2006-06-30 08:55:32 +00:00
|
|
|
/*
|
|
|
|
* linux/mm/vmstat.c
|
|
|
|
*
|
|
|
|
* Manages VM statistics
|
|
|
|
* Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds
|
2006-06-30 08:55:33 +00:00
|
|
|
*
|
|
|
|
* zoned VM statistics
|
|
|
|
* Copyright (C) 2006 Silicon Graphics, Inc.,
|
|
|
|
* Christoph Lameter <christoph@lameter.com>
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
* Copyright (C) 2008-2014 Christoph Lameter
|
2006-06-30 08:55:32 +00:00
|
|
|
*/
|
2008-10-06 00:13:52 +00:00
|
|
|
#include <linux/fs.h>
|
2006-06-30 08:55:32 +00:00
|
|
|
#include <linux/mm.h>
|
2007-07-29 22:36:13 +00:00
|
|
|
#include <linux/err.h>
|
2006-06-30 08:55:33 +00:00
|
|
|
#include <linux/module.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
|
|
|
#include <linux/slab.h>
|
2006-09-01 04:27:35 +00:00
|
|
|
#include <linux/cpu.h>
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
#include <linux/cpumask.h>
|
2008-07-24 04:27:03 +00:00
|
|
|
#include <linux/vmstat.h>
|
2015-02-10 22:09:43 +00:00
|
|
|
#include <linux/proc_fs.h>
|
|
|
|
#include <linux/seq_file.h>
|
|
|
|
#include <linux/debugfs.h>
|
Detach sched.h from mm.h
First thing mm.h does is including sched.h solely for can_do_mlock() inline
function which has "current" dereference inside. By dealing with can_do_mlock()
mm.h can be detached from sched.h which is good. See below, why.
This patch
a) removes unconditional inclusion of sched.h from mm.h
b) makes can_do_mlock() normal function in mm/mlock.c
c) exports can_do_mlock() to not break compilation
d) adds sched.h inclusions back to files that were getting it indirectly.
e) adds less bloated headers to some files (asm/signal.h, jiffies.h) that were
getting them indirectly
Net result is:
a) mm.h users would get less code to open, read, preprocess, parse, ... if
they don't need sched.h
b) sched.h stops being dependency for significant number of files:
on x86_64 allmodconfig touching sched.h results in recompile of 4083 files,
after patch it's only 3744 (-8.3%).
Cross-compile tested on
all arm defconfigs, all mips defconfigs, all powerpc defconfigs,
alpha alpha-up
arm
i386 i386-up i386-defconfig i386-allnoconfig
ia64 ia64-up
m68k
mips
parisc parisc-up
powerpc powerpc-up
s390 s390-up
sparc sparc-up
sparc64 sparc64-up
um-x86_64
x86_64 x86_64-up x86_64-defconfig x86_64-allnoconfig
as well as my two usual configs.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-05-20 21:22:52 +00:00
|
|
|
#include <linux/sched.h>
|
2010-05-24 21:32:26 +00:00
|
|
|
#include <linux/math64.h>
|
2010-10-26 21:21:36 +00:00
|
|
|
#include <linux/writeback.h>
|
2010-10-26 21:22:04 +00:00
|
|
|
#include <linux/compaction.h>
|
2013-09-11 21:22:36 +00:00
|
|
|
#include <linux/mm_inline.h>
|
mm/page_owner: keep track of page owners
This is the page owner tracking code which is introduced so far ago. It
is resident on Andrew's tree, though, nobody tried to upstream so it
remain as is. Our company uses this feature actively to debug memory leak
or to find a memory hogger so I decide to upstream this feature.
This functionality help us to know who allocates the page. When
allocating a page, we store some information about allocation in extra
memory. Later, if we need to know status of all pages, we can get and
analyze it from this stored information.
In previous version of this feature, extra memory is statically defined in
struct page, but, in this version, extra memory is allocated outside of
struct page. It enables us to turn on/off this feature at boottime
without considerable memory waste.
Although we already have tracepoint for tracing page allocation/free,
using it to analyze page owner is rather complex. We need to enlarge the
trace buffer for preventing overlapping until userspace program launched.
And, launched program continually dump out the trace buffer for later
analysis and it would change system behaviour with more possibility rather
than just keeping it in memory, so bad for debug.
Moreover, we can use page_owner feature further for various purposes. For
example, we can use it for fragmentation statistics implemented in this
patch. And, I also plan to implement some CMA failure debugging feature
using this interface.
I'd like to give the credit for all developers contributed this feature,
but, it's not easy because I don't know exact history. Sorry about that.
Below is people who has "Signed-off-by" in the patches in Andrew's tree.
Contributor:
Alexander Nyberg <alexn@dsv.su.se>
Mel Gorman <mgorman@suse.de>
Dave Hansen <dave@linux.vnet.ibm.com>
Minchan Kim <minchan@kernel.org>
Michal Nazarewicz <mina86@mina86.com>
Andrew Morton <akpm@linux-foundation.org>
Jungsoo Son <jungsoo.son@lge.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Dave Hansen <dave@sr71.net>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Jungsoo Son <jungsoo.son@lge.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:01 +00:00
|
|
|
#include <linux/page_ext.h>
|
|
|
|
#include <linux/page_owner.h>
|
2013-09-11 21:22:36 +00:00
|
|
|
|
|
|
|
#include "internal.h"
|
2006-06-30 08:55:32 +00:00
|
|
|
|
2006-06-30 08:55:45 +00:00
|
|
|
#ifdef CONFIG_VM_EVENT_COUNTERS
|
|
|
|
DEFINE_PER_CPU(struct vm_event_state, vm_event_states) = {{0}};
|
|
|
|
EXPORT_PER_CPU_SYMBOL(vm_event_states);
|
|
|
|
|
2010-08-10 00:18:59 +00:00
|
|
|
static void sum_vm_events(unsigned long *ret)
|
2006-06-30 08:55:45 +00:00
|
|
|
{
|
2008-02-05 06:29:22 +00:00
|
|
|
int cpu;
|
2006-06-30 08:55:45 +00:00
|
|
|
int i;
|
|
|
|
|
|
|
|
memset(ret, 0, NR_VM_EVENT_ITEMS * sizeof(unsigned long));
|
|
|
|
|
2010-08-10 00:18:59 +00:00
|
|
|
for_each_online_cpu(cpu) {
|
2006-06-30 08:55:45 +00:00
|
|
|
struct vm_event_state *this = &per_cpu(vm_event_states, cpu);
|
|
|
|
|
|
|
|
for (i = 0; i < NR_VM_EVENT_ITEMS; i++)
|
|
|
|
ret[i] += this->event[i];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Accumulate the vm event counters across all CPUs.
|
|
|
|
* The result is unavoidably approximate - it can change
|
|
|
|
* during and after execution of this function.
|
|
|
|
*/
|
|
|
|
void all_vm_events(unsigned long *ret)
|
|
|
|
{
|
2008-05-12 21:02:06 +00:00
|
|
|
get_online_cpus();
|
2010-08-10 00:18:59 +00:00
|
|
|
sum_vm_events(ret);
|
2008-05-12 21:02:06 +00:00
|
|
|
put_online_cpus();
|
2006-06-30 08:55:45 +00:00
|
|
|
}
|
2006-07-10 11:44:31 +00:00
|
|
|
EXPORT_SYMBOL_GPL(all_vm_events);
|
2006-06-30 08:55:45 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Fold the foreign cpu events into our own.
|
|
|
|
*
|
|
|
|
* This is adding to the events on one processor
|
|
|
|
* but keeps the global counts constant.
|
|
|
|
*/
|
|
|
|
void vm_events_fold_cpu(int cpu)
|
|
|
|
{
|
|
|
|
struct vm_event_state *fold_state = &per_cpu(vm_event_states, cpu);
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < NR_VM_EVENT_ITEMS; i++) {
|
|
|
|
count_vm_events(i, fold_state->event[i]);
|
|
|
|
fold_state->event[i] = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* CONFIG_VM_EVENT_COUNTERS */
|
|
|
|
|
2006-06-30 08:55:33 +00:00
|
|
|
/*
|
|
|
|
* Manage combined zone based / global counters
|
|
|
|
*
|
|
|
|
* vm_stat contains the global counters
|
|
|
|
*/
|
2011-11-01 00:09:46 +00:00
|
|
|
atomic_long_t vm_stat[NR_VM_ZONE_STAT_ITEMS] __cacheline_aligned_in_smp;
|
2006-06-30 08:55:33 +00:00
|
|
|
EXPORT_SYMBOL(vm_stat);
|
|
|
|
|
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
|
2011-01-13 23:45:43 +00:00
|
|
|
int calculate_pressure_threshold(struct zone *zone)
|
mm: page allocator: adjust the per-cpu counter threshold when memory is low
Commit aa45484 ("calculate a better estimate of NR_FREE_PAGES when memory
is low") noted that watermarks were based on the vmstat NR_FREE_PAGES. To
avoid synchronization overhead, these counters are maintained on a per-cpu
basis and drained both periodically and when a threshold is above a
threshold. On large CPU systems, the difference between the estimate and
real value of NR_FREE_PAGES can be very high. The system can get into a
case where pages are allocated far below the min watermark potentially
causing livelock issues. The commit solved the problem by taking a better
reading of NR_FREE_PAGES when memory was low.
Unfortately, as reported by Shaohua Li this accurate reading can consume a
large amount of CPU time on systems with many sockets due to cache line
bouncing. This patch takes a different approach. For large machines
where counter drift might be unsafe and while kswapd is awake, the per-cpu
thresholds for the target pgdat are reduced to limit the level of drift to
what should be a safe level. This incurs a performance penalty in heavy
memory pressure by a factor that depends on the workload and the machine
but the machine should function correctly without accidentally exhausting
all memory on a node. There is an additional cost when kswapd wakes and
sleeps but the event is not expected to be frequent - in Shaohua's test
case, there was one recorded sleep and wake event at least.
To ensure that kswapd wakes up, a safe version of zone_watermark_ok() is
introduced that takes a more accurate reading of NR_FREE_PAGES when called
from wakeup_kswapd, when deciding whether it is really safe to go back to
sleep in sleeping_prematurely() and when deciding if a zone is really
balanced or not in balance_pgdat(). We are still using an expensive
function but limiting how often it is called.
When the test case is reproduced, the time spent in the watermark
functions is reduced. The following report is on the percentage of time
spent cumulatively spent in the functions zone_nr_free_pages(),
zone_watermark_ok(), __zone_watermark_ok(), zone_watermark_ok_safe(),
zone_page_state_snapshot(), zone_page_state().
vanilla 11.6615%
disable-threshold 0.2584%
David said:
: We had to pull aa454840 "mm: page allocator: calculate a better estimate
: of NR_FREE_PAGES when memory is low and kswapd is awake" from 2.6.36
: internally because tests showed that it would cause the machine to stall
: as the result of heavy kswapd activity. I merged it back with this fix as
: it is pending in the -mm tree and it solves the issue we were seeing, so I
: definitely think this should be pushed to -stable (and I would seriously
: consider it for 2.6.37 inclusion even at this late date).
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reported-by: Shaohua Li <shaohua.li@intel.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Tested-by: Nicolas Bareil <nico@chdir.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: <stable@kernel.org> [2.6.37.1, 2.6.36.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:45:41 +00:00
|
|
|
{
|
|
|
|
int threshold;
|
|
|
|
int watermark_distance;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* As vmstats are not up to date, there is drift between the estimated
|
|
|
|
* and real values. For high thresholds and a high number of CPUs, it
|
|
|
|
* is possible for the min watermark to be breached while the estimated
|
|
|
|
* value looks fine. The pressure threshold is a reduced value such
|
|
|
|
* that even the maximum amount of drift will not accidentally breach
|
|
|
|
* the min watermark
|
|
|
|
*/
|
|
|
|
watermark_distance = low_wmark_pages(zone) - min_wmark_pages(zone);
|
|
|
|
threshold = max(1, (int)(watermark_distance / num_online_cpus()));
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Maximum threshold is 125
|
|
|
|
*/
|
|
|
|
threshold = min(125, threshold);
|
|
|
|
|
|
|
|
return threshold;
|
|
|
|
}
|
|
|
|
|
2011-01-13 23:45:43 +00:00
|
|
|
int calculate_normal_threshold(struct zone *zone)
|
2006-09-01 04:27:35 +00:00
|
|
|
{
|
|
|
|
int threshold;
|
|
|
|
int mem; /* memory in 128 MB units */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The threshold scales with the number of processors and the amount
|
|
|
|
* of memory per zone. More memory means that we can defer updates for
|
|
|
|
* longer, more processors could lead to more contention.
|
|
|
|
* fls() is used to have a cheap way of logarithmic scaling.
|
|
|
|
*
|
|
|
|
* Some sample thresholds:
|
|
|
|
*
|
|
|
|
* Threshold Processors (fls) Zonesize fls(mem+1)
|
|
|
|
* ------------------------------------------------------------------
|
|
|
|
* 8 1 1 0.9-1 GB 4
|
|
|
|
* 16 2 2 0.9-1 GB 4
|
|
|
|
* 20 2 2 1-2 GB 5
|
|
|
|
* 24 2 2 2-4 GB 6
|
|
|
|
* 28 2 2 4-8 GB 7
|
|
|
|
* 32 2 2 8-16 GB 8
|
|
|
|
* 4 2 2 <128M 1
|
|
|
|
* 30 4 3 2-4 GB 5
|
|
|
|
* 48 4 3 8-16 GB 8
|
|
|
|
* 32 8 4 1-2 GB 4
|
|
|
|
* 32 8 4 0.9-1GB 4
|
|
|
|
* 10 16 5 <128M 1
|
|
|
|
* 40 16 5 900M 4
|
|
|
|
* 70 64 7 2-4 GB 5
|
|
|
|
* 84 64 7 4-8 GB 6
|
|
|
|
* 108 512 9 4-8 GB 6
|
|
|
|
* 125 1024 10 8-16 GB 8
|
|
|
|
* 125 1024 10 16-32 GB 9
|
|
|
|
*/
|
|
|
|
|
2013-02-23 00:33:52 +00:00
|
|
|
mem = zone->managed_pages >> (27 - PAGE_SHIFT);
|
2006-09-01 04:27:35 +00:00
|
|
|
|
|
|
|
threshold = 2 * fls(num_online_cpus()) * (1 + fls(mem));
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Maximum threshold is 125
|
|
|
|
*/
|
|
|
|
threshold = min(125, threshold);
|
|
|
|
|
|
|
|
return threshold;
|
|
|
|
}
|
2006-06-30 08:55:33 +00:00
|
|
|
|
|
|
|
/*
|
2006-09-01 04:27:35 +00:00
|
|
|
* Refresh the thresholds for each zone.
|
2006-06-30 08:55:33 +00:00
|
|
|
*/
|
2011-05-25 00:11:33 +00:00
|
|
|
void refresh_zone_stat_thresholds(void)
|
2006-06-30 08:55:33 +00:00
|
|
|
{
|
2006-09-01 04:27:35 +00:00
|
|
|
struct zone *zone;
|
|
|
|
int cpu;
|
|
|
|
int threshold;
|
|
|
|
|
2009-03-31 22:19:31 +00:00
|
|
|
for_each_populated_zone(zone) {
|
2010-09-09 23:38:17 +00:00
|
|
|
unsigned long max_drift, tolerate_drift;
|
|
|
|
|
2011-01-13 23:45:43 +00:00
|
|
|
threshold = calculate_normal_threshold(zone);
|
2006-09-01 04:27:35 +00:00
|
|
|
|
|
|
|
for_each_online_cpu(cpu)
|
2010-01-05 06:34:51 +00:00
|
|
|
per_cpu_ptr(zone->pageset, cpu)->stat_threshold
|
|
|
|
= threshold;
|
2010-09-09 23:38:17 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Only set percpu_drift_mark if there is a danger that
|
|
|
|
* NR_FREE_PAGES reports the low watermark is ok when in fact
|
|
|
|
* the min watermark could be breached by an allocation
|
|
|
|
*/
|
|
|
|
tolerate_drift = low_wmark_pages(zone) - min_wmark_pages(zone);
|
|
|
|
max_drift = num_online_cpus() * threshold;
|
|
|
|
if (max_drift > tolerate_drift)
|
|
|
|
zone->percpu_drift_mark = high_wmark_pages(zone) +
|
|
|
|
max_drift;
|
2006-09-01 04:27:35 +00:00
|
|
|
}
|
2006-06-30 08:55:33 +00:00
|
|
|
}
|
|
|
|
|
2011-01-13 23:45:43 +00:00
|
|
|
void set_pgdat_percpu_threshold(pg_data_t *pgdat,
|
|
|
|
int (*calculate_pressure)(struct zone *))
|
mm: page allocator: adjust the per-cpu counter threshold when memory is low
Commit aa45484 ("calculate a better estimate of NR_FREE_PAGES when memory
is low") noted that watermarks were based on the vmstat NR_FREE_PAGES. To
avoid synchronization overhead, these counters are maintained on a per-cpu
basis and drained both periodically and when a threshold is above a
threshold. On large CPU systems, the difference between the estimate and
real value of NR_FREE_PAGES can be very high. The system can get into a
case where pages are allocated far below the min watermark potentially
causing livelock issues. The commit solved the problem by taking a better
reading of NR_FREE_PAGES when memory was low.
