2005-01-07 02:29:27 +00:00
|
|
|
/*-
|
1994-05-24 10:09:53 +00:00
|
|
|
* Copyright (c) 1982, 1986, 1989, 1993
|
|
|
|
|
* The Regents of the University of California. All rights reserved.
|
|
|
|
|
*
|
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
|
* are met:
|
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
|
* 4. Neither the name of the University nor the names of its contributors
|
|
|
|
|
* may be used to endorse or promote products derived from this software
|
|
|
|
|
* without specific prior written permission.
|
|
|
|
|
*
|
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
|
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
|
*
|
|
|
|
|
* @(#)vm_meter.c 8.4 (Berkeley) 1/4/94
|
|
|
|
|
*/
|
|
|
|
|
|
2003-06-11 23:50:51 +00:00
|
|
|
#include <sys/cdefs.h>
|
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/param.h>
|
|
|
|
|
#include <sys/systm.h>
|
|
|
|
|
#include <sys/kernel.h>
|
2001-05-01 08:13:21 +00:00
|
|
|
#include <sys/lock.h>
|
|
|
|
|
#include <sys/mutex.h>
|
|
|
|
|
#include <sys/proc.h>
|
1998-03-28 10:33:27 +00:00
|
|
|
#include <sys/resource.h>
|
2001-03-28 11:52:56 +00:00
|
|
|
#include <sys/sx.h>
|
1995-12-07 12:48:31 +00:00
|
|
|
#include <sys/vmmeter.h>
|
2002-04-04 21:38:47 +00:00
|
|
|
#include <sys/smp.h>
|
1995-12-07 12:48:31 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <vm/vm.h>
|
1998-10-31 17:21:31 +00:00
|
|
|
#include <vm/vm_page.h>
|
1995-12-10 14:52:10 +00:00
|
|
|
#include <vm/vm_extern.h>
|
1995-12-07 12:48:31 +00:00
|
|
|
#include <vm/vm_param.h>
|
|
|
|
|
#include <vm/pmap.h>
|
|
|
|
|
#include <vm/vm_map.h>
|
|
|
|
|
#include <vm/vm_object.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/sysctl.h>
|
|
|
|
|
|
1995-07-29 11:44:31 +00:00
|
|
|
struct vmmeter cnt;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2001-05-18 00:08:38 +00:00
|
|
|
int maxslp = MAXSLP;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, VM_V_FREE_MIN, v_free_min,
|
1995-11-14 09:29:34 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_free_min, 0, "");
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, VM_V_FREE_TARGET, v_free_target,
|
1995-11-14 09:29:34 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_free_target, 0, "");
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, VM_V_FREE_RESERVED, v_free_reserved,
|
1995-11-14 09:29:34 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_free_reserved, 0, "");
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, VM_V_INACTIVE_TARGET, v_inactive_target,
|
1995-11-14 09:29:34 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_inactive_target, 0, "");
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, VM_V_CACHE_MIN, v_cache_min,
|
1995-11-14 09:29:34 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_cache_min, 0, "");
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, VM_V_CACHE_MAX, v_cache_max,
|
1995-11-14 09:29:34 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_cache_max, 0, "");
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, VM_V_PAGEOUT_FREE_MIN, v_pageout_free_min,
|
1995-11-14 09:29:34 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_pageout_free_min, 0, "");
|
2000-07-05 07:46:41 +00:00
|
|
|
SYSCTL_UINT(_vm, OID_AUTO, v_free_severe,
|
1999-09-17 04:56:40 +00:00
