Date: Tue, 04 Nov 2014 17:48:59 +0000 From: Steven Hartland <killing@multiplay.co.uk> To: George Kola <georgekola@gmail.com> Cc: freebsd-stable@freebsd.org Subject: Re: ARC size limit Message-ID: <5459118B.7090904@multiplay.co.uk> In-Reply-To: <F6BFA688-9D13-4DD9-8B8A-22F4ACFD7622@gmail.com> References: <3B70EA0C-0976-49D6-8418-6B5D22ED7E65@gmail.com> <54589722.3080803@multiplay.co.uk> <F6BFA688-9D13-4DD9-8B8A-22F4ACFD7622@gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------040704030000030500010906
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Try the attached, its the patch we'll be using when we roll out 10.1.
1. Apply with: cd /usr/src && patch < zfs-arc-refactor.patch
2. Rebuild your kernel
3. Install the new kernel
4. Reboot
Regards
Steve
On 04/11/2014 16:48, George Kola wrote:
> Thanks Steven. We are running 10.1rc3. Is the best way to get this patch to just compile 10-stable kernel ? We are new to running FreeBSD and hence the newbie question.
>
>
> Thanks,
> George
>
>
>
>> On Nov 4, 2014, at 1:06 AM, Steven Hartland <killing@multiplay.co.uk> wrote:
>>
>> You need https://svnweb.freebsd.org/base?view=revision&revision=272875
>> On 04/11/2014 06:29, George Kola wrote:
>>> Hi All,
>>> This is my first post to freebsd-stable fresh of Meet BSD California 2014. We are switching our entire production to FreeBSD. Our storage servers have 256 GB of RAM , 4 TB of SSD and 40 TB of spinning disks. We are running ZFS root and the SSD is configured as L2ARC. We are running FreeBSD 10.1 RC3
>>> I am finding that on all our machines, ARC is somehow limited to < 64 GB of memory and we have a huge inactive memory (180 G). The surprising thing is that ARC seems to have almost the same limit (< 64 GB) on all of our storage boxes and ARC is not growing even though L2ARC hit shows that there is advantage in growing ARC.
>>> Any help/pointers is appreciated.
>>> What I am trying to do is to tune ZFS for our workload. We are hoping that we get a high hit rate.
>>> Thanks to Justin Gibbs and Allan Jude for initial pointers and help. They suggested posting to the mailing list to get further help.
>>>
>>> I have pasted top output and zfs-stats output below and yes UMA is enabled.
>>>
>>> Thanks,
>>> George
>>>
>>>
>>> top
>>> last pid: 27458; load averages: 3.30, 5.42, 5.34 up 6+09:59:30 05:38:49
>>> 71 processes: 1 running, 70 sleeping
>>> CPU: 4.2% user, 0.0% nice, 4.6% system, 0.2% interrupt, 90.9% idle
>>> Mem: 11G Active, 181G Inact, 52G Wired, 1368M Cache, 4266M Free
>>> ARC: 47G Total, 1555M MFU, 41G MRU, 35M Anon, 3984M Header, 709M Other
>>> Swap: 64G Total, 2874M Used, 61G Free, 4% Inuse
>>>
>>>
>>>
>>> sysctl vfs.zfs.zio.use_uma
>>> vfs.zfs.zio.use_uma: 1
>>>
>>>
>>>
>>>
>>> zfs-mon -a output
>>>
>>> ZFS real-time cache activity monitor
>>> Seconds elapsed: 62
>>>
>>> Cache hits and misses:
>>> 1s 10s 60s tot
>>> ARC hits: 124 126 103 101
>>> ARC misses: 35 46 29 28
>>> ARC demand data hits: 55 90 61 61
>>> ARC demand data misses: 20 32 18 17
>>> ARC demand metadata hits: 69 36 42 40
>>> ARC demand metadata misses: 9 13 10 9
>>> ARC prefetch data hits: 0 0 0 0
>>> ARC prefetch data misses: 6 1 1 1
>>> ARC prefetch metadata hits: 0 0 0 0
>>> ARC prefetch metadata misses: 0 0 0 0
>>> L2ARC hits: 16 28 14 14
>>> L2ARC misses: 19 18 15 14
>>> ZFETCH hits: 592 2842 2098 2047
>>> ZFETCH misses: 308 1326 507 494
>>>
>>> Cache efficiency percentage:
>>> 10s 60s tot
>>> ARC: 73.26 78.03 78.29
>>> ARC demand data: 73.77 77.22 78.21
>>> ARC demand metadata: 73.47 80.77 81.63
>>> ARC prefetch data: 0.00 0.00 0.00
>>> ARC prefetch metadata: 0.00 0.00 0.00
>>> L2ARC: 60.87 48.28 50.00
>>> ZFETCH: 68.19 80.54 80.56
>>>
>>>
>>>