Unfortately, as reported by Shaohua Li this accurate reading can consume a
large amount of CPU time on systems with many sockets due to cache line
bouncing. This patch takes a different approach. For large machines
where counter drift might be unsafe and while kswapd is awake, the per-cpu
thresholds for the target pgdat are reduced to limit the level of drift to
what should be a safe level. This incurs a performance penalty in heavy
memory pressure by a factor that depends on the workload and the machine
but the machine should function correctly without accidentally exhausting
all memory on a node. There is an additional cost when kswapd wakes and
sleeps but the event is not expected to be frequent - in Shaohua's test
case, there was one recorded sleep and wake event at least.
To ensure that kswapd wakes up, a safe version of zone_watermark_ok() is
introduced that takes a more accurate reading of NR_FREE_PAGES when called
from wakeup_kswapd, when deciding whether it is really safe to go back to
sleep in sleeping_prematurely() and when deciding if a zone is really
balanced or not in balance_pgdat(). We are still using an expensive
function but limiting how often it is called.
When the test case is reproduced, the time spent in the watermark
functions is reduced. The following report is on the percentage of time
spent cumulatively spent in the functions zone_nr_free_pages(),
zone_watermark_ok(), __zone_watermark_ok(), zone_watermark_ok_safe(),
zone_page_state_snapshot(), zone_page_state().
vanilla 11.6615%
disable-threshold 0.2584%
David said:
: We had to pull aa454840 "mm: page allocator: calculate a better estimate
: of NR_FREE_PAGES when memory is low and kswapd is awake" from 2.6.36
: internally because tests showed that it would cause the machine to stall
: as the result of heavy kswapd activity. I merged it back with this fix as
: it is pending in the -mm tree and it solves the issue we were seeing, so I
: definitely think this should be pushed to -stable (and I would seriously
: consider it for 2.6.37 inclusion even at this late date).
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reported-by: Shaohua Li <shaohua.li@intel.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Tested-by: Nicolas Bareil <nico@chdir.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: <stable@kernel.org> [2.6.37.1, 2.6.36.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:45:41 +00:00
|
|
|
{
|
|
|
|
struct zone *zone;
|
|
|
|
int cpu;
|
|
|
|
int threshold;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < pgdat->nr_zones; i++) {
|
|
|
|
zone = &pgdat->node_zones[i];
|
|
|
|
if (!zone->percpu_drift_mark)
|
|
|
|
continue;
|
|
|
|
|
2011-01-13 23:45:43 +00:00
|
|
|
threshold = (*calculate_pressure)(zone);
|
2014-08-06 23:07:18 +00:00
|
|
|
for_each_online_cpu(cpu)
|
mm: page allocator: adjust the per-cpu counter threshold when memory is low
Commit aa45484 ("calculate a better estimate of NR_FREE_PAGES when memory
is low") noted that watermarks were based on the vmstat NR_FREE_PAGES. To
avoid synchronization overhead, these counters are maintained on a per-cpu
basis and drained both periodically and when a threshold is above a
threshold. On large CPU systems, the difference between the estimate and
real value of NR_FREE_PAGES can be very high. The system can get into a
case where pages are allocated far below the min watermark potentially
causing livelock issues. The commit solved the problem by taking a better
reading of NR_FREE_PAGES when memory was low.
Unfortately, as reported by Shaohua Li this accurate reading can consume a
large amount of CPU time on systems with many sockets due to cache line
bouncing. This patch takes a different approach. For large machines
where counter drift might be unsafe and while kswapd is awake, the per-cpu
thresholds for the target pgdat are reduced to limit the level of drift to
what should be a safe level. This incurs a performance penalty in heavy
memory pressure by a factor that depends on the workload and the machine
but the machine should function correctly without accidentally exhausting
all memory on a node. There is an additional cost when kswapd wakes and
sleeps but the event is not expected to be frequent - in Shaohua's test
case, there was one recorded sleep and wake event at least.
To ensure that kswapd wakes up, a safe version of zone_watermark_ok() is
introduced that takes a more accurate reading of NR_FREE_PAGES when called
from wakeup_kswapd, when deciding whether it is really safe to go back to
sleep in sleeping_prematurely() and when deciding if a zone is really
balanced or not in balance_pgdat(). We are still using an expensive
function but limiting how often it is called.
When the test case is reproduced, the time spent in the watermark
functions is reduced. The following report is on the percentage of time
spent cumulatively spent in the functions zone_nr_free_pages(),
zone_watermark_ok(), __zone_watermark_ok(), zone_watermark_ok_safe(),
zone_page_state_snapshot(), zone_page_state().
vanilla 11.6615%
disable-threshold 0.2584%
David said:
: We had to pull aa454840 "mm: page allocator: calculate a better estimate
: of NR_FREE_PAGES when memory is low and kswapd is awake" from 2.6.36
: internally because tests showed that it would cause the machine to stall
: as the result of heavy kswapd activity. I merged it back with this fix as
: it is pending in the -mm tree and it solves the issue we were seeing, so I
: definitely think this should be pushed to -stable (and I would seriously
: consider it for 2.6.37 inclusion even at this late date).
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reported-by: Shaohua Li <shaohua.li@intel.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Tested-by: Nicolas Bareil <nico@chdir.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: <stable@kernel.org> [2.6.37.1, 2.6.36.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:45:41 +00:00
|
|
|
per_cpu_ptr(zone->pageset, cpu)->stat_threshold
|
|
|
|
= threshold;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-06-30 08:55:33 +00:00
|
|
|
/*
|
2014-06-04 23:09:51 +00:00
|
|
|
* For use when we know that interrupts are disabled,
|
|
|
|
* or when we know that preemption is disabled and that
|
|
|
|
* particular counter cannot be updated from interrupt context.
|
2006-06-30 08:55:33 +00:00
|
|
|
*/
|
|
|
|
void __mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
|
2015-12-29 22:54:32 +00:00
|
|
|
long delta)
|
2006-06-30 08:55:33 +00:00
|
|
|
{
|
2010-12-06 17:16:20 +00:00
|
|
|
struct per_cpu_pageset __percpu *pcp = zone->pageset;
|
|
|
|
s8 __percpu *p = pcp->vm_stat_diff + item;
|
2006-06-30 08:55:33 +00:00
|
|
|
long x;
|
2010-12-06 17:16:20 +00:00
|
|
|
long t;
|
|
|
|
|
|
|
|
x = delta + __this_cpu_read(*p);
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2010-12-06 17:16:20 +00:00
|
|
|
t = __this_cpu_read(pcp->stat_threshold);
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2010-12-06 17:16:20 +00:00
|
|
|
if (unlikely(x > t || x < -t)) {
|
2006-06-30 08:55:33 +00:00
|
|
|
zone_page_state_add(x, zone, item);
|
|
|
|
x = 0;
|
|
|
|
}
|
2010-12-06 17:16:20 +00:00
|
|
|
__this_cpu_write(*p, x);
|
2006-06-30 08:55:33 +00:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(__mod_zone_page_state);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Optimized increment and decrement functions.
|
|
|
|
*
|
|
|
|
* These are only for a single page and therefore can take a struct page *
|
|
|
|
* argument instead of struct zone *. This allows the inclusion of the code
|
|
|
|
* generated for page_zone(page) into the optimized functions.
|
|
|
|
*
|
|
|
|
* No overflow check is necessary and therefore the differential can be
|
|
|
|
* incremented or decremented in place which may allow the compilers to
|
|
|
|
* generate better code.
|
|
|
|
* The increment or decrement is known and therefore one boundary check can
|
|
|
|
* be omitted.
|
|
|
|
*
|
2006-09-01 04:27:35 +00:00
|
|
|
* NOTE: These functions are very performance sensitive. Change only
|
|
|
|
* with care.
|
|
|
|
*
|
2006-06-30 08:55:33 +00:00
|
|
|
* Some processors have inc/dec instructions that are atomic vs an interrupt.
|
|
|
|
* However, the code must first determine the differential location in a zone
|
|
|
|
* based on the processor number and then inc/dec the counter. There is no
|
|
|
|
* guarantee without disabling preemption that the processor will not change
|
|
|
|
* in between and therefore the atomicity vs. interrupt cannot be exploited
|
|
|
|
* in a useful way here.
|
|
|
|
*/
|
2007-02-10 09:43:01 +00:00
|
|
|
void __inc_zone_state(struct zone *zone, enum zone_stat_item item)
|
2006-06-30 08:55:33 +00:00
|
|
|
{
|
2010-12-06 17:16:20 +00:00
|
|
|
struct per_cpu_pageset __percpu *pcp = zone->pageset;
|
|
|
|
s8 __percpu *p = pcp->vm_stat_diff + item;
|
|
|
|
s8 v, t;
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2010-12-06 17:40:02 +00:00
|
|
|
v = __this_cpu_inc_return(*p);
|
2010-12-06 17:16:20 +00:00
|
|
|
t = __this_cpu_read(pcp->stat_threshold);
|
|
|
|
if (unlikely(v > t)) {
|
|
|
|
s8 overstep = t >> 1;
|
2006-09-01 04:27:35 +00:00
|
|
|
|
2010-12-06 17:16:20 +00:00
|
|
|
zone_page_state_add(v + overstep, zone, item);
|
|
|
|
__this_cpu_write(*p, -overstep);
|
2006-06-30 08:55:33 +00:00
|
|
|
}
|
|
|
|
}
|
2006-06-30 08:55:44 +00:00
|
|
|
|
|
|
|
void __inc_zone_page_state(struct page *page, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
__inc_zone_state(page_zone(page), item);
|
|
|
|
}
|
2006-06-30 08:55:33 +00:00
|
|
|
EXPORT_SYMBOL(__inc_zone_page_state);
|
|
|
|
|
2007-02-10 09:43:01 +00:00
|
|
|
void __dec_zone_state(struct zone *zone, enum zone_stat_item item)
|
2006-06-30 08:55:33 +00:00
|
|
|
{
|
2010-12-06 17:16:20 +00:00
|
|
|
struct per_cpu_pageset __percpu *pcp = zone->pageset;
|
|
|
|
s8 __percpu *p = pcp->vm_stat_diff + item;
|
|
|
|
s8 v, t;
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2010-12-06 17:40:02 +00:00
|
|
|
v = __this_cpu_dec_return(*p);
|
2010-12-06 17:16:20 +00:00
|
|
|
t = __this_cpu_read(pcp->stat_threshold);
|
|
|
|
if (unlikely(v < - t)) {
|
|
|
|
s8 overstep = t >> 1;
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2010-12-06 17:16:20 +00:00
|
|
|
zone_page_state_add(v - overstep, zone, item);
|
|
|
|
__this_cpu_write(*p, overstep);
|
2006-06-30 08:55:33 +00:00
|
|
|
}
|
|
|
|
}
|
2007-02-10 09:43:01 +00:00
|
|
|
|
|
|
|
void __dec_zone_page_state(struct page *page, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
__dec_zone_state(page_zone(page), item);
|
|
|
|
}
|
2006-06-30 08:55:33 +00:00
|
|
|
EXPORT_SYMBOL(__dec_zone_page_state);
|
|
|
|
|
2012-01-13 01:17:30 +00:00
|
|
|
#ifdef CONFIG_HAVE_CMPXCHG_LOCAL
|
2010-12-14 16:28:46 +00:00
|
|
|
/*
|
|
|
|
* If we have cmpxchg_local support then we do not need to incur the overhead
|
|
|
|
* that comes with local_irq_save/restore if we use this_cpu_cmpxchg.
|
|
|
|
*
|
|
|
|
* mod_state() modifies the zone counter state through atomic per cpu
|
|
|
|
* operations.
|
|
|
|
*
|
|
|
|
* Overstep mode specifies how overstep should handled:
|
|
|
|
* 0 No overstepping
|
|
|
|
* 1 Overstepping half of threshold
|
|
|
|
* -1 Overstepping minus half of threshold
|
|
|
|
*/
|
2015-12-29 22:54:32 +00:00
|
|
|
static inline void mod_state(struct zone *zone, enum zone_stat_item item,
|
|
|
|
long delta, int overstep_mode)
|
2010-12-14 16:28:46 +00:00
|
|
|
{
|
|
|
|
struct per_cpu_pageset __percpu *pcp = zone->pageset;
|
|
|
|
s8 __percpu *p = pcp->vm_stat_diff + item;
|
|
|
|
long o, n, t, z;
|
|
|
|
|
|
|
|
do {
|
|
|
|
z = 0; /* overflow to zone counters */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The fetching of the stat_threshold is racy. We may apply
|
|
|
|
* a counter threshold to the wrong the cpu if we get
|
2011-04-14 22:21:58 +00:00
|
|
|
* rescheduled while executing here. However, the next
|
|
|
|
* counter update will apply the threshold again and
|
|
|
|
* therefore bring the counter under the threshold again.
|
|
|
|
*
|
|
|
|
* Most of the time the thresholds are the same anyways
|
|
|
|
* for all cpus in a zone.
