2005-01-07 02:29:27 +00:00
|
|
|
/*-
|
1994-05-24 10:09:53 +00:00
|
|
|
* Copyright (c) 1991, 1993
|
|
|
|
* The Regents of the University of California. All rights reserved.
|
|
|
|
*
|
|
|
|
* This code is derived from software contributed to Berkeley by
|
|
|
|
* The Mach Operating System project at Carnegie-Mellon University.
|
|
|
|
*
|
|
|
|
* 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.
|
|
|
|
*
|
1994-08-02 07:55:43 +00:00
|
|
|
* from: @(#)vm_page.h 8.2 (Berkeley) 12/13/93
|
1994-05-24 10:09:53 +00:00
|
|
|
*
|
|
|
|
*
|
|
|
|
* Copyright (c) 1987, 1990 Carnegie-Mellon University.
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* Authors: Avadis Tevanian, Jr., Michael Wayne Young
|
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
|
|
|
*
|
1994-05-24 10:09:53 +00:00
|
|
|
* Permission to use, copy, modify and distribute this software and
|
|
|
|
* its documentation is hereby granted, provided that both the copyright
|
|
|
|
* notice and this permission notice appear in all copies of the
|
|
|
|
* software, derivative works or modified versions, and any portions
|
|
|
|
* thereof, and that both notices appear in supporting documentation.
|
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
|
|
|
*
|
|
|
|
* CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
|
|
|
|
* CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND
|
1994-05-24 10:09:53 +00:00
|
|
|
* FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
|
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
|
|
|
*
|
1994-05-24 10:09:53 +00:00
|
|
|
* Carnegie Mellon requests users of this software to return to
|
|
|
|
*
|
|
|
|
* Software Distribution Coordinator or Software.Distribution@CS.CMU.EDU
|
|
|
|
* School of Computer Science
|
|
|
|
* Carnegie Mellon University
|
|
|
|
* Pittsburgh PA 15213-3890
|
|
|
|
*
|
|
|
|
* any improvements or extensions that they make and grant Carnegie the
|
|
|
|
* rights to redistribute these changes.
|
1994-08-02 07:55:43 +00:00
|
|
|
*
|
1999-08-28 01:08:13 +00:00
|
|
|
* $FreeBSD$
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Resident memory system definitions.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#ifndef _VM_PAGE_
|
|
|
|
#define _VM_PAGE_
|
|
|
|
|
1995-03-01 23:30:04 +00:00
|
|
|
#include <vm/pmap.h>
|
1998-08-24 08:39:39 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Management of resident (logical) pages.
|
|
|
|
*
|
|
|
|
* A small structure is kept for each resident
|
|
|
|
* page, indexed by page number. Each structure
|
2013-06-04 05:44:52 +00:00
|
|
|
* is an element of several collections:
|
1994-05-24 10:09:53 +00:00
|
|
|
*
|
2013-06-04 05:44:52 +00:00
|
|
|
* A radix tree used to quickly
|
1994-05-24 10:09:53 +00:00
|
|
|
* perform object/offset lookups
|
|
|
|
*
|
|
|
|
* A list of all pages for a given object,
|
|
|
|
* so they can be quickly deactivated at
|
|
|
|
* time of deallocation.
|
|
|
|
*
|
|
|
|
* An ordered list of pages due for pageout.
|
|
|
|
*
|
|
|
|
* In addition, the structure contains the object
|
|
|
|
* and offset to which this page belongs (for pageout),
|
|
|
|
* and sundry status bits.
|
|
|
|
*
|
2011-06-19 19:13:24 +00:00
|
|
|
* In general, operations on this structure's mutable fields are
|
|
|
|
* synchronized using either one of or a combination of the lock on the
|
|
|
|
* object that the page belongs to (O), the pool lock for the page (P),
|
2012-11-13 02:50:39 +00:00
|
|
|
* or the lock for either the free or paging queue (Q). If a field is
|
2011-06-19 19:13:24 +00:00
|
|
|
* annotated below with two of these locks, then holding either lock is
|
2011-09-12 15:11:34 +00:00
|
|
|
* sufficient for read access, but both locks are required for write
|
2011-06-19 19:13:24 +00:00
|
|
|
* access.
|
|
|
|
*
|
2011-09-12 15:11:34 +00:00
|
|
|
* In contrast, the synchronization of accesses to the page's
|
|
|
|
* dirty field is machine dependent (M). In the
|
|
|
|
* machine-independent layer, the lock on the object that the
|
|
|
|
* page belongs to must be held in order to operate on the field.
|
|
|
|
* However, the pmap layer is permitted to set all bits within
|
|
|
|
* the field without holding that lock. If the underlying
|
|
|
|
* architecture does not support atomic read-modify-write
|
|
|
|
* operations on the field's type, then the machine-independent
|
2011-09-29 09:22:54 +00:00
|
|
|
* layer uses a 32-bit atomic on the aligned 32-bit word that
|
2011-09-12 15:11:34 +00:00
|
|
|
* contains the dirty field. In the machine-independent layer,
|
|
|
|
* the implementation of read-modify-write operations on the
|
|
|
|
* field is encapsulated in vm_page_clear_dirty_mask().
