] mm: update_hiwaters just in time
update_mem_hiwater has attracted various criticisms, in particular from those
concerned with mm scalability.  Originally it was called whenever rss or
total_vm got raised.  Then many of those callsites were replaced by a timer
tick call from account_system_time.  Now Frank van Maarseveen reports that to
be found inadequate.  How about this?  Works for Frank.

Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros
update_hiwater_rss and update_hiwater_vm.  Don't attempt to keep
mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually
by 1): those are hot paths.  Do the opposite, update only when about to lower
rss (usually by many), or just before final accounting in do_exit.  Handle
mm->hiwater_vm in the same way, though it's much less of an issue.  Demand
that whoever collects these hiwater statistics do the work of taking the
maximum with rss or total_vm.

And there has been no collector of these hiwater statistics in the tree.  The
new convention needs an example, so match Frank's usage by adding a VmPeak
line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS
(High-Water-Mark or High-Water-Memory).

There was a particular anomaly during mremap move, that hiwater_vm might be
captured too high.  A fleeting such anomaly remains, but it's quickly
corrected now, whereas before it would stick.

What locking?  None: if the app is racy then these statistics will be racy,
it's not worth any overhead to make them exact.  But whenever it suits,
hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under
page_table_lock (for now) or with preemption disabled (later on): without
going to any trouble, minimize the time between reading current values and
updating, to minimize those occasions when a racing thread bumps a count up
and back down in between.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-29 21:40:39 -07:00
..
array.c [PATCH] mm: rss = file_rss + anon_rss 2005-10-29 21:40:38 -07:00
base.c [PATCH] nommu build error fix 2005-10-14 17:10:13 -07:00
generic.c [PATCH] proc: link count fix 2005-09-07 16:57:28 -07:00
inode-alloc.txt Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
inode.c [PATCH] update filesystems for new delete_inode behavior 2005-09-09 13:57:27 -07:00
internal.h Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
kcore.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
kmsg.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
Makefile [PATCH] kdump: Access dump file in elf format (/proc/vmcore) 2005-06-25 16:24:53 -07:00
mmu.c [PATCH] fix impossible VmallocChunk 2005-05-17 07:59:10 -07:00
nommu.c [PATCH] output of /proc/maps on nommu systems is incomplete 2005-10-17 17:03:57 -07:00
proc_devtree.c [PATCH] ppc32/ppc64: cleanup /proc/device-tree 2005-06-01 07:54:14 -07:00
proc_misc.c [PATCH] kdump: Access dump file in elf format (/proc/vmcore) 2005-06-25 16:24:53 -07:00
proc_tty.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
root.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
task_mmu.c [PATCH] mm: update_hiwaters just in time 2005-10-29 21:40:39 -07:00
task_nommu.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
vmcore.c [PATCH] kdump: Parse elf32 headers and export through /proc/vmcore 2005-06-25 16:24:53 -07:00