|
2010-12-14 16:28:46 +00:00
|
|
|
*/
|
|
|
|
t = this_cpu_read(pcp->stat_threshold);
|
|
|
|
|
|
|
|
o = this_cpu_read(*p);
|
|
|
|
n = delta + o;
|
|
|
|
|
|
|
|
if (n > t || n < -t) {
|
|
|
|
int os = overstep_mode * (t >> 1) ;
|
|
|
|
|
|
|
|
/* Overflow must be added to zone counters */
|
|
|
|
z = n + os;
|
|
|
|
n = -os;
|
|
|
|
}
|
|
|
|
} while (this_cpu_cmpxchg(*p, o, n) != o);
|
|
|
|
|
|
|
|
if (z)
|
|
|
|
zone_page_state_add(z, zone, item);
|
|
|
|
}
|
|
|
|
|
|
|
|
void mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
|
2015-12-29 22:54:32 +00:00
|
|
|
long delta)
|
2010-12-14 16:28:46 +00:00
|
|
|
{
|
|
|
|
mod_state(zone, item, delta, 0);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(mod_zone_page_state);
|
|
|
|
|
|
|
|
void inc_zone_state(struct zone *zone, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
mod_state(zone, item, 1, 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
void inc_zone_page_state(struct page *page, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
mod_state(page_zone(page), item, 1, 1);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(inc_zone_page_state);
|
|
|
|
|
|
|
|
void dec_zone_page_state(struct page *page, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
mod_state(page_zone(page), item, -1, -1);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(dec_zone_page_state);
|
|
|
|
#else
|
|
|
|
/*
|
|
|
|
* Use interrupt disable to serialize counter updates
|
|
|
|
*/
|
|
|
|
void mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
|
2015-12-29 22:54:32 +00:00
|
|
|
long delta)
|
2010-12-14 16:28:46 +00:00
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
local_irq_save(flags);
|
|
|
|
__mod_zone_page_state(zone, item, delta);
|
|
|
|
local_irq_restore(flags);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(mod_zone_page_state);
|
|
|
|
|
2006-06-30 08:55:44 +00:00
|
|
|
void inc_zone_state(struct zone *zone, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
local_irq_save(flags);
|
|
|
|
__inc_zone_state(zone, item);
|
|
|
|
local_irq_restore(flags);
|
|
|
|
}
|
|
|
|
|
2006-06-30 08:55:33 +00:00
|
|
|
void inc_zone_page_state(struct page *page, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
struct zone *zone;
|
|
|
|
|
|
|
|
zone = page_zone(page);
|
|
|
|
local_irq_save(flags);
|
2006-06-30 08:55:44 +00:00
|
|
|
__inc_zone_state(zone, item);
|
2006-06-30 08:55:33 +00:00
|
|
|
local_irq_restore(flags);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(inc_zone_page_state);
|
|
|
|
|
|
|
|
void dec_zone_page_state(struct page *page, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
local_irq_save(flags);
|
2006-09-01 04:27:34 +00:00
|
|
|
__dec_zone_page_state(page, item);
|
2006-06-30 08:55:33 +00:00
|
|
|
local_irq_restore(flags);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(dec_zone_page_state);
|
2010-12-14 16:28:46 +00:00
|
|
|
#endif
|
2006-06-30 08:55:33 +00:00
|
|
|
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Fold a differential into the global counters.
|
|
|
|
* Returns the number of counters updated.
|
|
|
|
*/
|
|
|
|
static int fold_diff(int *diff)
|
2013-09-11 21:21:31 +00:00
|
|
|
{
|
|
|
|
int i;
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
int changes = 0;
|
2013-09-11 21:21:31 +00:00
|
|
|
|
|
|
|
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
if (diff[i]) {
|
2013-09-11 21:21:31 +00:00
|
|
|
atomic_long_add(diff[i], &vm_stat[i]);
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
changes++;
|
|
|
|
}
|
|
|
|
return changes;
|
2013-09-11 21:21:31 +00:00
|
|
|
}
|
|
|
|
|
2006-06-30 08:55:33 +00:00
|
|
|
/*
|
2013-09-11 21:21:30 +00:00
|
|
|
* Update the zone counters for the current cpu.
|
2008-02-05 06:29:16 +00:00
|
|
|
*
|
2007-05-09 09:35:14 +00:00
|
|
|
* Note that refresh_cpu_vm_stats strives to only access
|
|
|
|
* node local memory. The per cpu pagesets on remote zones are placed
|
|
|
|
* in the memory local to the processor using that pageset. So the
|
|
|
|
* loop over all zones will access a series of cachelines local to
|
|
|
|
* the processor.
|
|
|
|
*
|
|
|
|
* The call to zone_page_state_add updates the cachelines with the
|
|
|
|
* statistics in the remote zone struct as well as the global cachelines
|
|
|
|
* with the global counters. These could cause remote node cache line
|
|
|
|
* bouncing and will have to be only done when necessary.
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
*
|
|
|
|
* The function returns the number of global counters updated.
|
2006-06-30 08:55:33 +00:00
|
|
|
*/
|
2016-01-14 23:21:40 +00:00
|
|
|
static int refresh_cpu_vm_stats(bool do_pagesets)
|
2006-06-30 08:55:33 +00:00
|
|
|
{
|
|
|
|
struct zone *zone;
|
|
|
|
int i;
|
2008-02-05 06:29:16 +00:00
|
|
|
int global_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
int changes = 0;
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2009-03-31 22:19:31 +00:00
|
|
|
for_each_populated_zone(zone) {
|
2013-09-11 21:21:32 +00:00
|
|
|
struct per_cpu_pageset __percpu *p = zone->pageset;
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2013-09-11 21:21:32 +00:00
|
|
|
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++) {
|
|
|
|
int v;
|
2006-06-30 08:55:33 +00:00
|
|
|
|
2013-09-11 21:21:32 +00:00
|
|
|
v = this_cpu_xchg(p->vm_stat_diff[i], 0);
|
|
|
|
if (v) {
|
2008-02-05 06:29:16 +00:00
|
|
|
|
|
|
|
atomic_long_add(v, &zone->vm_stat[i]);
|
|
|
|
global_diff[i] += v;
|
2007-05-09 09:35:14 +00:00
|
|
|
#ifdef CONFIG_NUMA
|
|
|
|
/* 3 seconds idle till flush */
|
2013-09-11 21:21:32 +00:00
|
|
|
__this_cpu_write(p->expire, 3);
|
2007-05-09 09:35:14 +00:00
|
|
|
#endif
|
2006-06-30 08:55:33 +00:00
|
|
|
}
|
2013-09-11 21:21:32 +00:00
|
|
|
}
|
2007-05-09 09:35:14 +00:00
|
|
|
#ifdef CONFIG_NUMA
|
2016-01-14 23:21:40 +00:00
|
|
|
if (do_pagesets) {
|
|
|
|
cond_resched();
|
|
|
|
/*
|
|
|
|
* Deal with draining the remote pageset of this
|
|
|
|
* processor
|
|
|
|
*
|
|
|
|
* Check if there are pages remaining in this pageset
|
|
|
|
* if not then there is nothing to expire.
|
|
|
|
*/
|
|
|
|
if (!__this_cpu_read(p->expire) ||
|
2013-09-11 21:21:32 +00:00
|
|
|
!__this_cpu_read(p->pcp.count))
|
2016-01-14 23:21:40 +00:00
|
|
|
continue;
|
2007-05-09 09:35:14 +00:00
|
|
|
|
2016-01-14 23:21:40 +00:00
|
|
|
/*
|
|
|
|
* We never drain zones local to this processor.
|
|
|
|
*/
|
|
|
|
if (zone_to_nid(zone) == numa_node_id()) {
|
|
|
|
__this_cpu_write(p->expire, 0);
|
|
|
|
continue;
|
|
|
|
}
|
2007-05-09 09:35:14 +00:00
|
|
|
|
2016-01-14 23:21:40 +00:00
|
|
|
if (__this_cpu_dec_return(p->expire))
|
|
|
|
continue;
|
2007-05-09 09:35:14 +00:00
|
|
|
|
2016-01-14 23:21:40 +00:00
|
|
|
if (__this_cpu_read(p->pcp.count)) {
|
|
|
|
drain_zone_pages(zone, this_cpu_ptr(&p->pcp));
|
|
|
|
changes++;
|
|
|
|
}
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
}
|
2007-05-09 09:35:14 +00:00
|
|
|
#endif
|
2006-06-30 08:55:33 +00:00
|
|
|
}
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
changes += fold_diff(global_diff);
|
|
|
|
return changes;
|
2006-06-30 08:55:33 +00:00
|
|
|
}
|
|
|
|
|
2013-09-11 21:21:30 +00:00
|
|
|
/*
|
|
|
|
* Fold the data for an offline cpu into the global array.
|
|
|
|
* There cannot be any access by the offline cpu and therefore
|
|
|
|
* synchronization is simplified.
|
|
|
|
*/
|
|
|
|
void cpu_vm_stats_fold(int cpu)
|
|
|
|
{
|
|
|
|
struct zone *zone;
|
|
|
|
int i;
|
|
|
|
int global_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
|
|
|
|
|
|
|
|
for_each_populated_zone(zone) {
|
|
|
|
struct per_cpu_pageset *p;
|
|
|
|
|
|
|
|
p = per_cpu_ptr(zone->pageset, cpu);
|
|
|
|
|
|
|
|
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
|
|
|
|
if (p->vm_stat_diff[i]) {
|
|
|
|
int v;
|
|
|
|
|
|
|
|
v = p->vm_stat_diff[i];
|
|
|
|
p->vm_stat_diff[i] = 0;
|
|
|
|
atomic_long_add(v, &zone->vm_stat[i]);
|
|
|
|
global_diff[i] += v;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-09-11 21:21:31 +00:00
|
|
|
fold_diff(global_diff);
|
2013-09-11 21:21:30 +00:00
|
|
|
}
|
|
|
|
|
2013-04-29 22:08:38 +00:00
|
|
|
/*
|
|
|
|
* this is only called if !populated_zone(zone), which implies no other users of
|
|
|
|
* pset->vm_stat_diff[] exsist.
|
|
|
|
*/
|
2012-10-08 23:33:39 +00:00
|
|
|
void drain_zonestat(struct zone *zone, struct per_cpu_pageset *pset)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
|
|
|
|
if (pset->vm_stat_diff[i]) {
|
|
|
|
int v = pset->vm_stat_diff[i];
|
|
|
|
pset->vm_stat_diff[i] = 0;
|
|
|
|
atomic_long_add(v, &zone->vm_stat[i]);
|
|
|
|
atomic_long_add(v, &vm_stat[i]);
|
|
|
|
}
|
|
|
|
}
|
2006-06-30 08:55:33 +00:00
|
|
|
#endif
|
|
|
|
|
2006-06-30 08:55:44 +00:00
|
|
|
#ifdef CONFIG_NUMA
|
|
|
|
/*
|
|
|
|
* zonelist = the list of zones passed to the allocator
|
|
|
|
* z = the zone from which the allocation occurred.
|
|
|
|
*
|
|
|
|
* Must be called with interrupts disabled.
|
2011-03-22 23:33:12 +00:00
|
|
|
*
|
|
|
|
* When __GFP_OTHER_NODE is set assume the node of the preferred
|
|
|
|
* zone is the local node. This is useful for daemons who allocate
|
|
|
|
* memory on behalf of other processes.
|
2006-06-30 08:55:44 +00:00
|
|
|
*/
|
2011-03-22 23:33:12 +00:00
|
|
|
void zone_statistics(struct zone *preferred_zone, struct zone *z, gfp_t flags)
|
2006-06-30 08:55:44 +00:00
|
|
|
{
|
2008-04-28 09:12:14 +00:00
|
|
|
if (z->zone_pgdat == preferred_zone->zone_pgdat) {
|
2006-06-30 08:55:44 +00:00
|
|
|
__inc_zone_state(z, NUMA_HIT);
|
|
|
|
} else {
|
|
|
|
__inc_zone_state(z, NUMA_MISS);
|
2008-04-28 09:12:14 +00:00
|
|
|
__inc_zone_state(preferred_zone, NUMA_FOREIGN);
|
2006-06-30 08:55:44 +00:00
|
|
|
}
|
2011-03-22 23:33:12 +00:00
|
|
|
if (z->node == ((flags & __GFP_OTHER_NODE) ?
|
|
|
|
preferred_zone->node : numa_node_id()))
|
2006-06-30 08:55:44 +00:00
|
|
|
__inc_zone_state(z, NUMA_LOCAL);
|
|
|
|
else
|
|
|
|
__inc_zone_state(z, NUMA_OTHER);
|
|
|
|
}
|
2015-11-06 02:48:43 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Determine the per node value of a stat item.
|
|
|
|
*/
|
|
|
|
unsigned long node_page_state(int node, enum zone_stat_item item)
|
|
|
|
{
|
|
|
|
struct zone *zones = NODE_DATA(node)->node_zones;
|
|
|
|
|
|
|
|
return
|
|
|
|
#ifdef CONFIG_ZONE_DMA
|
|
|
|
zone_page_state(&zones[ZONE_DMA], item) +
|
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_ZONE_DMA32
|
|
|
|
zone_page_state(&zones[ZONE_DMA32], item) +
|
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_HIGHMEM
|
|
|
|
zone_page_state(&zones[ZONE_HIGHMEM], item) +
|
|
|
|
#endif
|
|
|
|
zone_page_state(&zones[ZONE_NORMAL], item) +
|
|
|
|
zone_page_state(&zones[ZONE_MOVABLE], item);
|
|
|
|
}
|
|
|
|
|
2006-06-30 08:55:44 +00:00
|
|
|
#endif
|
|
|
|
|
2010-05-24 21:32:25 +00:00
|
|
|
#ifdef CONFIG_COMPACTION
|
2010-10-26 21:22:04 +00:00
|
|
|
|
2010-05-24 21:32:25 +00:00
|
|
|
struct contig_page_info {
|
|
|
|
unsigned long free_pages;
|
|
|
|
unsigned long free_blocks_total;
|
|
|
|
unsigned long free_blocks_suitable;
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Calculate the number of free pages in a zone, how many contiguous
|
|
|
|
* pages are free and how many are large enough to satisfy an allocation of
|
|
|
|
* the target size. Note that this function makes no attempt to estimate
|
|
|
|
* how many suitable free blocks there *might* be if MOVABLE pages were
|
|
|
|
* migrated. Calculating that is possible, but expensive and can be
|
|
|
|
* figured out from userspace
|
|
|
|
*/
|
|
|
|
static void fill_contig_page_info(struct zone *zone,
|
|
|
|
unsigned int suitable_order,
|
|
|
|
struct contig_page_info *info)
|
|
|
|
{
|
|
|
|
unsigned int order;
|
|
|
|
|
|
|
|
info->free_pages = 0;
|
|
|
|
info->free_blocks_total = 0;
|
|
|
|
info->free_blocks_suitable = 0;
|
|
|
|
|
|
|
|
for (order = 0; order < MAX_ORDER; order++) {
|
|
|
|
unsigned long blocks;
|
|
|
|
|
|
|
|
/* Count number of free blocks */
|
|
|
|
blocks = zone->free_area[order].nr_free;
|
|
|
|
info->free_blocks_total += blocks;
|
|
|
|
|
|
|
|
/* Count free base pages */
|
|
|
|
info->free_pages += blocks << order;
|
|
|
|
|
|
|
|
/* Count the suitable free blocks */
|
|
|
|
if (order >= suitable_order)
|
|
|
|
info->free_blocks_suitable += blocks <<
|
|
|
|
(order - suitable_order);
|
|
|
|
}
|
|
|
|
}
|
2010-05-24 21:32:26 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* A fragmentation index only makes sense if an allocation of a requested
|
|
|
|
* size would fail. If that is true, the fragmentation index indicates
|
|
|
|
* whether external fragmentation or a lack of memory was the problem.