|
|
|
CTLFLAG_RW, &cnt.v_free_severe, 0, "");
|
1995-11-14 09:29:34 +00:00
|
|
|
|
2004-10-11 22:04:16 +00:00
|
|
|
static int
|
|
|
|
|
sysctl_vm_loadavg(SYSCTL_HANDLER_ARGS)
|
|
|
|
|
{
|
|
|
|
|
#ifdef SCTL_MASK32
|
|
|
|
|
u_int32_t la[4];
|
|
|
|
|
|
|
|
|
|
if (req->flags & SCTL_MASK32) {
|
|
|
|
|
la[0] = averunnable.ldavg[0];
|
|
|
|
|
la[1] = averunnable.ldavg[1];
|
|
|
|
|
la[2] = averunnable.ldavg[2];
|
|
|
|
|
la[3] = averunnable.fscale;
|
|
|
|
|
return SYSCTL_OUT(req, la, sizeof(la));
|
|
|
|
|
} else
|
|
|
|
|
#endif
|
|
|
|
|
return SYSCTL_OUT(req, &averunnable, sizeof(averunnable));
|
|
|
|
|
}
|
|
|
|
|
SYSCTL_PROC(_vm, VM_LOADAVG, loadavg, CTLTYPE_STRUCT|CTLFLAG_RD,
|
|
|
|
|
NULL, 0, sysctl_vm_loadavg, "S,loadavg", "Machine loadaverage history");
|
1995-11-14 09:29:34 +00:00
|
|
|
|
|
|
|
|
static int
|
2000-07-04 11:25:35 +00:00
|
|
|
vmtotal(SYSCTL_HANDLER_ARGS)
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
/* XXXKSE almost completely broken */
|
1995-11-14 09:29:34 +00:00
|
|
|
struct proc *p;
|
2006-11-20 08:33:55 +00:00
|
|
|
struct vmtotal total;
|
1995-11-14 09:29:34 +00:00
|
|
|
vm_map_entry_t entry;
|
|
|
|
|
vm_object_t object;
|
|
|
|
|
vm_map_t map;
|
1994-05-24 10:09:53 +00:00
|
|
|
int paging;
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
struct thread *td;
|
2006-05-29 21:28:56 +00:00
|
|
|
struct vmspace *vm;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2006-11-20 08:33:55 +00:00
|
|
|
bzero(&total, sizeof(total));
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
|
* Mark all objects as inactive.
|
|
|
|
|
*/
|
2001-07-04 16:20:28 +00:00
|
|
|
GIANT_REQUIRED;
|
2002-04-20 07:23:22 +00:00
|
|
|
mtx_lock(&vm_object_list_mtx);
|
2003-01-02 09:09:27 +00:00
|
|
|
TAILQ_FOREACH(object, &vm_object_list, object_list) {
|
2004-01-02 19:38:25 +00:00
|
|
|
if (!VM_OBJECT_TRYLOCK(object)) {
|
|
|
|
|
/*
|
|
|
|
|
* Avoid a lock-order reversal. Consequently,
|
|
|
|
|
* the reported number of active pages may be
|
|
|
|
|
* greater than the actual number.
|
|
|
|
|
*/
|
|
|
|
|
continue;
|
|
|
|
|
}
|
1998-08-24 08:39:39 +00:00
|
|
|
vm_object_clear_flag(object, OBJ_ACTIVE);
|
2003-04-13 19:36:18 +00:00
|
|
|
VM_OBJECT_UNLOCK(object);
|
2003-01-02 09:09:27 +00:00
|
|
|
}
|
2002-04-20 07:23:22 +00:00
|
|
|
mtx_unlock(&vm_object_list_mtx);
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
|
* Calculate process statistics.
|
|
|
|
|
*/
|
2001-03-28 11:52:56 +00:00
|
|
|
sx_slock(&allproc_lock);
|
2001-09-12 08:38:13 +00:00
|
|
|
FOREACH_PROC_IN_SYSTEM(p) {
|
1994-05-24 10:09:53 +00:00
|
|
|
if (p->p_flag & P_SYSTEM)
|
|
|
|
|
continue;
|
Change and clean the mutex lock interface.
mtx_enter(lock, type) becomes:
mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks)
mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized)
similarily, for releasing a lock, we now have:
mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN.
We change the caller interface for the two different types of locks
because the semantics are entirely different for each case, and this
makes it explicitly clear and, at the same time, it rids us of the
extra `type' argument.
The enter->lock and exit->unlock change has been made with the idea
that we're "locking data" and not "entering locked code" in mind.