>>>
>>> zfs-stats -a output
>>>
>>> ZFS real-time cache activity monitor
>>> Seconds elapsed: 62
>>>
>>> Cache hits and misses:
>>> 1s 10s 60s tot
>>> ARC hits: 124 126 103 101
>>> ARC misses: 35 46 29 28
>>> ARC demand data hits: 55 90 61 61
>>> ARC demand data misses: 20 32 18 17
>>> ARC demand metadata hits: 69 36 42 40
>>> ARC demand metadata misses: 9 13 10 9
>>> ARC prefetch data hits: 0 0 0 0
>>> ARC prefetch data misses: 6 1 1 1
>>> ARC prefetch metadata hits: 0 0 0 0
>>> ARC prefetch metadata misses: 0 0 0 0
>>> L2ARC hits: 16 28 14 14
>>> L2ARC misses: 19 18 15 14
>>> ZFETCH hits: 592 2842 2098 2047
>>> ZFETCH misses: 308 1326 507 494
>>>
>>> Cache efficiency percentage:
>>> 10s 60s tot
>>> ARC: 73.26 78.03 78.29
>>> ARC demand data: 73.77 77.22 78.21
>>> ARC demand metadata: 73.47 80.77 81.63
>>> ARC prefetch data: 0.00 0.00 0.00
>>> ARC prefetch metadata: 0.00 0.00 0.00
>>> L2ARC: 60.87 48.28 50.00
>>> ZFETCH: 68.19 80.54 80.56
>>>
>>>
>>>
>>> _______________________________________________
>>> freebsd-stable@freebsd.org mailing list
>>> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
>>> To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"
>> _______________________________________________
>> freebsd-stable@freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
>> To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"
--------------040704030000030500010906
Content-Type: text/plain; charset=windows-1252;
name="zfs-arc-refactor.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename="zfs-arc-refactor.patch"
Index: sys/cddl/compat/opensolaris/kern/opensolaris_kmem.c
===================================================================
--- sys/cddl/compat/opensolaris/kern/opensolaris_kmem.c (revision 274056)
+++ sys/cddl/compat/opensolaris/kern/opensolaris_kmem.c (working copy)
@@ -133,13 +133,6 @@ kmem_size(void)
return (kmem_size_val);
}
-uint64_t
-kmem_used(void)
-{
-
- return (vmem_size(kmem_arena, VMEM_ALLOC));
-}
-
static int
kmem_std_constructor(void *mem, int size __unused, void *private, int flags)
{
Index: sys/cddl/compat/opensolaris/sys/kmem.h
===================================================================
--- sys/cddl/compat/opensolaris/sys/kmem.h (revision 274056)
+++ sys/cddl/compat/opensolaris/sys/kmem.h (working copy)
@@ -66,7 +66,6 @@ typedef struct kmem_cache {
void *zfs_kmem_alloc(size_t size, int kmflags);
void zfs_kmem_free(void *buf, size_t size);
uint64_t kmem_size(void);
-uint64_t kmem_used(void);
kmem_cache_t *kmem_cache_create(char *name, size_t bufsize, size_t align,
int (*constructor)(void *, void *, int), void (*destructor)(void *, void *),
void (*reclaim)(void *) __unused, void *private, vmem_t *vmp, int cflags);
@@ -78,6 +77,9 @@ void kmem_reap(void);
int kmem_debugging(void);
void *calloc(size_t n, size_t s);
+#define freemem (cnt.v_free_count + cnt.v_cache_count)
+#define minfree cnt.v_free_min
+#define heap_arena kmem_arena
#define kmem_alloc(size, kmflags) zfs_kmem_alloc((size), (kmflags))
#define kmem_zalloc(size, kmflags) zfs_kmem_alloc((size), (kmflags) | M_ZERO)
#define kmem_free(buf, size) zfs_kmem_free((buf), (size))
Index: sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
===================================================================
--- sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c (revision 274056)
+++ sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c (working copy)
@@ -138,6 +138,7 @@
#include <sys/sdt.h>
#include <vm/vm_pageout.h>
+#include <machine/vmparam.h>
#ifdef illumos
#ifndef _KERNEL
@@ -193,9 +194,6 @@ extern int zfs_prefetch_disable;
*/
static boolean_t arc_warm;
-/*
- * These tunables are for performance analysis.