|
|
|
|
* The value can be used to determine if page reclaim or compaction
|
|
|
|
* should be used
|
|
|
|
*/
|
2010-05-24 21:32:30 +00:00
|
|
|
static int __fragmentation_index(unsigned int order, struct contig_page_info *info)
|
2010-05-24 21:32:26 +00:00
|
|
|
{
|
|
|
|
unsigned long requested = 1UL << order;
|
|
|
|
|
|
|
|
if (!info->free_blocks_total)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* Fragmentation index only makes sense when a request would fail */
|
|
|
|
if (info->free_blocks_suitable)
|
|
|
|
return -1000;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Index is between 0 and 1 so return within 3 decimal places
|
|
|
|
*
|
|
|
|
* 0 => allocation would fail due to lack of memory
|
|
|
|
* 1 => allocation would fail due to fragmentation
|
|
|
|
*/
|
|
|
|
return 1000 - div_u64( (1000+(div_u64(info->free_pages * 1000ULL, requested))), info->free_blocks_total);
|
|
|
|
}
|
2010-05-24 21:32:30 +00:00
|
|
|
|
|
|
|
/* Same as __fragmentation index but allocs contig_page_info on stack */
|
|
|
|
int fragmentation_index(struct zone *zone, unsigned int order)
|
|
|
|
{
|
|
|
|
struct contig_page_info info;
|
|
|
|
|
|
|
|
fill_contig_page_info(zone, order, &info);
|
|
|
|
return __fragmentation_index(order, &info);
|
|
|
|
}
|
2010-05-24 21:32:25 +00:00
|
|
|
#endif
|
|
|
|
|
2011-09-14 23:21:05 +00:00
|
|
|
#if defined(CONFIG_PROC_FS) || defined(CONFIG_SYSFS) || defined(CONFIG_NUMA)
|
2011-05-25 00:11:28 +00:00
|
|
|
#ifdef CONFIG_ZONE_DMA
|
|
|
|
#define TEXT_FOR_DMA(xx) xx "_dma",
|
|
|
|
#else
|
|
|
|
#define TEXT_FOR_DMA(xx)
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifdef CONFIG_ZONE_DMA32
|
|
|
|
#define TEXT_FOR_DMA32(xx) xx "_dma32",
|
|
|
|
#else
|
|
|
|
#define TEXT_FOR_DMA32(xx)
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifdef CONFIG_HIGHMEM
|
|
|
|
#define TEXT_FOR_HIGHMEM(xx) xx "_high",
|
|
|
|
#else
|
|
|
|
#define TEXT_FOR_HIGHMEM(xx)
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#define TEXTS_FOR_ZONES(xx) TEXT_FOR_DMA(xx) TEXT_FOR_DMA32(xx) xx "_normal", \
|
|
|
|
TEXT_FOR_HIGHMEM(xx) xx "_movable",
|
|
|
|
|
|
|
|
const char * const vmstat_text[] = {
|
2014-10-09 22:29:32 +00:00
|
|
|
/* enum zone_stat_item countes */
|
2011-05-25 00:11:28 +00:00
|
|
|
"nr_free_pages",
|
mm: page_alloc: fair zone allocator policy
Each zone that holds userspace pages of one workload must be aged at a
speed proportional to the zone size. Otherwise, the time an individual
page gets to stay in memory depends on the zone it happened to be
allocated in. Asymmetry in the zone aging creates rather unpredictable
aging behavior and results in the wrong pages being reclaimed, activated
etc.
But exactly this happens right now because of the way the page allocator
and kswapd interact. The page allocator uses per-node lists of all zones
in the system, ordered by preference, when allocating a new page. When
the first iteration does not yield any results, kswapd is woken up and the
allocator retries. Due to the way kswapd reclaims zones below the high
watermark while a zone can be allocated from when it is above the low
watermark, the allocator may keep kswapd running while kswapd reclaim
ensures that the page allocator can keep allocating from the first zone in
the zonelist for extended periods of time. Meanwhile the other zones
rarely see new allocations and thus get aged much slower in comparison.
The result is that the occasional page placed in lower zones gets
relatively more time in memory, even gets promoted to the active list
after its peers have long been evicted. Meanwhile, the bulk of the
working set may be thrashing on the preferred zone even though there may
be significant amounts of memory available in the lower zones.
Even the most basic test -- repeatedly reading a file slightly bigger than
memory -- shows how broken the zone aging is. In this scenario, no single
page should be able stay in memory long enough to get referenced twice and
activated, but activation happens in spades:
$ grep active_file /proc/zoneinfo
nr_inactive_file 0
nr_active_file 0
nr_inactive_file 0
nr_active_file 8
nr_inactive_file 1582
nr_active_file 11994
$ cat data data data data >/dev/null
$ grep active_file /proc/zoneinfo
nr_inactive_file 0
nr_active_file 70
nr_inactive_file 258753
nr_active_file 443214
nr_inactive_file 149793
nr_active_file 12021
Fix this with a very simple round robin allocator. Each zone is allowed a
batch of allocations that is proportional to the zone's size, after which
it is treated as full. The batch counters are reset when all zones have
been tried and the allocator enters the slowpath and kicks off kswapd
reclaim. Allocation and reclaim is now fairly spread out to all
available/allowable zones:
$ grep active_file /proc/zoneinfo
nr_inactive_file 0
nr_active_file 0
nr_inactive_file 174
nr_active_file 4865
nr_inactive_file 53
nr_active_file 860
$ cat data data data data >/dev/null
$ grep active_file /proc/zoneinfo
nr_inactive_file 0
nr_active_file 0
nr_inactive_file 666622
nr_active_file 4988
nr_inactive_file 190969
nr_active_file 937
When zone_reclaim_mode is enabled, allocations will now spread out to all
zones on the local node, not just the first preferred zone (which on a 4G
node might be a tiny Normal zone).
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Paul Bolle <paul.bollee@gmail.com>
Cc: Zlatko Calusic <zcalusic@bitsync.net>
Tested-by: Kevin Hilman <khilman@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-11 21:20:47 +00:00
|
|
|
"nr_alloc_batch",
|
2011-05-25 00:11:28 +00:00
|
|
|
"nr_inactive_anon",
|
|
|
|
"nr_active_anon",
|
|
|
|
"nr_inactive_file",
|
|
|
|
"nr_active_file",
|
|
|
|
"nr_unevictable",
|
|
|
|
"nr_mlock",
|
|
|
|
"nr_anon_pages",
|
|
|
|
"nr_mapped",
|
|
|
|
"nr_file_pages",
|
|
|
|
"nr_dirty",
|
|
|
|
"nr_writeback",
|
|
|
|
"nr_slab_reclaimable",
|
|
|
|
"nr_slab_unreclaimable",
|
|
|
|
"nr_page_table_pages",
|
|
|
|
"nr_kernel_stack",
|
|
|
|
"nr_unstable",
|
|
|
|
"nr_bounce",
|
|
|
|
"nr_vmscan_write",
|
2011-11-01 00:07:59 +00:00
|
|
|
"nr_vmscan_immediate_reclaim",
|
2011-05-25 00:11:28 +00:00
|
|
|
"nr_writeback_temp",
|
|
|
|
"nr_isolated_anon",
|
|
|
|
"nr_isolated_file",
|
|
|
|
"nr_shmem",
|
|
|
|
"nr_dirtied",
|
|
|
|
"nr_written",
|
2014-08-06 23:07:16 +00:00
|
|
|
"nr_pages_scanned",
|
2011-05-25 00:11:28 +00:00
|
|
|
|
|
|
|
#ifdef CONFIG_NUMA
|
|
|
|
"numa_hit",
|
|
|
|
"numa_miss",
|
|
|
|
"numa_foreign",
|
|
|
|
"numa_interleave",
|
|
|
|
"numa_local",
|
|
|
|
"numa_other",
|
|
|
|
#endif
|
2014-04-03 21:47:51 +00:00
|
|
|
"workingset_refault",
|
|
|
|
"workingset_activate",
|
mm: keep page cache radix tree nodes in check
Previously, page cache radix tree nodes were freed after reclaim emptied
out their page pointers. But now reclaim stores shadow entries in their
place, which are only reclaimed when the inodes themselves are
reclaimed. This is problematic for bigger files that are still in use
after they have a significant amount of their cache reclaimed, without
any of those pages actually refaulting. The shadow entries will just
sit there and waste memory. In the worst case, the shadow entries will
accumulate until the machine runs out of memory.
To get this under control, the VM will track radix tree nodes
exclusively containing shadow entries on a per-NUMA node list. Per-NUMA
rather than global because we expect the radix tree nodes themselves to
be allocated node-locally and we want to reduce cross-node references of
otherwise independent cache workloads. A simple shrinker will then
reclaim these nodes on memory pressure.
A few things need to be stored in the radix tree node to implement the
shadow node LRU and allow tree deletions coming from the list:
1. There is no index available that would describe the reverse path
from the node up to the tree root, which is needed to perform a
deletion. To solve this, encode in each node its offset inside the
parent. This can be stored in the unused upper bits of the same
member that stores the node's height at no extra space cost.
2. The number of shadow entries needs to be counted in addition to the
regular entries, to quickly detect when the node is ready to go to
the shadow node LRU list. The current entry count is an unsigned
int but the maximum number of entries is 64, so a shadow counter
can easily be stored in the unused upper bits.
3. Tree modification needs tree lock and tree root, which are located
in the address space, so store an address_space backpointer in the
node. The parent pointer of the node is in a union with the 2-word
rcu_head, so the backpointer comes at no extra cost as well.
4. The node needs to be linked to an LRU list, which requires a list
head inside the node. This does increase the size of the node, but
it does not change the number of objects that fit into a slab page.
[akpm@linux-foundation.org: export the right function]
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Jan Kara <jack@suse.cz>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Luigi Semenzato <semenzato@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Metin Doslu <metin@citusdata.com>
Cc: Michel Lespinasse <walken@google.com>
Cc: Ozgun Erdogan <ozgun@citusdata.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roman Gushchin <klamm@yandex-team.ru>
Cc: Ryan Mallon <rmallon@gmail.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-04-03 21:47:56 +00:00
|
|
|
"workingset_nodereclaim",
|
2011-05-25 00:11:28 +00:00
|
|
|
"nr_anon_transparent_hugepages",
|
2012-10-08 23:32:02 +00:00
|
|
|
"nr_free_cma",
|
2014-10-09 22:29:32 +00:00
|
|
|
|
|
|
|
/* enum writeback_stat_item counters */
|
2011-05-25 00:11:28 +00:00
|
|
|
"nr_dirty_threshold",
|
|
|
|
"nr_dirty_background_threshold",
|
|
|
|
|
|
|
|
#ifdef CONFIG_VM_EVENT_COUNTERS
|
2014-10-09 22:29:32 +00:00
|
|
|
/* enum vm_event_item counters */
|
2011-05-25 00:11:28 +00:00
|
|
|
"pgpgin",
|
|
|
|
"pgpgout",
|
|
|
|
"pswpin",
|
|
|
|
"pswpout",
|
|
|
|
|
|
|
|
TEXTS_FOR_ZONES("pgalloc")
|
|
|
|
|
|
|
|
"pgfree",
|
|
|
|
"pgactivate",
|
|
|
|
"pgdeactivate",
|
|
|
|
|
|
|
|
"pgfault",
|
|
|
|
"pgmajfault",
|
mm: support madvise(MADV_FREE)
Linux doesn't have an ability to free pages lazy while other OS already
have been supported that named by madvise(MADV_FREE).
The gain is clear that kernel can discard freed pages rather than
swapping out or OOM if memory pressure happens.
Without memory pressure, freed pages would be reused by userspace
without another additional overhead(ex, page fault + allocation +
zeroing).
Jason Evans said:
: Facebook has been using MAP_UNINITIALIZED
: (https://lkml.org/lkml/2012/1/18/308) in some of its applications for
: several years, but there are operational costs to maintaining this
: out-of-tree in our kernel and in jemalloc, and we are anxious to retire it
: in favor of MADV_FREE. When we first enabled MAP_UNINITIALIZED it
: increased throughput for much of our workload by ~5%, and although the
: benefit has decreased using newer hardware and kernels, there is still
: enough benefit that we cannot reasonably retire it without a replacement.
:
: Aside from Facebook operations, there are numerous broadly used
: applications that would benefit from MADV_FREE. The ones that immediately
: come to mind are redis, varnish, and MariaDB. I don't have much insight
: into Android internals and development process, but I would hope to see
: MADV_FREE support eventually end up there as well to benefit applications
: linked with the integrated jemalloc.
:
: jemalloc will use MADV_FREE once it becomes available in the Linux kernel.
: In fact, jemalloc already uses MADV_FREE or equivalent everywhere it's
: available: *BSD, OS X, Windows, and Solaris -- every platform except Linux
: (and AIX, but I'm not sure it even compiles on AIX). The lack of
: MADV_FREE on Linux forced me down a long series of increasingly
: sophisticated heuristics for madvise() volume reduction, and even so this
: remains a common performance issue for people using jemalloc on Linux.
: Please integrate MADV_FREE; many people will benefit substantially.
How it works:
When madvise syscall is called, VM clears dirty bit of ptes of the
range. If memory pressure happens, VM checks dirty bit of page table
and if it found still "clean", it means it's a "lazyfree pages" so VM
could discard the page instead of swapping out. Once there was store
operation for the page before VM peek a page to reclaim, dirty bit is
set so VM can swap out the page instead of discarding.
One thing we should notice is that basically, MADV_FREE relies on dirty
bit in page table entry to decide whether VM allows to discard the page
or not. IOW, if page table entry includes marked dirty bit, VM
shouldn't discard the page.
However, as a example, if swap-in by read fault happens, page table
entry doesn't have dirty bit so MADV_FREE could discard the page
wrongly.
For avoiding the problem, MADV_FREE did more checks with PageDirty and
PageSwapCache. It worked out because swapped-in page lives on swap
cache and since it is evicted from the swap cache, the page has PG_dirty
flag. So both page flags check effectively prevent wrong discarding by
MADV_FREE.
However, a problem in above logic is that swapped-in page has PG_dirty
still after they are removed from swap cache so VM cannot consider the
page as freeable any more even if madvise_free is called in future.
Look at below example for detail.
ptr = malloc();
memset(ptr);
..
..
.. heavy memory pressure so all of pages are swapped out
..
..
var = *ptr; -> a page swapped-in and could be removed from
swapcache. Then, page table doesn't mark
dirty bit and page descriptor includes PG_dirty
..
..
madvise_free(ptr); -> It doesn't clear PG_dirty of the page.
..
..
..
.. heavy memory pressure again.
.. In this time, VM cannot discard the page because the page
.. has *PG_dirty*
To solve the problem, this patch clears PG_dirty if only the page is
owned exclusively by current process when madvise is called because
PG_dirty represents ptes's dirtiness in several processes so we could
clear it only if we own it exclusively.
Firstly, heavy users would be general allocators(ex, jemalloc, tcmalloc
and hope glibc supports it) and jemalloc/tcmalloc already have supported
the feature for other OS(ex, FreeBSD)
barrios@blaptop:~/benchmark/ebizzy$ lscpu
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 12
On-line CPU(s) list: 0-11
Thread(s) per core: 1
Core(s) per socket: 1
Socket(s): 12
NUMA node(s): 1
Vendor ID: GenuineIntel
CPU family: 6
Model: 2
Stepping: 3
CPU MHz: 3200.185
BogoMIPS: 6400.53
Virtualization: VT-x
Hypervisor vendor: KVM
Virtualization type: full
L1d cache: 32K
L1i cache: 32K
L2 cache: 4096K
NUMA node0 CPU(s): 0-11
ebizzy benchmark(./ebizzy -S 10 -n 512)
Higher avg is better.
vanilla-jemalloc MADV_free-jemalloc
1 thread
records: 10 records: 10
avg: 2961.90 avg: 12069.70
std: 71.96(2.43%) std: 186.68(1.55%)
max: 3070.00 max: 12385.00
min: 2796.00 min: 11746.00
2 thread
records: 10 records: 10
avg: 5020.00 avg: 17827.00
std: 264.87(5.28%) std: 358.52(2.01%)
max: 5244.00 max: 18760.00
min: 4251.00 min: 17382.00
4 thread
records: 10 records: 10
avg: 8988.80 avg: 27930.80
std: 1175.33(13.08%) std: 3317.33(11.88%)
max: 9508.00 max: 30879.00
min: 5477.00 min: 21024.00
8 thread
records: 10 records: 10
avg: 13036.50 avg: 33739.40
std: 170.67(1.31%) std: 5146.22(15.25%)
max: 13371.00 max: 40572.00
min: 12785.00 min: 24088.00
16 thread
records: 10 records: 10
avg: 11092.40 avg: 31424.20
std: 710.60(6.41%) std: 3763.89(11.98%)
max: 12446.00 max: 36635.00
min: 9949.00 min: 25669.00
32 thread
records: 10 records: 10
avg: 11067.00 avg: 34495.80
std: 971.06(8.77%) std: 2721.36(7.89%)
max: 12010.00 max: 38598.00
min: 9002.00 min: 30636.00
In summary, MADV_FREE is about much faster than MADV_DONTNEED.