Further, remove all additional "flags" previously passed to the
lock acquire/release routines with the exception of two:
MTX_QUIET and MTX_NOSWITCH
The functionality of these flags is preserved and they can be passed
to the lock/unlock routines by calling the corresponding wrappers:
mtx_{lock, unlock}_flags(lock, flag(s)) and
mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN
locks, respectively.
Re-inline some lock acq/rel code; in the sleep lock case, we only
inline the _obtain_lock()s in order to ensure that the inlined code
fits into a cache line. In the spin lock case, we inline recursion and
actually only perform a function call if we need to spin. This change
has been made with the idea that we generally tend to avoid spin locks
and that also the spin locks that we do have and are heavily used
(i.e. sched_lock) do recurse, and therefore in an effort to reduce
function call overhead for some architectures (such as alpha), we
inline recursion for this case.
Create a new malloc type for the witness code and retire from using
the M_DEV type. The new type is called M_WITNESS and is only declared
if WITNESS is enabled.
Begin cleaning up some machdep/mutex.h code - specifically updated the
"optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN
and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently
need those.
Finally, caught up to the interface changes in all sys code.
Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
|
|
|
mtx_lock_spin(&sched_lock);
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
switch (p->p_state) {
|
|
|
|
|
case PRS_NEW:
|
Change and clean the mutex lock interface.
mtx_enter(lock, type) becomes:
mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks)
mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized)
similarily, for releasing a lock, we now have:
mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN.
We change the caller interface for the two different types of locks
because the semantics are entirely different for each case, and this
makes it explicitly clear and, at the same time, it rids us of the
extra `type' argument.
The enter->lock and exit->unlock change has been made with the idea
that we're "locking data" and not "entering locked code" in mind.
Further, remove all additional "flags" previously passed to the
lock acquire/release routines with the exception of two:
MTX_QUIET and MTX_NOSWITCH
The functionality of these flags is preserved and they can be passed
to the lock/unlock routines by calling the corresponding wrappers:
mtx_{lock, unlock}_flags(lock, flag(s)) and
mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN
locks, respectively.
Re-inline some lock acq/rel code; in the sleep lock case, we only
inline the _obtain_lock()s in order to ensure that the inlined code
fits into a cache line. In the spin lock case, we inline recursion and
actually only perform a function call if we need to spin. This change
has been made with the idea that we generally tend to avoid spin locks
and that also the spin locks that we do have and are heavily used
(i.e. sched_lock) do recurse, and therefore in an effort to reduce
function call overhead for some architectures (such as alpha), we
inline recursion for this case.
Create a new malloc type for the witness code and retire from using
the M_DEV type. The new type is called M_WITNESS and is only declared
if WITNESS is enabled.
Begin cleaning up some machdep/mutex.h code - specifically updated the
"optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN
and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently
need those.
Finally, caught up to the interface changes in all sys code.
Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
1994-05-24 10:09:53 +00:00
|
|
|
continue;
|
|
|
|
|
break;
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
default:
|
|
|
|
|
FOREACH_THREAD_IN_PROC(p, td) {
|
2002-09-11 08:13:56 +00:00
|
|
|
/* Need new statistics XXX */
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
switch (td->td_state) {
|
2002-09-11 08:13:56 +00:00
|
|
|
case TDS_INHIBITED:
|
2002-10-02 20:31:47 +00:00
|
|
|
if (TD_ON_LOCK(td) ||
|
2002-09-11 08:13:56 +00:00
|
|
|
(td->td_inhibitors ==
|
|
|
|
|
TDI_SWAPPED)) {
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_sw++;
|
2002-09-11 08:13:56 +00:00
|
|
|
} else if (TD_IS_SLEEPING(td) ||
|
|
|
|
|
TD_AWAITING_INTR(td) ||
|
|
|
|
|
TD_IS_SUSPENDED(td)) {
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
if (td->td_priority <= PZERO)
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_dw++;
|
2002-09-11 08:13:56 +00:00
|
|
|
else
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_sl++;
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
}
|
|
|
|
|
break;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2002-09-11 08:13:56 +00:00
|
|
|
case TDS_CAN_RUN:
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_sw++;
|
2002-09-11 08:13:56 +00:00
|
|
|
break;
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
case TDS_RUNQ:
|
|
|
|
|
case TDS_RUNNING:
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_rq++;
|
Part 1 of KSE-III
The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)
Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)
NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..