- */
uint64_t zfs_arc_max;
uint64_t zfs_arc_min;
uint64_t zfs_arc_meta_limit = 0;
@@ -204,7 +202,20 @@ int zfs_arc_shrink_shift = 0;
int zfs_arc_p_min_shift = 0;
int zfs_disable_dup_eviction = 0;
uint64_t zfs_arc_average_blocksize = 8 * 1024; /* 8KB */
+u_int zfs_arc_free_target = 0;
+static int sysctl_vfs_zfs_arc_free_target(SYSCTL_HANDLER_ARGS);
+
+#ifdef _KERNEL
+static void
+arc_free_target_init(void *unused __unused)
+{
+
+ zfs_arc_free_target = vm_pageout_wakeup_thresh;
+}
+SYSINIT(arc_free_target_init, SI_SUB_KTHREAD_PAGE, SI_ORDER_ANY,
+ arc_free_target_init, NULL);
+
TUNABLE_QUAD("vfs.zfs.arc_max", &zfs_arc_max);
TUNABLE_QUAD("vfs.zfs.arc_min", &zfs_arc_min);
TUNABLE_QUAD("vfs.zfs.arc_meta_limit", &zfs_arc_meta_limit);
@@ -217,7 +228,37 @@ SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, arc_min, CTLFLAG_
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, arc_average_blocksize, CTLFLAG_RDTUN,
&zfs_arc_average_blocksize, 0,
"ARC average blocksize");
+/*
+ * We don't have a tunable for arc_free_target due to the dependency on
+ * pagedaemon initialisation.
+ */
+SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_free_target,
+ CTLTYPE_UINT | CTLFLAG_MPSAFE | CTLFLAG_RW, 0, sizeof(u_int),
+ sysctl_vfs_zfs_arc_free_target, "IU",
+ "Desired number of free pages below which ARC triggers reclaim");
+static int
+sysctl_vfs_zfs_arc_free_target(SYSCTL_HANDLER_ARGS)
+{
+ u_int val;
+ int err;
+
+ val = zfs_arc_free_target;
+ err = sysctl_handle_int(oidp, &val, 0, req);
+ if (err != 0 || req->newptr == NULL)
+ return (err);
+
+ if (val < minfree)
+ return (EINVAL);
+ if (val > cnt.v_page_count)
+ return (EINVAL);
+
+ zfs_arc_free_target = val;
+
+ return (0);
+}
+#endif
+
/*
* Note that buffers can be in one of 6 states:
* ARC_anon - anonymous (discussed below)
@@ -2421,9 +2462,12 @@ arc_flush(spa_t *spa)
void
arc_shrink(void)
{
+
if (arc_c > arc_c_min) {
uint64_t to_free;
+ DTRACE_PROBE4(arc__shrink, uint64_t, arc_c, uint64_t,
+ arc_c_min, uint64_t, arc_p, uint64_t, to_free);
#ifdef _KERNEL
to_free = arc_c >> arc_shrink_shift;
#else
@@ -2439,12 +2483,19 @@ arc_shrink(void)
arc_c = MAX(arc_size, arc_c_min);
if (arc_p > arc_c)
arc_p = (arc_c >> 1);
+
+ DTRACE_PROBE2(arc__shrunk, uint64_t, arc_c, uint64_t,
+ arc_p);
+
ASSERT(arc_c >= arc_c_min);
ASSERT((int64_t)arc_p >= 0);
}
- if (arc_size > arc_c)
+ if (arc_size > arc_c) {
+ DTRACE_PROBE2(arc__shrink_adjust, uint64_t, arc_size,
+ uint64_t, arc_c);
arc_adjust();
+ }
}
static int needfree = 0;
@@ -2455,15 +2506,20 @@ arc_reclaim_needed(void)
#ifdef _KERNEL
- if (needfree)
+ if (needfree) {
+ DTRACE_PROBE(arc__reclaim_needfree);
return (1);
+ }
/*
* Cooperate with pagedaemon when it's time for it to scan
* and reclaim some pages.