This patch (of 12):
Add core MADV_FREE implementation.
[akpm@linux-foundation.org: small cleanups]
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Mika Penttil <mika.penttila@nextfour.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Jason Evans <je@fb.com>
Cc: Daniel Micay <danielmicay@gmail.com>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Shaohua Li <shli@kernel.org>
Cc: <yalin.wang2010@gmail.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: "James E.J. Bottomley" <jejb@parisc-linux.org>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: "Shaohua Li" <shli@kernel.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Chen Gang <gang.chen.5i5j@gmail.com>
Cc: Chris Zankel <chris@zankel.net>
Cc: Darrick J. Wong <darrick.wong@oracle.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Helge Deller <deller@gmx.de>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Roland Dreier <roland@kernel.org>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Shaohua Li <shli@kernel.org>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:54:53 +00:00
|
|
|
"pglazyfreed",
|
2011-05-25 00:11:28 +00:00
|
|
|
|
|
|
|
TEXTS_FOR_ZONES("pgrefill")
|
2012-04-25 23:01:48 +00:00
|
|
|
TEXTS_FOR_ZONES("pgsteal_kswapd")
|
|
|
|
TEXTS_FOR_ZONES("pgsteal_direct")
|
2011-05-25 00:11:28 +00:00
|
|
|
TEXTS_FOR_ZONES("pgscan_kswapd")
|
|
|
|
TEXTS_FOR_ZONES("pgscan_direct")
|
2012-07-31 23:44:39 +00:00
|
|
|
"pgscan_direct_throttle",
|
2011-05-25 00:11:28 +00:00
|
|
|
|
|
|
|
#ifdef CONFIG_NUMA
|
|
|
|
"zone_reclaim_failed",
|
|
|
|
#endif
|
|
|
|
"pginodesteal",
|
|
|
|
"slabs_scanned",
|
|
|
|
"kswapd_inodesteal",
|
|
|
|
"kswapd_low_wmark_hit_quickly",
|
|
|
|
"kswapd_high_wmark_hit_quickly",
|
|
|
|
"pageoutrun",
|
|
|
|
"allocstall",
|
|
|
|
|
|
|
|
"pgrotated",
|
|
|
|
|
2014-04-03 21:48:19 +00:00
|
|
|
"drop_pagecache",
|
|
|
|
"drop_slab",
|
|
|
|
|
2012-11-02 14:52:48 +00:00
|
|
|
#ifdef CONFIG_NUMA_BALANCING
|
|
|
|
"numa_pte_updates",
|
2013-11-12 23:08:32 +00:00
|
|
|
"numa_huge_pte_updates",
|
2012-11-02 14:52:48 +00:00
|
|
|
"numa_hint_faults",
|
|
|
|
"numa_hint_faults_local",
|
|
|
|
"numa_pages_migrated",
|
|
|
|
#endif
|
2012-10-19 09:46:20 +00:00
|
|
|
#ifdef CONFIG_MIGRATION
|
|
|
|
"pgmigrate_success",
|
|
|
|
"pgmigrate_fail",
|
|
|
|
#endif
|
2011-05-25 00:11:28 +00:00
|
|
|
#ifdef CONFIG_COMPACTION
|
2012-10-19 11:00:10 +00:00
|
|
|
"compact_migrate_scanned",
|
|
|
|
"compact_free_scanned",
|
|
|
|
"compact_isolated",
|
2011-05-25 00:11:28 +00:00
|
|
|
"compact_stall",
|
|
|
|
"compact_fail",
|
|
|
|
"compact_success",
|
mm, compaction: introduce kcompactd
Memory compaction can be currently performed in several contexts:
- kswapd balancing a zone after a high-order allocation failure
- direct compaction to satisfy a high-order allocation, including THP
page fault attemps
- khugepaged trying to collapse a hugepage
- manually from /proc
The purpose of compaction is two-fold. The obvious purpose is to
satisfy a (pending or future) high-order allocation, and is easy to
evaluate. The other purpose is to keep overal memory fragmentation low
and help the anti-fragmentation mechanism. The success wrt the latter
purpose is more
The current situation wrt the purposes has a few drawbacks:
- compaction is invoked only when a high-order page or hugepage is not
available (or manually). This might be too late for the purposes of
keeping memory fragmentation low.
- direct compaction increases latency of allocations. Again, it would
be better if compaction was performed asynchronously to keep
fragmentation low, before the allocation itself comes.
- (a special case of the previous) the cost of compaction during THP
page faults can easily offset the benefits of THP.
- kswapd compaction appears to be complex, fragile and not working in
some scenarios. It could also end up compacting for a high-order
allocation request when it should be reclaiming memory for a later
order-0 request.
To improve the situation, we should be able to benefit from an
equivalent of kswapd, but for compaction - i.e. a background thread
which responds to fragmentation and the need for high-order allocations
(including hugepages) somewhat proactively.
One possibility is to extend the responsibilities of kswapd, which could
however complicate its design too much. It should be better to let
kswapd handle reclaim, as order-0 allocations are often more critical
than high-order ones.
Another possibility is to extend khugepaged, but this kthread is a
single instance and tied to THP configs.
This patch goes with the option of a new set of per-node kthreads called
kcompactd, and lays the foundations, without introducing any new
tunables. The lifecycle mimics kswapd kthreads, including the memory
hotplug hooks.
For compaction, kcompactd uses the standard compaction_suitable() and
ompact_finished() criteria and the deferred compaction functionality.
Unlike direct compaction, it uses only sync compaction, as there's no
allocation latency to minimize.
This patch doesn't yet add a call to wakeup_kcompactd. The kswapd
compact/reclaim loop for high-order pages will be replaced by waking up
kcompactd in the next patch with the description of what's wrong with
the old approach.
Waking up of the kcompactd threads is also tied to kswapd activity and
follows these rules:
- we don't want to affect any fastpaths, so wake up kcompactd only from
the slowpath, as it's done for kswapd
- if kswapd is doing reclaim, it's more important than compaction, so
don't invoke kcompactd until kswapd goes to sleep
- the target order used for kswapd is passed to kcompactd
Future possible future uses for kcompactd include the ability to wake up
kcompactd on demand in special situations, such as when hugepages are
not available (currently not done due to __GFP_NO_KSWAPD) or when a
fragmentation event (i.e. __rmqueue_fallback()) occurs. It's also
possible to perform periodic compaction with kcompactd.
[arnd@arndb.de: fix build errors with kcompactd]
[paul.gortmaker@windriver.com: don't use modular references for non modular code]
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: David Rientjes <rientjes@google.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-17 21:18:08 +00:00
|
|
|
"compact_daemon_wake",
|
2011-05-25 00:11:28 +00:00
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifdef CONFIG_HUGETLB_PAGE
|
|
|
|
"htlb_buddy_alloc_success",
|
|
|
|
"htlb_buddy_alloc_fail",
|
|
|
|
#endif
|
|
|
|
"unevictable_pgs_culled",
|
|
|
|
"unevictable_pgs_scanned",
|
|
|
|
"unevictable_pgs_rescued",
|
|
|
|
"unevictable_pgs_mlocked",
|
|
|
|
"unevictable_pgs_munlocked",
|
|
|
|
"unevictable_pgs_cleared",
|
|
|
|
"unevictable_pgs_stranded",
|
|
|
|
|
|
|
|
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
|
|
|
|
"thp_fault_alloc",
|
|
|
|
"thp_fault_fallback",
|
|
|
|
"thp_collapse_alloc",
|
|
|
|
"thp_collapse_alloc_failed",
|
2016-01-16 00:52:46 +00:00
|
|
|
"thp_split_page",
|
|
|
|
"thp_split_page_failed",
|
2016-03-17 21:18:45 +00:00
|
|
|
"thp_deferred_split_page",
|
2016-01-16 00:52:46 +00:00
|
|
|
"thp_split_pmd",
|
2012-12-12 21:51:09 +00:00
|
|
|
"thp_zero_page_alloc",
|
|
|
|
"thp_zero_page_alloc_failed",
|
2011-05-25 00:11:28 +00:00
|
|
|
#endif
|
2014-10-09 22:29:32 +00:00
|
|
|
#ifdef CONFIG_MEMORY_BALLOON
|
|
|
|
"balloon_inflate",
|
|
|
|
"balloon_deflate",
|
|
|
|
#ifdef CONFIG_BALLOON_COMPACTION
|
|
|
|
"balloon_migrate",
|
|
|
|
#endif
|
|
|
|
#endif /* CONFIG_MEMORY_BALLOON */
|
2014-01-21 22:33:16 +00:00
|
|
|
#ifdef CONFIG_DEBUG_TLBFLUSH
|
2013-09-11 21:20:24 +00:00
|
|
|
#ifdef CONFIG_SMP
|
2013-09-11 21:20:23 +00:00
|
|
|
"nr_tlb_remote_flush",
|
|
|
|
"nr_tlb_remote_flush_received",
|
2014-01-21 22:33:16 +00:00
|
|
|
#endif /* CONFIG_SMP */
|
2013-09-11 21:20:23 +00:00
|
|
|
"nr_tlb_local_flush_all",
|
|
|
|
"nr_tlb_local_flush_one",
|
2014-01-21 22:33:16 +00:00
|
|
|
#endif /* CONFIG_DEBUG_TLBFLUSH */
|
2011-05-25 00:11:28 +00:00
|
|
|
|
2014-06-04 23:06:46 +00:00
|
|
|
#ifdef CONFIG_DEBUG_VM_VMACACHE
|
|
|
|
"vmacache_find_calls",
|
|
|
|
"vmacache_find_hits",
|
2014-12-13 00:56:10 +00:00
|
|
|
"vmacache_full_flushes",
|
2014-06-04 23:06:46 +00:00
|
|
|
#endif
|
2011-05-25 00:11:28 +00:00
|
|
|
#endif /* CONFIG_VM_EVENTS_COUNTERS */
|
|
|
|
};
|
2011-09-14 23:21:05 +00:00
|
|
|
#endif /* CONFIG_PROC_FS || CONFIG_SYSFS || CONFIG_NUMA */
|
2011-05-25 00:11:28 +00:00
|
|
|
|
|
|
|
|
2015-02-10 22:09:43 +00:00
|
|
|
#if (defined(CONFIG_DEBUG_FS) && defined(CONFIG_COMPACTION)) || \
|
|
|
|
defined(CONFIG_PROC_FS)
|
|
|
|
static void *frag_start(struct seq_file *m, loff_t *pos)
|
|
|
|
{
|
|
|
|
pg_data_t *pgdat;
|
|
|
|
loff_t node = *pos;
|
|
|
|
|
|
|
|
for (pgdat = first_online_pgdat();
|
|
|
|
pgdat && node;
|
|
|
|
pgdat = next_online_pgdat(pgdat))
|
|
|
|
--node;
|
|
|
|
|
|
|
|
return pgdat;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void *frag_next(struct seq_file *m, void *arg, loff_t *pos)
|
|
|
|
{
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
|
|
|
|
(*pos)++;
|
|
|
|
return next_online_pgdat(pgdat);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void frag_stop(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Walk all the zones in a node and print using a callback */
|
|
|
|
static void walk_zones_in_node(struct seq_file *m, pg_data_t *pgdat,
|
|
|
|
void (*print)(struct seq_file *m, pg_data_t *, struct zone *))
|
|
|
|
{
|
|
|
|
struct zone *zone;
|
|
|
|
struct zone *node_zones = pgdat->node_zones;
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
for (zone = node_zones; zone - node_zones < MAX_NR_ZONES; ++zone) {
|
|
|
|
if (!populated_zone(zone))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&zone->lock, flags);
|
|
|
|
print(m, pgdat, zone);
|
|
|
|
spin_unlock_irqrestore(&zone->lock, flags);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2010-05-24 21:32:25 +00:00
|
|
|
#ifdef CONFIG_PROC_FS
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
static void frag_show_print(struct seq_file *m, pg_data_t *pgdat,
|
|
|
|
struct zone *zone)
|
|
|
|
{
|
|
|
|
int order;
|
|
|
|
|
|
|
|
seq_printf(m, "Node %d, zone %8s ", pgdat->node_id, zone->name);
|
|
|
|
for (order = 0; order < MAX_ORDER; ++order)
|
|
|
|
seq_printf(m, "%6lu ", zone->free_area[order].nr_free);
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This walks the free areas for each zone.
|
|
|
|
*/
|
|
|
|
static int frag_show(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
walk_zones_in_node(m, pgdat, frag_show_print);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void pagetypeinfo_showfree_print(struct seq_file *m,
|
|
|
|
pg_data_t *pgdat, struct zone *zone)
|
|
|
|
{
|
|
|
|
int order, mtype;
|
|
|
|
|
|
|
|
for (mtype = 0; mtype < MIGRATE_TYPES; mtype++) {
|
|
|
|
seq_printf(m, "Node %4d, zone %8s, type %12s ",
|
|
|
|
pgdat->node_id,
|
|
|
|
zone->name,
|
|
|
|
migratetype_names[mtype]);
|
|
|
|
for (order = 0; order < MAX_ORDER; ++order) {
|
|
|
|
unsigned long freecount = 0;
|
|
|
|
struct free_area *area;
|
|
|
|
struct list_head *curr;
|
|
|
|
|
|
|
|
area = &(zone->free_area[order]);
|
|
|
|
|
|
|
|
list_for_each(curr, &area->free_list[mtype])
|
|
|
|
freecount++;
|
|
|
|
seq_printf(m, "%6lu ", freecount);
|
|
|
|
}
|
2006-06-30 08:55:32 +00:00
|
|
|
seq_putc(m, '\n');
|
|
|
|
}
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Print out the free pages at each order for each migatetype */
|
|
|
|
static int pagetypeinfo_showfree(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
int order;
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
|
|
|
|
/* Print header */
|
|
|
|
seq_printf(m, "%-43s ", "Free pages count per migrate type at order");
|
|
|
|
for (order = 0; order < MAX_ORDER; ++order)
|
|
|
|
seq_printf(m, "%6d ", order);
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
|
|
|
|
walk_zones_in_node(m, pgdat, pagetypeinfo_showfree_print);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void pagetypeinfo_showblockcount_print(struct seq_file *m,
|
|
|
|
pg_data_t *pgdat, struct zone *zone)
|
|
|
|
{
|
|
|
|
int mtype;
|
|
|
|
unsigned long pfn;
|
|
|
|
unsigned long start_pfn = zone->zone_start_pfn;
|
2013-02-23 00:35:23 +00:00
|
|
|
unsigned long end_pfn = zone_end_pfn(zone);
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
unsigned long count[MIGRATE_TYPES] = { 0, };
|
|
|
|
|
|
|
|
for (pfn = start_pfn; pfn < end_pfn; pfn += pageblock_nr_pages) {
|
|
|
|
struct page *page;
|
|
|
|
|
|
|
|
if (!pfn_valid(pfn))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
page = pfn_to_page(pfn);
|
2009-05-13 16:34:48 +00:00
|
|
|
|
|
|
|
/* Watch for unexpected holes punched in the memmap */
|
|
|
|
if (!memmap_valid_within(pfn, page, zone))
|
2008-08-14 10:10:14 +00:00
|
|
|
continue;
|
2009-05-13 16:34:48 +00:00
|
|
|
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
mtype = get_pageblock_migratetype(page);
|
|
|
|
|
2008-08-14 10:10:14 +00:00
|
|
|
if (mtype < MIGRATE_TYPES)
|
|
|
|
count[mtype]++;
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Print counts */
|
|
|
|
seq_printf(m, "Node %d, zone %8s ", pgdat->node_id, zone->name);
|
|
|
|
for (mtype = 0; mtype < MIGRATE_TYPES; mtype++)
|
|
|
|
seq_printf(m, "%12lu ", count[mtype]);
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Print out the free pages at each order for each migratetype */
|
|
|
|
static int pagetypeinfo_showblockcount(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
int mtype;
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
|
|
|
|
seq_printf(m, "\n%-23s", "Number of blocks type ");
|
|
|
|
for (mtype = 0; mtype < MIGRATE_TYPES; mtype++)
|
|
|
|
seq_printf(m, "%12s ", migratetype_names[mtype]);
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
walk_zones_in_node(m, pgdat, pagetypeinfo_showblockcount_print);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
mm/page_owner: keep track of page owners
This is the page owner tracking code which is introduced so far ago. It
is resident on Andrew's tree, though, nobody tried to upstream so it
remain as is. Our company uses this feature actively to debug memory leak
or to find a memory hogger so I decide to upstream this feature.