2002-06-29 17:26:22 +00:00
|
|
|
continue;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
2001-01-24 11:28:36 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
Change and clean the mutex lock interface.
mtx_enter(lock, type) becomes:
mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks)
mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized)
similarily, for releasing a lock, we now have:
mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN.
We change the caller interface for the two different types of locks
because the semantics are entirely different for each case, and this
makes it explicitly clear and, at the same time, it rids us of the
extra `type' argument.
The enter->lock and exit->unlock change has been made with the idea
that we're "locking data" and not "entering locked code" in mind.
Further, remove all additional "flags" previously passed to the
lock acquire/release routines with the exception of two:
MTX_QUIET and MTX_NOSWITCH
The functionality of these flags is preserved and they can be passed
to the lock/unlock routines by calling the corresponding wrappers:
mtx_{lock, unlock}_flags(lock, flag(s)) and
mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN
locks, respectively.
Re-inline some lock acq/rel code; in the sleep lock case, we only
inline the _obtain_lock()s in order to ensure that the inlined code
fits into a cache line. In the spin lock case, we inline recursion and
actually only perform a function call if we need to spin. This change
has been made with the idea that we generally tend to avoid spin locks
and that also the spin locks that we do have and are heavily used
(i.e. sched_lock) do recurse, and therefore in an effort to reduce
function call overhead for some architectures (such as alpha), we
inline recursion for this case.
Create a new malloc type for the witness code and retire from using
the M_DEV type. The new type is called M_WITNESS and is only declared
if WITNESS is enabled.
Begin cleaning up some machdep/mutex.h code - specifically updated the
"optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN
and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently
need those.
Finally, caught up to the interface changes in all sys code.
Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
|
|
|
mtx_unlock_spin(&sched_lock);
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
|
* Note active objects.
|
|
|
|
|
*/
|
|
|
|
|
paging = 0;
|
2006-05-29 21:28:56 +00:00
|
|
|
vm = vmspace_acquire_ref(p);
|
|
|
|
|
if (vm == NULL)
|
|
|
|
|
continue;
|
|
|
|
|
map = &vm->vm_map;
|
2003-01-03 05:52:02 +00:00
|
|
|
vm_map_lock_read(map);
|
|
|
|
|
for (entry = map->header.next;
|
These changes embody the support of the fully coherent merged VM buffer cache,
much higher filesystem I/O performance, and much better paging performance. It
represents the culmination of over 6 months of R&D.
The majority of the merged VM/cache work is by John Dyson.
The following highlights the most significant changes. Additionally, there are
(mostly minor) changes to the various filesystem modules (nfs, msdosfs, etc) to
support the new VM/buffer scheme.
vfs_bio.c:
Significant rewrite of most of vfs_bio to support the merged VM buffer cache
scheme. The scheme is almost fully compatible with the old filesystem
interface. Significant improvement in the number of opportunities for write
clustering.
vfs_cluster.c, vfs_subr.c
Upgrade and performance enhancements in vfs layer code to support merged
VM/buffer cache. Fixup of vfs_cluster to eliminate the bogus pagemove stuff.
vm_object.c:
Yet more improvements in the collapse code. Elimination of some windows that
can cause list corruption.
vm_pageout.c:
Fixed it, it really works better now. Somehow in 2.0, some "enhancements"
broke the code. This code has been reworked from the ground-up.
vm_fault.c, vm_page.c, pmap.c, vm_object.c
Support for small-block filesystems with merged VM/buffer cache scheme.
pmap.c vm_map.c
Dynamic kernel VM size, now we dont have to pre-allocate excessive numbers of
kernel PTs.
vm_glue.c
Much simpler and more effective swapping code. No more gratuitous swapping.
proc.h
Fixed the problem that the p_lock flag was not being cleared on a fork.
swap_pager.c, vnode_pager.c
Removal of old vfs_bio cruft to support the past pseudo-coherency. Now the
code doesn't need it anymore.
machdep.c
Changes to better support the parameter values for the merged VM/buffer cache
scheme.
machdep.c, kern_exec.c, vm_glue.c
Implemented a seperate submap for temporary exec string space and another one
to contain process upages. This eliminates all map fragmentation problems
that previously existed.
ffs_inode.c, ufs_inode.c, ufs_readwrite.c
Changes for merged VM/buffer cache. Add "bypass" support for sneaking in on
busy buffers.