*/
- if (vm_paging_needed())
+ if (freemem < zfs_arc_free_target) {
+ DTRACE_PROBE2(arc__reclaim_freemem, uint64_t,
+ freemem, uint64_t, zfs_arc_free_target);
return (1);
+ }
#ifdef sun
/*
@@ -2491,8 +2547,19 @@ arc_reclaim_needed(void)
if (availrmem < swapfs_minfree + swapfs_reserve + extra)
return (1);
-#if defined(__i386)
/*
+ * Check that we have enough availrmem that memory locking (e.g., via
+ * mlock(3C) or memcntl(2)) can still succeed. (pages_pp_maximum
+ * stores the number of pages that cannot be locked; when availrmem
+ * drops below pages_pp_maximum, page locking mechanisms such as
+ * page_pp_lock() will fail.)
+ */
+ if (availrmem <= pages_pp_maximum)
+ return (1);
+
+#endif /* sun */
+#if defined(__i386) || !defined(UMA_MD_SMALL_ALLOC)
+ /*
* If we're on an i386 platform, it's possible that we'll exhaust the
* kernel heap space before we ever run out of available physical
* memory. Most checks of the size of the heap_area compare against
@@ -2503,19 +2570,35 @@ arc_reclaim_needed(void)
* heap is allocated. (Or, in the calculation, if less than 1/4th is
* free)
*/
- if (btop(vmem_size(heap_arena, VMEM_FREE)) <
- (btop(vmem_size(heap_arena, VMEM_FREE | VMEM_ALLOC)) >> 2))
+ if (vmem_size(heap_arena, VMEM_FREE) <
+ (vmem_size(heap_arena, VMEM_FREE | VMEM_ALLOC) >> 2)) {
+ DTRACE_PROBE2(arc__reclaim_used, uint64_t,
+ vmem_size(heap_arena, VMEM_FREE), uint64_t,
+ (vmem_size(heap_arena, VMEM_FREE | VMEM_ALLOC)) >> 2);
return (1);
+ }
#endif
-#else /* !sun */
- if (kmem_used() > (kmem_size() * 3) / 4)
+#ifdef sun
+ /*
+ * If zio data pages are being allocated out of a separate heap segment,
+ * then enforce that the size of available vmem for this arena remains
+ * above about 1/16th free.
+ *
+ * Note: The 1/16th arena free requirement was put in place
+ * to aggressively evict memory from the arc in order to avoid
+ * memory fragmentation issues.
+ */
+ if (zio_arena != NULL &&
+ vmem_size(zio_arena, VMEM_FREE) <
+ (vmem_size(zio_arena, VMEM_ALLOC) >> 4))
return (1);
#endif /* sun */
-
-#else
+#else /* _KERNEL */
if (spa_get_random(100) == 0)
return (1);
-#endif
+#endif /* _KERNEL */
+ DTRACE_PROBE(arc__reclaim_no);
+
return (0);
}
@@ -2522,7 +2605,7 @@ arc_reclaim_needed(void)
extern kmem_cache_t *zio_buf_cache[];
extern kmem_cache_t *zio_data_buf_cache[];
-static void
+static void __noinline
arc_kmem_reap_now(arc_reclaim_strategy_t strat)
{
size_t i;
@@ -2529,6 +2612,7 @@ arc_kmem_reap_now(arc_reclaim_strategy_t strat)
kmem_cache_t *prev_cache = NULL;
kmem_cache_t *prev_data_cache = NULL;
+ DTRACE_PROBE(arc__kmem_reap_start);
#ifdef _KERNEL
if (arc_meta_used >= arc_meta_limit) {
/*
@@ -2564,6 +2648,16 @@ arc_kmem_reap_now(arc_reclaim_strategy_t strat)
}
kmem_cache_reap_now(buf_cache);
kmem_cache_reap_now(hdr_cache);
+
+#ifdef sun
+ /*
+ * Ask the vmem arena to reclaim unused memory from its
+ * quantum caches.