This functionality help us to know who allocates the page. When
allocating a page, we store some information about allocation in extra
memory. Later, if we need to know status of all pages, we can get and
analyze it from this stored information.
In previous version of this feature, extra memory is statically defined in
struct page, but, in this version, extra memory is allocated outside of
struct page. It enables us to turn on/off this feature at boottime
without considerable memory waste.
Although we already have tracepoint for tracing page allocation/free,
using it to analyze page owner is rather complex. We need to enlarge the
trace buffer for preventing overlapping until userspace program launched.
And, launched program continually dump out the trace buffer for later
analysis and it would change system behaviour with more possibility rather
than just keeping it in memory, so bad for debug.
Moreover, we can use page_owner feature further for various purposes. For
example, we can use it for fragmentation statistics implemented in this
patch. And, I also plan to implement some CMA failure debugging feature
using this interface.
I'd like to give the credit for all developers contributed this feature,
but, it's not easy because I don't know exact history. Sorry about that.
Below is people who has "Signed-off-by" in the patches in Andrew's tree.
Contributor:
Alexander Nyberg <alexn@dsv.su.se>
Mel Gorman <mgorman@suse.de>
Dave Hansen <dave@linux.vnet.ibm.com>
Minchan Kim <minchan@kernel.org>
Michal Nazarewicz <mina86@mina86.com>
Andrew Morton <akpm@linux-foundation.org>
Jungsoo Son <jungsoo.son@lge.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Dave Hansen <dave@sr71.net>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Jungsoo Son <jungsoo.son@lge.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:01 +00:00
|
|
|
#ifdef CONFIG_PAGE_OWNER
|
|
|
|
static void pagetypeinfo_showmixedcount_print(struct seq_file *m,
|
|
|
|
pg_data_t *pgdat,
|
|
|
|
struct zone *zone)
|
|
|
|
{
|
|
|
|
struct page *page;
|
|
|
|
struct page_ext *page_ext;
|
|
|
|
unsigned long pfn = zone->zone_start_pfn, block_end_pfn;
|
|
|
|
unsigned long end_pfn = pfn + zone->spanned_pages;
|
|
|
|
unsigned long count[MIGRATE_TYPES] = { 0, };
|
|
|
|
int pageblock_mt, page_mt;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
/* Scan block by block. First and last block may be incomplete */
|
|
|
|
pfn = zone->zone_start_pfn;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Walk the zone in pageblock_nr_pages steps. If a page block spans
|
|
|
|
* a zone boundary, it will be double counted between zones. This does
|
|
|
|
* not matter as the mixed block count will still be correct
|
|
|
|
*/
|
|
|
|
for (; pfn < end_pfn; ) {
|
|
|
|
if (!pfn_valid(pfn)) {
|
|
|
|
pfn = ALIGN(pfn + 1, MAX_ORDER_NR_PAGES);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
block_end_pfn = ALIGN(pfn + 1, pageblock_nr_pages);
|
|
|
|
block_end_pfn = min(block_end_pfn, end_pfn);
|
|
|
|
|
|
|
|
page = pfn_to_page(pfn);
|
|
|
|
pageblock_mt = get_pfnblock_migratetype(page, pfn);
|
|
|
|
|
|
|
|
for (; pfn < block_end_pfn; pfn++) {
|
|
|
|
if (!pfn_valid_within(pfn))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
page = pfn_to_page(pfn);
|
|
|
|
if (PageBuddy(page)) {
|
|
|
|
pfn += (1UL << page_order(page)) - 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (PageReserved(page))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
page_ext = lookup_page_ext(page);
|
|
|
|
|
|
|
|
if (!test_bit(PAGE_EXT_OWNER, &page_ext->flags))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
page_mt = gfpflags_to_migratetype(page_ext->gfp_mask);
|
|
|
|
if (pageblock_mt != page_mt) {
|
|
|
|
if (is_migrate_cma(pageblock_mt))
|
|
|
|
count[MIGRATE_MOVABLE]++;
|
|
|
|
else
|
|
|
|
count[pageblock_mt]++;
|
|
|
|
|
|
|
|
pfn = block_end_pfn;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
pfn += (1UL << page_ext->order) - 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Print counts */
|
|
|
|
seq_printf(m, "Node %d, zone %8s ", pgdat->node_id, zone->name);
|
|
|
|
for (i = 0; i < MIGRATE_TYPES; i++)
|
|
|
|
seq_printf(m, "%12lu ", count[i]);
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
}
|
|
|
|
#endif /* CONFIG_PAGE_OWNER */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Print out the number of pageblocks for each migratetype that contain pages
|
|
|
|
* of other types. This gives an indication of how well fallbacks are being
|
|
|
|
* contained by rmqueue_fallback(). It requires information from PAGE_OWNER
|
|
|
|
* to determine what is going on
|
|
|
|
*/
|
|
|
|
static void pagetypeinfo_showmixedcount(struct seq_file *m, pg_data_t *pgdat)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_PAGE_OWNER
|
|
|
|
int mtype;
|
|
|
|
|
2016-03-15 21:56:12 +00:00
|
|
|
if (!static_branch_unlikely(&page_owner_inited))
|
mm/page_owner: keep track of page owners
This is the page owner tracking code which is introduced so far ago. It
is resident on Andrew's tree, though, nobody tried to upstream so it
remain as is. Our company uses this feature actively to debug memory leak
or to find a memory hogger so I decide to upstream this feature.
This functionality help us to know who allocates the page. When
allocating a page, we store some information about allocation in extra
memory. Later, if we need to know status of all pages, we can get and
analyze it from this stored information.
In previous version of this feature, extra memory is statically defined in
struct page, but, in this version, extra memory is allocated outside of
struct page. It enables us to turn on/off this feature at boottime
without considerable memory waste.
Although we already have tracepoint for tracing page allocation/free,
using it to analyze page owner is rather complex. We need to enlarge the
trace buffer for preventing overlapping until userspace program launched.
And, launched program continually dump out the trace buffer for later
analysis and it would change system behaviour with more possibility rather
than just keeping it in memory, so bad for debug.
Moreover, we can use page_owner feature further for various purposes. For
example, we can use it for fragmentation statistics implemented in this
patch. And, I also plan to implement some CMA failure debugging feature
using this interface.
I'd like to give the credit for all developers contributed this feature,
but, it's not easy because I don't know exact history. Sorry about that.
Below is people who has "Signed-off-by" in the patches in Andrew's tree.
Contributor:
Alexander Nyberg <alexn@dsv.su.se>
Mel Gorman <mgorman@suse.de>
Dave Hansen <dave@linux.vnet.ibm.com>
Minchan Kim <minchan@kernel.org>
Michal Nazarewicz <mina86@mina86.com>
Andrew Morton <akpm@linux-foundation.org>
Jungsoo Son <jungsoo.son@lge.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Dave Hansen <dave@sr71.net>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Jungsoo Son <jungsoo.son@lge.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:01 +00:00
|
|
|
return;
|
|
|
|
|
|
|
|
drain_all_pages(NULL);
|
|
|
|
|
|
|
|
seq_printf(m, "\n%-23s", "Number of mixed blocks ");
|
|
|
|
for (mtype = 0; mtype < MIGRATE_TYPES; mtype++)
|
|
|
|
seq_printf(m, "%12s ", migratetype_names[mtype]);
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
|
|
|
|
walk_zones_in_node(m, pgdat, pagetypeinfo_showmixedcount_print);
|
|
|
|
#endif /* CONFIG_PAGE_OWNER */
|
|
|
|
}
|
|
|
|
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
/*
|
|
|
|
* This prints out statistics in relation to grouping pages by mobility.
|
|
|
|
* It is expensive to collect so do not constantly read the file.
|
|
|
|
*/
|
|
|
|
static int pagetypeinfo_show(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
|
2008-04-30 07:52:13 +00:00
|
|
|
/* check memoryless node */
|
2012-12-12 21:51:37 +00:00
|
|
|
if (!node_state(pgdat->node_id, N_MEMORY))
|
2008-04-30 07:52:13 +00:00
|
|
|
return 0;
|
|
|
|
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
seq_printf(m, "Page block order: %d\n", pageblock_order);
|
|
|
|
seq_printf(m, "Pages per block: %lu\n", pageblock_nr_pages);
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
pagetypeinfo_showfree(m, pgdat);
|
|
|
|
pagetypeinfo_showblockcount(m, pgdat);
|
mm/page_owner: keep track of page owners
This is the page owner tracking code which is introduced so far ago. It
is resident on Andrew's tree, though, nobody tried to upstream so it
remain as is. Our company uses this feature actively to debug memory leak
or to find a memory hogger so I decide to upstream this feature.
This functionality help us to know who allocates the page. When
allocating a page, we store some information about allocation in extra
memory. Later, if we need to know status of all pages, we can get and
analyze it from this stored information.
In previous version of this feature, extra memory is statically defined in
struct page, but, in this version, extra memory is allocated outside of
struct page. It enables us to turn on/off this feature at boottime
without considerable memory waste.
Although we already have tracepoint for tracing page allocation/free,
using it to analyze page owner is rather complex. We need to enlarge the
trace buffer for preventing overlapping until userspace program launched.
And, launched program continually dump out the trace buffer for later
analysis and it would change system behaviour with more possibility rather
than just keeping it in memory, so bad for debug.
Moreover, we can use page_owner feature further for various purposes. For
example, we can use it for fragmentation statistics implemented in this
patch. And, I also plan to implement some CMA failure debugging feature
using this interface.
I'd like to give the credit for all developers contributed this feature,
but, it's not easy because I don't know exact history. Sorry about that.
Below is people who has "Signed-off-by" in the patches in Andrew's tree.
Contributor:
Alexander Nyberg <alexn@dsv.su.se>
Mel Gorman <mgorman@suse.de>
Dave Hansen <dave@linux.vnet.ibm.com>
Minchan Kim <minchan@kernel.org>
Michal Nazarewicz <mina86@mina86.com>
Andrew Morton <akpm@linux-foundation.org>
Jungsoo Son <jungsoo.son@lge.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Dave Hansen <dave@sr71.net>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Jungsoo Son <jungsoo.son@lge.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:01 +00:00
|
|
|
pagetypeinfo_showmixedcount(m, pgdat);
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
|
2006-06-30 08:55:32 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-10-06 00:13:52 +00:00
|
|
|
static const struct seq_operations fragmentation_op = {
|
2006-06-30 08:55:32 +00:00
|
|
|
.start = frag_start,
|
|
|
|
.next = frag_next,
|
|
|
|
.stop = frag_stop,
|
|
|
|
.show = frag_show,
|
|
|
|
};
|
|
|
|
|
2008-10-06 00:13:52 +00:00
|
|
|
static int fragmentation_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return seq_open(file, &fragmentation_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations fragmentation_file_operations = {
|
|
|
|
.open = fragmentation_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.release = seq_release,
|
|
|
|
};
|
|
|
|
|
2008-10-06 00:15:36 +00:00
|
|
|
static const struct seq_operations pagetypeinfo_op = {
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
.start = frag_start,
|
|
|
|
.next = frag_next,
|
|
|
|
.stop = frag_stop,
|
|
|
|
.show = pagetypeinfo_show,
|
|
|
|
};
|
|
|
|
|
2008-10-06 00:15:36 +00:00
|
|
|
static int pagetypeinfo_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return seq_open(file, &pagetypeinfo_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations pagetypeinfo_file_ops = {
|
|
|
|
.open = pagetypeinfo_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.release = seq_release,
|
|
|
|
};
|
|
|
|
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
static void zoneinfo_show_print(struct seq_file *m, pg_data_t *pgdat,
|
|
|
|
struct zone *zone)
|
2006-06-30 08:55:32 +00:00
|
|
|
{
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
int i;
|
|
|
|
seq_printf(m, "Node %d, zone %8s", pgdat->node_id, zone->name);
|
|
|
|
seq_printf(m,
|
|
|
|
"\n pages free %lu"
|
|
|
|
"\n min %lu"
|
|
|
|
"\n low %lu"
|
|
|
|
"\n high %lu"
|
2009-06-16 22:32:30 +00:00
|
|
|
"\n scanned %lu"
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
"\n spanned %lu"
|
2012-12-12 21:52:12 +00:00
|
|
|
"\n present %lu"
|
|
|
|
"\n managed %lu",
|
mm: page allocator: adjust the per-cpu counter threshold when memory is low
Commit aa45484 ("calculate a better estimate of NR_FREE_PAGES when memory
is low") noted that watermarks were based on the vmstat NR_FREE_PAGES. To
avoid synchronization overhead, these counters are maintained on a per-cpu
basis and drained both periodically and when a threshold is above a
threshold. On large CPU systems, the difference between the estimate and
real value of NR_FREE_PAGES can be very high. The system can get into a
case where pages are allocated far below the min watermark potentially
causing livelock issues. The commit solved the problem by taking a better
reading of NR_FREE_PAGES when memory was low.
Unfortately, as reported by Shaohua Li this accurate reading can consume a
large amount of CPU time on systems with many sockets due to cache line
bouncing. This patch takes a different approach. For large machines
where counter drift might be unsafe and while kswapd is awake, the per-cpu
thresholds for the target pgdat are reduced to limit the level of drift to
what should be a safe level. This incurs a performance penalty in heavy
memory pressure by a factor that depends on the workload and the machine
but the machine should function correctly without accidentally exhausting
all memory on a node. There is an additional cost when kswapd wakes and
sleeps but the event is not expected to be frequent - in Shaohua's test
case, there was one recorded sleep and wake event at least.
To ensure that kswapd wakes up, a safe version of zone_watermark_ok() is
introduced that takes a more accurate reading of NR_FREE_PAGES when called
from wakeup_kswapd, when deciding whether it is really safe to go back to
sleep in sleeping_prematurely() and when deciding if a zone is really
balanced or not in balance_pgdat(). We are still using an expensive
function but limiting how often it is called.