Submitted by: John Dyson and David Greenman
1995-01-09 16:06:02 +00:00
|
|
|
entry != &map->header; entry = entry->next) {
|
1999-02-07 21:48:23 +00:00
|
|
|
if ((entry->eflags & MAP_ENTRY_IS_SUB_MAP) ||
|
2003-01-03 05:52:02 +00:00
|
|
|
(object = entry->object.vm_object) == NULL)
|
1994-05-24 10:09:53 +00:00
|
|
|
continue;
|
2003-04-13 19:36:18 +00:00
|
|
|
VM_OBJECT_LOCK(object);
|
2003-01-03 05:52:02 +00:00
|
|
|
vm_object_set_flag(object, OBJ_ACTIVE);
|
|
|
|
|
paging |= object->paging_in_progress;
|
2003-04-13 19:36:18 +00:00
|
|
|
VM_OBJECT_UNLOCK(object);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2003-01-03 05:52:02 +00:00
|
|
|
vm_map_unlock_read(map);
|
2006-05-29 21:28:56 +00:00
|
|
|
vmspace_free(vm);
|
1994-05-24 10:09:53 +00:00
|
|
|
if (paging)
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_pw++;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2001-03-28 11:52:56 +00:00
|
|
|
sx_sunlock(&allproc_lock);
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
|
* Calculate object memory usage statistics.
|
|
|
|
|
*/
|
2002-04-20 07:23:22 +00:00
|
|
|
mtx_lock(&vm_object_list_mtx);
|
2001-04-15 10:22:04 +00:00
|
|
|
TAILQ_FOREACH(object, &vm_object_list, object_list) {
|
1999-01-21 08:29:12 +00:00
|
|
|
/*
|
2004-01-02 19:38:25 +00:00
|
|
|
* Perform unsynchronized reads on the object to avoid
|
|
|
|
|
* a lock-order reversal. In this case, the lack of
|
|
|
|
|
* synchronization should not impair the accuracy of
|
|
|
|
|
* the reported statistics.
|
1999-01-21 08:29:12 +00:00
|
|
|
*/
|
2003-01-03 05:52:02 +00:00
|
|
|
if (object->type == OBJT_DEVICE) {
|
2004-01-02 19:38:25 +00:00
|
|
|
/*
|
|
|
|
|
* Devices, like /dev/mem, will badly skew our totals.
|
|
|
|
|
*/
|
1999-01-21 08:29:12 +00:00
|
|
|
continue;
|
2003-01-03 05:52:02 +00:00
|
|
|
}
|
2006-11-20 00:16:00 +00:00
|
|
|
if (object->ref_count == 0) {
|
|
|
|
|
/*
|
|
|
|
|
* Also skip unreferenced objects, including
|
|
|
|
|
* vnodes representing mounted file systems.
|
|
|
|
|
*/
|
|
|
|
|
continue;
|
|
|
|
|
}
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_vm += object->size;
|
|
|
|
|
total.t_rm += object->resident_page_count;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (object->flags & OBJ_ACTIVE) {
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_avm += object->size;
|
|
|
|
|
total.t_arm += object->resident_page_count;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1996-09-08 20:44:49 +00:00
|
|
|
if (object->shadow_count > 1) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/* shared object */
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_vmshr += object->size;
|
|
|
|
|
total.t_rmshr += object->resident_page_count;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (object->flags & OBJ_ACTIVE) {
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_avmshr += object->size;
|
|
|
|
|
total.t_armshr += object->resident_page_count;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
2002-04-20 07:23:22 +00:00
|
|
|
mtx_unlock(&vm_object_list_mtx);
|
2006-11-20 08:33:55 +00:00
|
|
|
total.t_free = cnt.v_free_count + cnt.v_cache_count;
|
|
|
|
|
return (sysctl_handle_opaque(oidp, &total, sizeof(total), req));
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1995-11-14 09:29:34 +00:00
|
|
|
|
2002-04-04 21:38:47 +00:00
|
|
|
/*
|
|
|
|
|
* vcnt() - accumulate statistics from all cpus and the global cnt
|
|
|
|
|
* structure.