+ */
+ if (zio_arena != NULL && strat == ARC_RECLAIM_AGGR)
+ vmem_qcache_reap(zio_arena);
+#endif
+ DTRACE_PROBE(arc__kmem_reap_end);
}
static void
@@ -2581,6 +2675,7 @@ arc_reclaim_thread(void *dummy __unused)
if (arc_no_grow) {
if (last_reclaim == ARC_RECLAIM_CONS) {
+ DTRACE_PROBE(arc__reclaim_aggr_no_grow);
last_reclaim = ARC_RECLAIM_AGGR;
} else {
last_reclaim = ARC_RECLAIM_CONS;
@@ -2588,6 +2683,7 @@ arc_reclaim_thread(void *dummy __unused)
} else {
arc_no_grow = TRUE;
last_reclaim = ARC_RECLAIM_AGGR;
+ DTRACE_PROBE(arc__reclaim_aggr);
membar_producer();
}
@@ -2692,6 +2788,7 @@ arc_adapt(int bytes, arc_state_t *state)
* cache size, increment the target cache size
*/
if (arc_size > arc_c - (2ULL << SPA_MAXBLOCKSHIFT)) {
+ DTRACE_PROBE1(arc__inc_adapt, int, bytes);
atomic_add_64(&arc_c, (int64_t)bytes);
if (arc_c > arc_c_max)
arc_c = arc_c_max;
@@ -2713,20 +2810,6 @@ arc_evict_needed(arc_buf_contents_t type)
if (type == ARC_BUFC_METADATA && arc_meta_used >= arc_meta_limit)
return (1);
-#ifdef sun
-#ifdef _KERNEL
- /*
- * If zio data pages are being allocated out of a separate heap segment,
- * then enforce that the size of available vmem for this area remains
- * above about 1/32nd free.
- */
- if (type == ARC_BUFC_DATA && zio_arena != NULL &&
- vmem_size(zio_arena, VMEM_FREE) <
- (vmem_size(zio_arena, VMEM_ALLOC) >> 5))
- return (1);
-#endif
-#endif /* sun */
-
if (arc_reclaim_needed())
return (1);
@@ -3885,20 +3968,16 @@ static int
arc_memory_throttle(uint64_t reserve, uint64_t txg)
{
#ifdef _KERNEL
- uint64_t available_memory =
- ptoa((uintmax_t)cnt.v_free_count + cnt.v_cache_count);
+ uint64_t available_memory = ptob(freemem);
static uint64_t page_load = 0;
static uint64_t last_txg = 0;
-#ifdef sun
-#if defined(__i386)
+#if defined(__i386) || !defined(UMA_MD_SMALL_ALLOC)
available_memory =
- MIN(available_memory, vmem_size(heap_arena, VMEM_FREE));
+ MIN(available_memory, ptob(vmem_size(heap_arena, VMEM_FREE)));
#endif
-#endif /* sun */
- if (cnt.v_free_count + cnt.v_cache_count >
- (uint64_t)physmem * arc_lotsfree_percent / 100)
+ if (freemem > (uint64_t)physmem * arc_lotsfree_percent / 100)
return (0);
if (txg > last_txg) {
@@ -3911,7 +3990,7 @@ arc_memory_throttle(uint64_t reserve, uint64_t txg
* continue to let page writes occur as quickly as possible.