When the test case is reproduced, the time spent in the watermark
functions is reduced. The following report is on the percentage of time
spent cumulatively spent in the functions zone_nr_free_pages(),
zone_watermark_ok(), __zone_watermark_ok(), zone_watermark_ok_safe(),
zone_page_state_snapshot(), zone_page_state().
vanilla 11.6615%
disable-threshold 0.2584%
David said:
: We had to pull aa454840 "mm: page allocator: calculate a better estimate
: of NR_FREE_PAGES when memory is low and kswapd is awake" from 2.6.36
: internally because tests showed that it would cause the machine to stall
: as the result of heavy kswapd activity. I merged it back with this fix as
: it is pending in the -mm tree and it solves the issue we were seeing, so I
: definitely think this should be pushed to -stable (and I would seriously
: consider it for 2.6.37 inclusion even at this late date).
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reported-by: Shaohua Li <shaohua.li@intel.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Tested-by: Nicolas Bareil <nico@chdir.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: <stable@kernel.org> [2.6.37.1, 2.6.36.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:45:41 +00:00
|
|
|
zone_page_state(zone, NR_FREE_PAGES),
|
2009-06-16 22:32:12 +00:00
|
|
|
min_wmark_pages(zone),
|
|
|
|
low_wmark_pages(zone),
|
|
|
|
high_wmark_pages(zone),
|
2014-08-06 23:07:16 +00:00
|
|
|
zone_page_state(zone, NR_PAGES_SCANNED),
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
zone->spanned_pages,
|
2012-12-12 21:52:12 +00:00
|
|
|
zone->present_pages,
|
|
|
|
zone->managed_pages);
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
|
|
|
|
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
|
|
|
|
seq_printf(m, "\n %-12s %lu", vmstat_text[i],
|
|
|
|
zone_page_state(zone, i));
|
|
|
|
|
|
|
|
seq_printf(m,
|
2014-08-06 23:07:14 +00:00
|
|
|
"\n protection: (%ld",
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
zone->lowmem_reserve[0]);
|
|
|
|
for (i = 1; i < ARRAY_SIZE(zone->lowmem_reserve); i++)
|
2014-08-06 23:07:14 +00:00
|
|
|
seq_printf(m, ", %ld", zone->lowmem_reserve[i]);
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
seq_printf(m,
|
|
|
|
")"
|
|
|
|
"\n pagesets");
|
|
|
|
for_each_online_cpu(i) {
|
|
|
|
struct per_cpu_pageset *pageset;
|
|
|
|
|
2010-01-05 06:34:51 +00:00
|
|
|
pageset = per_cpu_ptr(zone->pageset, i);
|
2008-02-05 06:29:19 +00:00
|
|
|
seq_printf(m,
|
|
|
|
"\n cpu: %i"
|
|
|
|
"\n count: %i"
|
|
|
|
"\n high: %i"
|
|
|
|
"\n batch: %i",
|
|
|
|
i,
|
|
|
|
pageset->pcp.count,
|
|
|
|
pageset->pcp.high,
|
|
|
|
pageset->pcp.batch);
|
2006-09-01 04:27:35 +00:00
|
|
|
#ifdef CONFIG_SMP
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
seq_printf(m, "\n vm stats threshold: %d",
|
|
|
|
pageset->stat_threshold);
|
2006-09-01 04:27:35 +00:00
|
|
|
#endif
|
2006-06-30 08:55:32 +00:00
|
|
|
}
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
seq_printf(m,
|
|
|
|
"\n all_unreclaimable: %u"
|
2008-10-19 03:26:34 +00:00
|
|
|
"\n start_pfn: %lu"
|
|
|
|
"\n inactive_ratio: %u",
|
2013-09-11 21:22:36 +00:00
|
|
|
!zone_reclaimable(zone),
|
2008-10-19 03:26:34 +00:00
|
|
|
zone->zone_start_pfn,
|
|
|
|
zone->inactive_ratio);
|
Print out statistics in relation to fragmentation avoidance to /proc/pagetypeinfo
This patch provides fragmentation avoidance statistics via /proc/pagetypeinfo.
The information is collected only on request so there is no runtime overhead.
The statistics are in three parts:
The first part prints information on the size of blocks that pages are
being grouped on and looks like
Page block order: 10
Pages per block: 1024
The second part is a more detailed version of /proc/buddyinfo and looks like
Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
Node 0, zone DMA, type Unmovable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reclaimable 1 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 0 0
Node 0, zone DMA, type Reserve 0 4 4 0 0 0 0 1 0 1 0
Node 0, zone Normal, type Unmovable 111 8 4 4 2 3 1 0 0 0 0
Node 0, zone Normal, type Reclaimable 293 89 8 0 0 0 0 0 0 0 0
Node 0, zone Normal, type Movable 1 6 13 9 7 6 3 0 0 0 0
Node 0, zone Normal, type Reserve 0 0 0 0 0 0 0 0 0 0 4
The third part looks like
Number of blocks type Unmovable Reclaimable Movable Reserve
Node 0, zone DMA 0 1 2 1
Node 0, zone Normal 3 17 94 4
To walk the zones within a node with interrupts disabled, walk_zones_in_node()
is introduced and shared between /proc/buddyinfo, /proc/zoneinfo and
/proc/pagetypeinfo to reduce code duplication. It seems specific to what
vmstat.c requires but could be broken out as a general utility function in
mmzone.c if there were other other potential users.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-16 08:26:02 +00:00
|
|
|
seq_putc(m, '\n');
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Output information about zones in @pgdat.
|
|
|
|
*/
|
|
|
|
static int zoneinfo_show(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
walk_zones_in_node(m, pgdat, zoneinfo_show_print);
|
2006-06-30 08:55:32 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-10-06 00:19:42 +00:00
|
|
|
static const struct seq_operations zoneinfo_op = {
|
2006-06-30 08:55:32 +00:00
|
|
|
.start = frag_start, /* iterate over all zones. The same as in
|
|
|
|
* fragmentation. */
|
|
|
|
.next = frag_next,
|
|
|
|
.stop = frag_stop,
|
|
|
|
.show = zoneinfo_show,
|
|
|
|
};
|
|
|
|
|
2008-10-06 00:19:42 +00:00
|
|
|
static int zoneinfo_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return seq_open(file, &zoneinfo_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations proc_zoneinfo_file_operations = {
|
|
|
|
.open = zoneinfo_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.release = seq_release,
|
|
|
|
};
|
|
|
|
|
2010-10-26 21:21:36 +00:00
|
|
|
enum writeback_stat_item {
|
|
|
|
NR_DIRTY_THRESHOLD,
|
|
|
|
NR_DIRTY_BG_THRESHOLD,
|
|
|
|
NR_VM_WRITEBACK_STAT_ITEMS,
|
|
|
|
};
|
|
|
|
|
2006-06-30 08:55:32 +00:00
|
|
|
static void *vmstat_start(struct seq_file *m, loff_t *pos)
|
|
|
|
{
|
2006-06-30 08:55:33 +00:00
|
|
|
unsigned long *v;
|
2010-10-26 21:21:36 +00:00
|
|
|
int i, stat_items_size;
|
2006-06-30 08:55:32 +00:00
|
|
|
|
|
|
|
if (*pos >= ARRAY_SIZE(vmstat_text))
|
|
|
|
return NULL;
|
2010-10-26 21:21:36 +00:00
|
|
|
stat_items_size = NR_VM_ZONE_STAT_ITEMS * sizeof(unsigned long) +
|
|
|
|
NR_VM_WRITEBACK_STAT_ITEMS * sizeof(unsigned long);
|
2006-06-30 08:55:32 +00:00
|
|
|
|
2006-06-30 08:55:45 +00:00
|
|
|
#ifdef CONFIG_VM_EVENT_COUNTERS
|
2010-10-26 21:21:36 +00:00
|
|
|
stat_items_size += sizeof(struct vm_event_state);
|
2006-06-30 08:55:45 +00:00
|
|
|
#endif
|
2010-10-26 21:21:36 +00:00
|
|
|
|
|
|
|
v = kmalloc(stat_items_size, GFP_KERNEL);
|
2006-06-30 08:55:33 +00:00
|
|
|
m->private = v;
|
|
|
|
if (!v)
|
2006-06-30 08:55:32 +00:00
|
|
|
return ERR_PTR(-ENOMEM);
|
2006-06-30 08:55:33 +00:00
|
|
|
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
|
|
|
|
v[i] = global_page_state(i);
|
2010-10-26 21:21:36 +00:00
|
|
|
v += NR_VM_ZONE_STAT_ITEMS;
|
|
|
|
|
|
|
|
global_dirty_limits(v + NR_DIRTY_BG_THRESHOLD,
|
|
|
|
v + NR_DIRTY_THRESHOLD);
|
|
|
|
v += NR_VM_WRITEBACK_STAT_ITEMS;
|
|
|
|
|
2006-06-30 08:55:45 +00:00
|
|
|
#ifdef CONFIG_VM_EVENT_COUNTERS
|
2010-10-26 21:21:36 +00:00
|
|
|
all_vm_events(v);
|
|
|
|
v[PGPGIN] /= 2; /* sectors -> kbytes */
|
|
|
|
v[PGPGOUT] /= 2;
|
2006-06-30 08:55:45 +00:00
|
|
|
#endif
|
2010-11-03 17:56:49 +00:00
|
|
|
return (unsigned long *)m->private + *pos;
|
2006-06-30 08:55:32 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void *vmstat_next(struct seq_file *m, void *arg, loff_t *pos)
|
|
|
|
{
|
|
|
|
(*pos)++;
|
|
|
|
if (*pos >= ARRAY_SIZE(vmstat_text))
|
|
|
|
return NULL;
|
|
|
|
return (unsigned long *)m->private + *pos;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int vmstat_show(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
unsigned long *l = arg;
|
|
|
|
unsigned long off = l - (unsigned long *)m->private;
|
|
|
|
|
|
|
|
seq_printf(m, "%s %lu\n", vmstat_text[off], *l);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void vmstat_stop(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
kfree(m->private);
|
|
|
|
m->private = NULL;
|
|
|
|
}
|
|
|
|
|
2008-10-06 00:17:48 +00:00
|
|
|
static const struct seq_operations vmstat_op = {
|
2006-06-30 08:55:32 +00:00
|
|
|
.start = vmstat_start,
|
|
|
|
.next = vmstat_next,
|
|
|
|
.stop = vmstat_stop,
|
|
|
|
.show = vmstat_show,
|
|
|
|
};
|
|
|
|
|
2008-10-06 00:17:48 +00:00
|
|
|
static int vmstat_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return seq_open(file, &vmstat_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations proc_vmstat_file_operations = {
|
|
|
|
.open = vmstat_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.release = seq_release,
|
|
|
|
};
|
2006-06-30 08:55:32 +00:00
|
|
|
#endif /* CONFIG_PROC_FS */
|
|
|
|
|
2006-09-01 04:27:35 +00:00
|
|
|
#ifdef CONFIG_SMP
|
2015-12-11 21:40:32 +00:00
|
|
|
static struct workqueue_struct *vmstat_wq;
|
2007-05-09 09:35:12 +00:00
|
|
|
static DEFINE_PER_CPU(struct delayed_work, vmstat_work);
|
2007-05-09 09:35:13 +00:00
|
|
|
int sysctl_stat_interval __read_mostly = HZ;
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
static cpumask_var_t cpu_stat_off;
|
2007-05-09 09:35:12 +00:00
|
|
|
|
|
|
|
static void vmstat_update(struct work_struct *w)
|
|
|
|
{
|
2016-01-14 23:21:40 +00:00
|
|
|
if (refresh_cpu_vm_stats(true)) {
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
/*
|
|
|
|
* Counters were updated so we expect more updates
|
|
|
|
* to occur in the future. Keep on running the
|
|
|
|
* update worker thread.
|
2016-02-05 23:36:24 +00:00
|
|
|
* If we were marked on cpu_stat_off clear the flag
|
|
|
|
* so that vmstat_shepherd doesn't schedule us again.
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
*/
|
2016-02-05 23:36:24 +00:00
|
|
|
if (!cpumask_test_and_clear_cpu(smp_processor_id(),
|
|
|
|
cpu_stat_off)) {
|
|
|
|
queue_delayed_work_on(smp_processor_id(), vmstat_wq,
|
|
|
|
this_cpu_ptr(&vmstat_work),
|
|
|
|
round_jiffies_relative(sysctl_stat_interval));
|
|
|
|
}
|
2015-10-15 20:01:50 +00:00
|
|
|
} else {
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
/*
|
|
|
|
* We did not update any counters so the app may be in
|
|
|
|
* a mode where it does not cause counter updates.
|
|
|
|
* We may be uselessly running vmstat_update.
|
|
|
|
* Defer the checking for differentials to the
|
|
|
|
* shepherd thread on a different processor.
|
|
|
|
*/
|
2016-01-22 16:46:14 +00:00
|
|
|
cpumask_set_cpu(smp_processor_id(), cpu_stat_off);
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-01-14 23:21:40 +00:00
|
|
|
/*
|
|
|
|
* Switch off vmstat processing and then fold all the remaining differentials
|
|
|
|
* until the diffs stay at zero. The function is used by NOHZ and can only be
|
|
|
|
* invoked when tick processing is not active.
|
|
|
|
*/
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
/*
|
|
|
|
* Check if the diffs for a certain cpu indicate that
|
|
|
|
* an update is needed.
|
|
|
|
*/
|
|
|
|
static bool need_update(int cpu)
|
|
|
|
{
|
|
|
|
struct zone *zone;
|
|
|
|
|
|
|
|
for_each_populated_zone(zone) {
|
|
|
|
struct per_cpu_pageset *p = per_cpu_ptr(zone->pageset, cpu);
|
|
|
|
|
|
|
|
BUILD_BUG_ON(sizeof(p->vm_stat_diff[0]) != 1);
|
|
|
|
/*
|
|
|
|
* The fast way of checking if there are any vmstat diffs.
|
|
|
|
* This works because the diffs are byte sized items.
|
|
|
|
*/
|
|
|
|
if (memchr_inv(p->vm_stat_diff, 0, NR_VM_ZONE_STAT_ITEMS))
|
|
|
|
return true;
|
|
|
|
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2016-02-05 23:36:24 +00:00
|
|
|
void quiet_vmstat(void)
|
|
|
|
{
|
|
|
|
if (system_state != SYSTEM_RUNNING)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If we are already in hands of the shepherd then there
|
|
|
|
* is nothing for us to do here.
|
|
|
|
*/
|
|
|
|
if (cpumask_test_and_set_cpu(smp_processor_id(), cpu_stat_off))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (!need_update(smp_processor_id()))
|
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Just refresh counters and do not care about the pending delayed
|
|
|
|
* vmstat_update. It doesn't fire that often to matter and canceling
|
|
|
|
* it would be too expensive from this path.
|
|
|
|
* vmstat_shepherd will take care about that for us.
|
|
|
|
*/
|
|
|
|
refresh_cpu_vm_stats(false);
|
|
|
|
}
|
|
|
|
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Shepherd worker thread that checks the
|
|
|
|
* differentials of processors that have their worker
|
|
|
|
* threads for vm statistics updates disabled because of
|
|
|
|
* inactivity.