|
|
|
|
|
*
|
|
|
|
|
* The vmmeter structure is now per-cpu as well as global. Those
|
|
|
|
|
* statistics which can be kept on a per-cpu basis (to avoid cache
|
|
|
|
|
* stalls between cpus) can be moved to the per-cpu vmmeter. Remaining
|
|
|
|
|
* statistics, such as v_free_reserved, are left in the global
|
|
|
|
|
* structure.
|
|
|
|
|
*
|
|
|
|
|
* (sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req)
|
|
|
|
|
*/
|
|
|
|
|
static int
|
|
|
|
|
vcnt(SYSCTL_HANDLER_ARGS)
|
|
|
|
|
{
|
|
|
|
|
int count = *(int *)arg1;
|
2005-04-12 23:15:28 +00:00
|
|
|
int offset = (char *)arg1 - (char *)&cnt;
|
2002-04-04 21:38:47 +00:00
|
|
|
#ifdef SMP
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < mp_ncpus; ++i) {
|
|
|
|
|
struct pcpu *pcpu = pcpu_find(i);
|
|
|
|
|
count += *(int *)((char *)&pcpu->pc_cnt + offset);
|
|
|
|
|
}
|
2002-12-22 05:04:30 +00:00
|
|
|
#else
|
|
|
|
|
count += *(int *)((char *)PCPU_PTR(cnt) + offset);
|
2002-04-04 21:38:47 +00:00
|
|
|
#endif
|
2005-04-12 23:15:28 +00:00
|
|
|
return (SYSCTL_OUT(req, &count, sizeof(int)));
|
2002-04-04 21:38:47 +00:00
|
|
|
}
|
|
|
|
|
|
2003-01-11 07:29:47 +00:00
|
|
|
SYSCTL_PROC(_vm, VM_TOTAL, vmtotal, CTLTYPE_OPAQUE|CTLFLAG_RD,
|
1999-05-03 23:57:32 +00:00
|
|
|
0, sizeof(struct vmtotal), vmtotal, "S,vmtotal",
|
|
|
|
|
"System virtual memory statistics");
|
2005-02-11 16:34:14 +00:00
|
|
|
SYSCTL_NODE(_vm, OID_AUTO, stats, CTLFLAG_RW, 0, "VM meter stats");
|
2005-02-10 12:18:36 +00:00
|
|
|
static SYSCTL_NODE(_vm_stats, OID_AUTO, sys, CTLFLAG_RW, 0,
|
|
|
|
|
"VM meter sys stats");
|
|
|
|
|
static SYSCTL_NODE(_vm_stats, OID_AUTO, vm, CTLFLAG_RW, 0,
|
|
|
|
|
"VM meter vm stats");
|
1998-10-31 17:21:31 +00:00
|
|
|
SYSCTL_NODE(_vm_stats, OID_AUTO, misc, CTLFLAG_RW, 0, "VM meter misc stats");
|
2002-04-04 21:38:47 +00:00
|
|
|
|
|
|
|
|
SYSCTL_PROC(_vm_stats_sys, OID_AUTO, v_swtch, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_swtch, 0, vcnt, "IU", "Context switches");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_sys, OID_AUTO, v_trap, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_trap, 0, vcnt, "IU", "Traps");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_sys, OID_AUTO, v_syscall, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_syscall, 0, vcnt, "IU", "Syscalls");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_sys, OID_AUTO, v_intr, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_intr, 0, vcnt, "IU", "Hardware interrupts");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_sys, OID_AUTO, v_soft, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_soft, 0, vcnt, "IU", "Software interrupts");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_vm_faults, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_vm_faults, 0, vcnt, "IU", "VM faults");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_cow_faults, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_cow_faults, 0, vcnt, "IU", "COW faults");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_cow_optim, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_cow_optim, 0, vcnt, "IU", "Optimized COW faults");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_zfod, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_zfod, 0, vcnt, "IU", "Zero fill");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_ozfod, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_ozfod, 0, vcnt, "IU", "Optimized zero fill");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_swapin, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_swapin, 0, vcnt, "IU", "Swapin operations");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_swapout, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_swapout, 0, vcnt, "IU", "Swapout operations");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_swappgsin, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_swappgsin, 0, vcnt, "IU", "Swapin pages");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_swappgsout, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_swappgsout, 0, vcnt, "IU", "Swapout pages");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_vnodein, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_vnodein, 0, vcnt, "IU", "Vnodein operations");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_vnodeout, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_vnodeout, 0, vcnt, "IU", "Vnodeout operations");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_vnodepgsin, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_vnodepgsin, 0, vcnt, "IU", "Vnodein pages");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_vnodepgsout, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_vnodepgsout, 0, vcnt, "IU", "Vnodeout pages");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_intrans, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_intrans, 