*/
if (curproc == pageproc) {
- if (page_load > available_memory / 4)
+ if (page_load > MAX(ptob(minfree), available_memory) / 4)
return (SET_ERROR(ERESTART));
/* Note: reserve is inflated, so we deflate */
page_load += reserve / 8;
@@ -3939,8 +4018,10 @@ arc_tempreserve_space(uint64_t reserve, uint64_t t
int error;
uint64_t anon_size;
- if (reserve > arc_c/4 && !arc_no_grow)
+ if (reserve > arc_c/4 && !arc_no_grow) {
arc_c = MIN(arc_c_max, reserve * 4);
+ DTRACE_PROBE1(arc__set_reserve, uint64_t, arc_c);
+ }
if (reserve > arc_c)
return (SET_ERROR(ENOMEM));
@@ -3994,6 +4075,7 @@ arc_lowmem(void *arg __unused, int howto __unused)
mutex_enter(&arc_lowmem_lock);
mutex_enter(&arc_reclaim_thr_lock);
needfree = 1;
+ DTRACE_PROBE(arc__needfree);
cv_signal(&arc_reclaim_thr_cv);
/*
Index: sys/vm/vm_pageout.c
===================================================================
--- sys/vm/vm_pageout.c (revision 274056)
+++ sys/vm/vm_pageout.c (working copy)
@@ -76,6 +76,7 @@
__FBSDID("$FreeBSD$");
#include "opt_vm.h"
+#include "opt_kdtrace.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
@@ -89,6 +90,7 @@ __FBSDID("$FreeBSD$");
#include <sys/racct.h>
#include <sys/resourcevar.h>
#include <sys/sched.h>
+#include <sys/sdt.h>
#include <sys/signalvar.h>
#include <sys/smp.h>
#include <sys/vnode.h>
@@ -115,10 +117,14 @@ __FBSDID("$FreeBSD$");
/* the kernel process "vm_pageout"*/
static void vm_pageout(void);
+static void vm_pageout_init(void);
static int vm_pageout_clean(vm_page_t);
static void vm_pageout_scan(struct vm_domain *vmd, int pass);
static void vm_pageout_mightbe_oom(struct vm_domain *vmd, int pass);
+SYSINIT(pagedaemon_init, SI_SUB_KTHREAD_PAGE, SI_ORDER_FIRST, vm_pageout_init,
+ NULL);
+
struct proc *pageproc;
static struct kproc_desc page_kp = {
@@ -126,9 +132,13 @@ static struct kproc_desc page_kp = {
vm_pageout,
&pageproc
};
-SYSINIT(pagedaemon, SI_SUB_KTHREAD_PAGE, SI_ORDER_FIRST, kproc_start,
+SYSINIT(pagedaemon, SI_SUB_KTHREAD_PAGE, SI_ORDER_SECOND, kproc_start,
&page_kp);
+SDT_PROVIDER_DEFINE(vm);
+SDT_PROBE_DEFINE(vm, , , vm__lowmem_cache);
+SDT_PROBE_DEFINE(vm, , , vm__lowmem_scan);
+
#if !defined(NO_SWAPPING)
/* the kernel process "vm_daemon"*/
static void vm_daemon(void);
@@ -663,6 +673,7 @@ vm_pageout_grow_cache(int tries, vm_paddr_t low, v
* may acquire locks and/or sleep, so they can only be invoked
* when "tries" is greater than zero.
*/
+ SDT_PROBE0(vm, , , vm__lowmem_cache);
EVENTHANDLER_INVOKE(vm_lowmem, 0);
/*
@@ -925,6 +936,7 @@ vm_pageout_scan(struct vm_domain *vmd, int pass)
/*
* Decrease registered cache sizes.
*/
+ SDT_PROBE0(vm, , , vm__lowmem_scan);
EVENTHANDLER_INVOKE(vm_lowmem, 0);
/*
* We do this explicitly after the caches have been
@@ -1650,15 +1662,11 @@ vm_pageout_worker(void *arg)
}
/*
- * vm_pageout is the high level pageout daemon.
+ * vm_pageout_init initialises basic pageout daemon settings.
*/
static void
-vm_pageout(void)
+vm_pageout_init(void)
{
-#if MAXMEMDOM > 1
- int error, i;
-#endif
-
/*
* Initialize some paging parameters.
*/
@@ -1704,7 +1712,18 @@ static void
/* XXX does not really belong here */
if (vm_page_max_wired == 0)
vm_page_max_wired = cnt.v_free_count / 3;
+}
+/*
+ * vm_pageout is the high level pageout daemon.
+ */
+static void
+vm_pageout(void)
+{
+#if MAXMEMDOM > 1
+ int error, i;
+#endif
+
swap_pager_swap_init();
#if MAXMEMDOM > 1
for (i = 1; i < vm_ndomains; i++) {
Index: .
===================================================================
--- . (revision 274056)
+++ . (working copy)
Property changes on: .
___________________________________________________________________
Modified: svn:mergeinfo
Merged /head:r270759,270861,272483
Merged /stable/10:r272875
--------------040704030000030500010906--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5459118B.7090904>