|
|
|
|
*/
|
|
|
|
static void vmstat_shepherd(struct work_struct *w);
|
|
|
|
|
2016-01-14 23:21:40 +00:00
|
|
|
static DECLARE_DEFERRABLE_WORK(shepherd, vmstat_shepherd);
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
|
|
|
|
static void vmstat_shepherd(struct work_struct *w)
|
|
|
|
{
|
|
|
|
int cpu;
|
|
|
|
|
|
|
|
get_online_cpus();
|
|
|
|
/* Check processors whose vmstat worker threads have been disabled */
|
2016-02-05 23:36:24 +00:00
|
|
|
for_each_cpu(cpu, cpu_stat_off) {
|
|
|
|
struct delayed_work *dw = &per_cpu(vmstat_work, cpu);
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
|
2016-02-05 23:36:24 +00:00
|
|
|
if (need_update(cpu)) {
|
|
|
|
if (cpumask_test_and_clear_cpu(cpu, cpu_stat_off))
|
|
|
|
queue_delayed_work_on(cpu, vmstat_wq, dw, 0);
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* Cancel the work if quiet_vmstat has put this
|
|
|
|
* cpu on cpu_stat_off because the work item might
|
|
|
|
* be still scheduled
|
|
|
|
*/
|
|
|
|
cancel_delayed_work(dw);
|
|
|
|
}
|
|
|
|
}
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
put_online_cpus();
|
|
|
|
|
|
|
|
schedule_delayed_work(&shepherd,
|
2009-04-02 23:56:39 +00:00
|
|
|
round_jiffies_relative(sysctl_stat_interval));
|
2007-05-09 09:35:12 +00:00
|
|
|
}
|
|
|
|
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
static void __init start_shepherd_timer(void)
|
2007-05-09 09:35:12 +00:00
|
|
|
{
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
int cpu;
|
|
|
|
|
|
|
|
for_each_possible_cpu(cpu)
|
2016-02-05 23:36:27 +00:00
|
|
|
INIT_DEFERRABLE_WORK(per_cpu_ptr(&vmstat_work, cpu),
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
vmstat_update);
|
|
|
|
|
|
|
|
if (!alloc_cpumask_var(&cpu_stat_off, GFP_KERNEL))
|
|
|
|
BUG();
|
|
|
|
cpumask_copy(cpu_stat_off, cpu_online_mask);
|
2007-05-09 09:35:12 +00:00
|
|
|
|
2016-01-08 10:18:29 +00:00
|
|
|
vmstat_wq = alloc_workqueue("vmstat", WQ_FREEZABLE|WQ_MEM_RECLAIM, 0);
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
schedule_delayed_work(&shepherd,
|
|
|
|
round_jiffies_relative(sysctl_stat_interval));
|
2007-05-09 09:35:12 +00:00
|
|
|
}
|
|
|
|
|
2013-11-12 23:08:13 +00:00
|
|
|
static void vmstat_cpu_dead(int node)
|
|
|
|
{
|
|
|
|
int cpu;
|
|
|
|
|
|
|
|
get_online_cpus();
|
|
|
|
for_each_online_cpu(cpu)
|
|
|
|
if (cpu_to_node(cpu) == node)
|
|
|
|
goto end;
|
|
|
|
|
|
|
|
node_clear_state(node, N_CPU);
|
|
|
|
end:
|
|
|
|
put_online_cpus();
|
|
|
|
}
|
|
|
|
|
2006-09-01 04:27:35 +00:00
|
|
|
/*
|
|
|
|
* Use the cpu notifier to insure that the thresholds are recalculated
|
|
|
|
* when necessary.
|
|
|
|
*/
|
2013-06-19 18:53:51 +00:00
|
|
|
static int vmstat_cpuup_callback(struct notifier_block *nfb,
|
2006-09-01 04:27:35 +00:00
|
|
|
unsigned long action,
|
|
|
|
void *hcpu)
|
|
|
|
{
|
2007-05-09 09:35:12 +00:00
|
|
|
long cpu = (long)hcpu;
|
|
|
|
|
2006-09-01 04:27:35 +00:00
|
|
|
switch (action) {
|
2007-05-09 09:35:12 +00:00
|
|
|
case CPU_ONLINE:
|
|
|
|
case CPU_ONLINE_FROZEN:
|
2010-09-09 23:38:14 +00:00
|
|
|
refresh_zone_stat_thresholds();
|
2010-01-05 06:34:51 +00:00
|
|
|
node_set_state(cpu_to_node(cpu), N_CPU);
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
cpumask_set_cpu(cpu, cpu_stat_off);
|
2007-05-09 09:35:12 +00:00
|
|
|
break;
|
|
|
|
case CPU_DOWN_PREPARE:
|
|
|
|
case CPU_DOWN_PREPARE_FROZEN:
|
2010-12-14 15:21:17 +00:00
|
|
|
cancel_delayed_work_sync(&per_cpu(vmstat_work, cpu));
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
cpumask_clear_cpu(cpu, cpu_stat_off);
|
2007-05-09 09:35:12 +00:00
|
|
|
break;
|
|
|
|
case CPU_DOWN_FAILED:
|
|
|
|
case CPU_DOWN_FAILED_FROZEN:
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
cpumask_set_cpu(cpu, cpu_stat_off);
|
2007-05-09 09:35:12 +00:00
|
|
|
break;
|
2006-12-07 04:33:08 +00:00
|
|
|
case CPU_DEAD:
|
2007-05-09 09:35:10 +00:00
|
|
|
case CPU_DEAD_FROZEN:
|
2006-12-07 04:33:08 +00:00
|
|
|
refresh_zone_stat_thresholds();
|
2013-11-12 23:08:13 +00:00
|
|
|
vmstat_cpu_dead(cpu_to_node(cpu));
|
2006-12-07 04:33:08 +00:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
2006-09-01 04:27:35 +00:00
|
|
|
}
|
|
|
|
return NOTIFY_OK;
|
|
|
|
}
|
|
|
|
|
2013-06-19 18:53:51 +00:00
|
|
|
static struct notifier_block vmstat_notifier =
|
2006-09-01 04:27:35 +00:00
|
|
|
{ &vmstat_cpuup_callback, NULL, 0 };
|
2008-10-06 00:13:52 +00:00
|
|
|
#endif
|
2006-09-01 04:27:35 +00:00
|
|
|
|
2007-10-16 08:26:27 +00:00
|
|
|
static int __init setup_vmstat(void)
|
2006-09-01 04:27:35 +00:00
|
|
|
{
|
2008-10-06 00:13:52 +00:00
|
|
|
#ifdef CONFIG_SMP
|
2014-03-10 20:42:27 +00:00
|
|
|
cpu_notifier_register_begin();
|
|
|
|
__register_cpu_notifier(&vmstat_notifier);
|
2007-05-09 09:35:12 +00:00
|
|
|
|
vmstat: on-demand vmstat workers V8
vmstat workers are used for folding counter differentials into the zone,
per node and global counters at certain time intervals. They currently
run at defined intervals on all processors which will cause some holdoff
for processors that need minimal intrusion by the OS.
The current vmstat_update mechanism depends on a deferrable timer firing
every other second by default which registers a work queue item that runs
on the local CPU, with the result that we have 1 interrupt and one
additional schedulable task on each CPU every 2 seconds If a workload
indeed causes VM activity or multiple tasks are running on a CPU, then
there are probably bigger issues to deal with.
However, some workloads dedicate a CPU for a single CPU bound task. This
is done in high performance computing, in high frequency financial
applications, in networking (Intel DPDK, EZchip NPS) and with the advent
of systems with more and more CPUs over time, this may become more and
more common to do since when one has enough CPUs one cares less about
efficiently sharing a CPU with other tasks and more about efficiently
monopolizing a CPU per task.
The difference of having this timer firing and workqueue kernel thread
scheduled per second can be enormous. An artificial test measuring the
worst case time to do a simple "i++" in an endless loop on a bare metal
system and under Linux on an isolated CPU with dynticks and with and
without this patch, have Linux match the bare metal performance (~700
cycles) with this patch and loose by couple of orders of magnitude (~200k
cycles) without it[*]. The loss occurs for something that just calculates
statistics. For networking applications, for example, this could be the
difference between dropping packets or sustaining line rate.
Statistics are important and useful, but it would be great if there would
be a way to not cause statistics gathering produce a huge performance
difference. This patche does just that.
This patch creates a vmstat shepherd worker that monitors the per cpu
differentials on all processors. If there are differentials on a
processor then a vmstat worker local to the processors with the
differentials is created. That worker will then start folding the diffs
in regular intervals. Should the worker find that there is no work to be
done then it will make the shepherd worker monitor the differentials
again.
With this patch it is possible then to have periods longer than
2 seconds without any OS event on a "cpu" (hardware thread).
The patch shows a very minor increased in system performance.
hackbench -s 512 -l 2000 -g 15 -f 25 -P
Results before the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.992
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.971
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 5.063
Hackbench after the patch:
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.973
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.990
Running in process mode with 15 groups using 50 file descriptors each (== 750 tasks)
Each sender will pass 2000 messages of 512 bytes
Time: 4.993
[fengguang.wu@intel.com: cpu_stat_off can be static]
Signed-off-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tejun Heo <tj@kernel.org>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Mike Frysinger <vapier@gentoo.org>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: Hakan Akkan <hakanakkan@gmail.com>
Cc: Max Krasnyansky <maxk@qti.qualcomm.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:29:43 +00:00
|
|
|
start_shepherd_timer();
|
2014-03-10 20:42:27 +00:00
|
|
|
cpu_notifier_register_done();
|
2008-10-06 00:13:52 +00:00
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_PROC_FS
|
|
|
|
proc_create("buddyinfo", S_IRUGO, NULL, &fragmentation_file_operations);
|
2008-10-06 00:15:36 +00:00
|
|
|
proc_create("pagetypeinfo", S_IRUGO, NULL, &pagetypeinfo_file_ops);
|
2008-10-06 00:17:48 +00:00
|
|
|
proc_create("vmstat", S_IRUGO, NULL, &proc_vmstat_file_operations);
|
2008-10-06 00:19:42 +00:00
|
|
|
proc_create("zoneinfo", S_IRUGO, NULL, &proc_zoneinfo_file_operations);
|
2008-10-06 00:13:52 +00:00
|
|
|
#endif
|
2006-09-01 04:27:35 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
module_init(setup_vmstat)
|
2010-05-24 21:32:25 +00:00
|
|
|
|
|
|
|
#if defined(CONFIG_DEBUG_FS) && defined(CONFIG_COMPACTION)
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Return an index indicating how much of the available free memory is
|
|
|
|
* unusable for an allocation of the requested size.
|
|
|
|
*/
|
|
|
|
static int unusable_free_index(unsigned int order,
|
|
|
|
struct contig_page_info *info)
|
|
|
|
{
|
|
|
|
/* No free memory is interpreted as all free memory is unusable */
|
|
|
|
if (info->free_pages == 0)
|
|
|
|
return 1000;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Index should be a value between 0 and 1. Return a value to 3
|
|
|
|
* decimal places.
|
|
|
|
*
|
|
|
|
* 0 => no fragmentation
|
|
|
|
* 1 => high fragmentation
|
|
|
|
*/
|
|
|
|
return div_u64((info->free_pages - (info->free_blocks_suitable << order)) * 1000ULL, info->free_pages);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
static void unusable_show_print(struct seq_file *m,
|
|
|
|
pg_data_t *pgdat, struct zone *zone)
|
|
|
|
{
|
|
|
|
unsigned int order;
|
|
|
|
int index;
|
|
|
|
struct contig_page_info info;
|
|
|
|
|
|
|
|
seq_printf(m, "Node %d, zone %8s ",
|
|
|
|
pgdat->node_id,
|
|
|
|
zone->name);
|
|
|
|
for (order = 0; order < MAX_ORDER; ++order) {
|
|
|
|
fill_contig_page_info(zone, order, &info);
|
|
|
|
index = unusable_free_index(order, &info);
|
|
|
|
seq_printf(m, "%d.%03d ", index / 1000, index % 1000);
|
|
|
|
}
|
|
|
|
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Display unusable free space index
|
|
|
|
*
|
|
|
|
* The unusable free space index measures how much of the available free
|
|
|
|
* memory cannot be used to satisfy an allocation of a given size and is a
|
|
|
|
* value between 0 and 1. The higher the value, the more of free memory is
|
|
|
|
* unusable and by implication, the worse the external fragmentation is. This
|
|
|
|
* can be expressed as a percentage by multiplying by 100.
|
|
|
|
*/
|
|
|
|
static int unusable_show(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
|
|
|
|
/* check memoryless node */
|
2012-12-12 21:51:37 +00:00
|
|
|
if (!node_state(pgdat->node_id, N_MEMORY))
|
2010-05-24 21:32:25 +00:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
walk_zones_in_node(m, pgdat, unusable_show_print);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct seq_operations unusable_op = {
|
|
|
|
.start = frag_start,
|
|
|
|
.next = frag_next,
|
|
|
|
.stop = frag_stop,
|
|
|
|
.show = unusable_show,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int unusable_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return seq_open(file, &unusable_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations unusable_file_ops = {
|
|
|
|
.open = unusable_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.release = seq_release,
|
|
|
|
};
|
|
|
|
|
2010-05-24 21:32:26 +00:00
|
|
|
static void extfrag_show_print(struct seq_file *m,
|
|
|
|
pg_data_t *pgdat, struct zone *zone)
|
|
|
|
{
|
|
|
|
unsigned int order;
|
|
|
|
int index;
|
|
|
|
|
|
|
|
/* Alloc on stack as interrupts are disabled for zone walk */
|
|
|
|
struct contig_page_info info;
|
|
|
|
|
|
|
|
seq_printf(m, "Node %d, zone %8s ",
|
|
|
|
pgdat->node_id,
|
|
|
|
zone->name);
|
|
|
|
for (order = 0; order < MAX_ORDER; ++order) {
|
|
|
|
fill_contig_page_info(zone, order, &info);
|
2010-05-24 21:32:30 +00:00
|
|
|
index = __fragmentation_index(order, &info);
|
2010-05-24 21:32:26 +00:00
|
|
|
seq_printf(m, "%d.%03d ", index / 1000, index % 1000);
|
|
|
|
}
|
|
|
|
|
|
|
|
seq_putc(m, '\n');
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Display fragmentation index for orders that allocations would fail for
|
|
|
|
*/
|
|
|
|
static int extfrag_show(struct seq_file *m, void *arg)
|
|
|
|
{
|
|
|
|
pg_data_t *pgdat = (pg_data_t *)arg;
|
|
|
|
|
|
|
|
walk_zones_in_node(m, pgdat, extfrag_show_print);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct seq_operations extfrag_op = {
|
|
|
|
.start = frag_start,
|
|
|
|
.next = frag_next,
|
|
|
|
.stop = frag_stop,
|
|
|
|
.show = extfrag_show,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int extfrag_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return seq_open(file, &extfrag_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations extfrag_file_ops = {
|
|
|
|
.open = extfrag_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
|
|
|
.release = seq_release,
|
|
|
|
};
|
|
|
|
|
2010-05-24 21:32:25 +00:00
|
|
|
static int __init extfrag_debug_init(void)
|
|
|
|
{
|
2012-05-29 22:06:22 +00:00
|
|
|
struct dentry *extfrag_debug_root;
|
|
|
|
|
2010-05-24 21:32:25 +00:00
|
|
|
extfrag_debug_root = debugfs_create_dir("extfrag", NULL);
|
|
|
|
if (!extfrag_debug_root)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
if (!debugfs_create_file("unusable_index", 0444,
|
|
|
|
extfrag_debug_root, NULL, &unusable_file_ops))
|
2012-05-29 22:06:22 +00:00
|
|
|
goto fail;
|
2010-05-24 21:32:25 +00:00
|
|
|
|
2010-05-24 21:32:26 +00:00
|
|
|
if (!debugfs_create_file("extfrag_index", 0444,
|
|
|
|
extfrag_debug_root, NULL, &extfrag_file_ops))
|
2012-05-29 22:06:22 +00:00
|
|
|
goto fail;
|
2010-05-24 21:32:26 +00:00
|
|
|
|
2010-05-24 21:32:25 +00:00
|
|
|
return 0;
|
2012-05-29 22:06:22 +00:00
|
|
|
fail:
|
|
|
|
debugfs_remove_recursive(extfrag_debug_root);
|
|
|
|
return -ENOMEM;
|
2010-05-24 21:32:25 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
module_init(extfrag_debug_init);
|
|
|
|
#endif
|