0, vcnt, "IU", "In transit page blocking");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_reactivated, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_reactivated, 0, vcnt, "IU", "Reactivated pages");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_pdwakeups, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_pdwakeups, 0, vcnt, "IU", "Pagedaemon wakeups");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_pdpages, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_pdpages, 0, vcnt, "IU", "Pagedaemon page scans");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_dfree, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_dfree, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_pfree, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_pfree, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_tfree, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_tfree, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_page_size, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_page_size, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_page_count, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_page_count, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_free_reserved, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_free_reserved, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_free_target, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_free_target, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_free_min, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_free_min, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_free_count, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_free_count, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_wire_count, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_wire_count, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_active_count, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_active_count, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_inactive_target, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_inactive_target, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_inactive_count, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_inactive_count, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_cache_count, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_cache_count, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_cache_min, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_cache_min, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_cache_max, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_cache_max, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_pageout_free_min, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_pageout_free_min, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_interrupt_free_min, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_interrupt_free_min, 0, vcnt, "IU", "");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_forks, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_forks, 0, vcnt, "IU", "Number of fork() calls");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_vforks, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_vforks, 0, vcnt, "IU", "Number of vfork() calls");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_rforks, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_rforks, 0, vcnt, "IU", "Number of rfork() calls");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_kthreads, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_kthreads, 0, vcnt, "IU", "Number of fork() calls by kernel");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_forkpages, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_forkpages, 0, vcnt, "IU", "VM pages affected by fork()");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_vforkpages, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_vforkpages, 0, vcnt, "IU", "VM pages affected by vfork()");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_rforkpages, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_rforkpages, 0, vcnt, "IU", "VM pages affected by rfork()");
|
|
|
|
|
SYSCTL_PROC(_vm_stats_vm, OID_AUTO, v_kthreadpages, CTLTYPE_UINT|CTLFLAG_RD,
|
|
|
|
|
&cnt.v_kthreadpages, 0, vcnt, "IU", "VM pages affected by fork() by kernel");
|
|
|
|
|
|
1998-10-31 17:21:31 +00:00
|
|
|
SYSCTL_INT(_vm_stats_misc, OID_AUTO,
|
|
|
|
|
zero_page_count, CTLFLAG_RD, &vm_page_zero_count, 0, "");
|
|
|
|
|
#if 0
|
|
|
|
|
SYSCTL_INT(_vm_stats_misc, OID_AUTO,
|
|
|
|
|
page_mask, CTLFLAG_RD, &page_mask, 0, "");
|
|
|
|
|
SYSCTL_INT(_vm_stats_misc, OID_AUTO,
|
|
|
|
|
page_shift, CTLFLAG_RD, &page_shift, 0, "");
|
|
|
|
|
SYSCTL_INT(_vm_stats_misc, OID_AUTO,
|
|
|
|
|
first_page, CTLFLAG_RD, &first_page, 0, "");
|
|
|
|
|
SYSCTL_INT(_vm_stats_misc, OID_AUTO,
|
|
|
|
|
last_page, CTLFLAG_RD, &last_page, 0, "");
|
|
|
|
|
#endif
|