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
|
2011-11-08 11:08:40 +00:00
|
|
|
#if PAGE_SIZE == 4096
|
|
|
|
#define VM_PAGE_BITS_ALL 0xffu
|
|
|
|
typedef uint8_t vm_page_bits_t;
|
|
|
|
#elif PAGE_SIZE == 8192
|
|
|
|
#define VM_PAGE_BITS_ALL 0xffffu
|
|
|
|
typedef uint16_t vm_page_bits_t;
|
|
|
|
#elif PAGE_SIZE == 16384
|
|
|
|
#define VM_PAGE_BITS_ALL 0xffffffffu
|
|
|
|
typedef uint32_t vm_page_bits_t;
|
|
|
|
#elif PAGE_SIZE == 32768
|
|
|
|
#define VM_PAGE_BITS_ALL 0xfffffffffffffffflu
|
|
|
|
typedef uint64_t vm_page_bits_t;
|
|
|
|
#endif
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
struct vm_page {
|
2013-08-10 17:36:42 +00:00
|
|
|
union {
|
|
|
|
TAILQ_ENTRY(vm_page) q; /* page queue or free list (Q) */
|
|
|
|
struct {
|
|
|
|
SLIST_ENTRY(vm_page) ss; /* private slists */
|
|
|
|
void *pv;
|
|
|
|
} s;
|
|
|
|
struct {
|
|
|
|
u_long p;
|
|
|
|
u_long v;
|
|
|
|
} memguard;
|
|
|
|
} plinks;
|
|
|
|
TAILQ_ENTRY(vm_page) listq; /* pages in same object (O) */
|
|
|
|
vm_object_t object; /* which object am I in (O,P) */
|
2011-01-16 18:04:01 +00:00
|
|
|
vm_pindex_t pindex; /* offset into object (O,P) */
|
2003-10-25 18:33:04 +00:00
|
|
|
vm_paddr_t phys_addr; /* physical address of page */
|
2000-05-21 12:50:18 +00:00
|
|
|
struct md_page md; /* machine dependant stuff */
|
2013-09-16 06:25:54 +00:00
|
|
|
u_int wire_count; /* wired down maps refs (P) */
|
|
|
|
volatile u_int busy_lock; /* busy owners lock */
|
|
|
|
uint16_t hold_count; /* page hold count (P) */
|
|
|
|
uint16_t flags; /* page PG_* flags (P) */
|
|
|
|
uint8_t aflags; /* access is atomic */
|
|
|
|
uint8_t oflags; /* page VPO_* flags (O) */
|
2010-05-06 16:39:43 +00:00
|
|
|
uint8_t queue; /* page queue index (P,Q) */
|
2009-01-03 13:24:08 +00:00
|
|
|
int8_t segind;
|
Enable the new physical memory allocator.
This allocator uses a binary buddy system with a twist. First and
foremost, this allocator is required to support the implementation of
superpages. As a side effect, it enables a more robust implementation
of contigmalloc(9). Moreover, this reimplementation of
contigmalloc(9) eliminates the acquisition of Giant by
contigmalloc(..., M_NOWAIT, ...).
The twist is that this allocator tries to reduce the number of TLB
misses incurred by accesses through a direct map to small, UMA-managed
objects and page table pages. Roughly speaking, the physical pages
that are allocated for such purposes are clustered together in the
physical address space. The performance benefits vary. In the most
extreme case, a uniprocessor kernel running on an Opteron, I measured
an 18% reduction in system time during a buildworld.
This allocator does not implement page coloring. The reason is that
superpages have much the same effect. The contiguous physical memory
allocation necessary for a superpage is inherently colored.
Finally, the one caveat is that this allocator does not effectively
support prezeroed pages. I hope this is temporary. On i386, this is
a slight pessimization. However, on amd64, the beneficial effects of
the direct-map optimization outweigh the ill effects. I speculate
that this is true in general of machines with a direct map.
Approved by: re
2007-06-16 04:57:06 +00:00
|
|
|
uint8_t order; /* index of the buddy queue */
|
|
|
|
uint8_t pool;
|
2013-04-08 20:02:27 +00:00
|
|
|
u_char act_count; /* page usage count (P) */
|
2013-08-10 17:36:42 +00:00
|
|
|
/* NOTE that these must support one bit per DEV_BSIZE in a page */
|
1996-01-19 04:00:31 +00:00
|
|
|
/* so, on normal X86 kernels, they must be at least 8 bits wide */
|
2011-11-08 11:08:40 +00:00
|
|
|
vm_page_bits_t valid; /* map of valid DEV_BSIZE chunks (O) */
|
|
|
|
vm_page_bits_t dirty; /* map of dirty DEV_BSIZE chunks (M) */
|
1994-05-24 10:09:53 +00:00
|
|
|
};
|
|
|
|
|
2006-08-09 17:43:27 +00:00
|
|
|
/*
|
|
|
|
* Page flags stored in oflags:
|
|
|
|
*
|
|
|
|
* Access to these page flags is synchronized by the lock on the object
|
|
|
|
* containing the page (O).
|
2011-08-09 21:01:36 +00:00
|
|
|
*
|
|
|
|
* Note: VPO_UNMANAGED (used by OBJT_DEVICE, OBJT_PHYS and OBJT_SG)
|
|
|
|
* indicates that the page is not under PV management but
|
|
|
|
* otherwise should be treated as a normal page. Pages not
|
|
|
|
* under PV management cannot be paged out via the
|
|
|
|
* object/vm_page_t because there is no knowledge of their pte
|
|
|
|
* mappings, and such pages are also not on any PQ queue.
|
|
|
|
*
|
2006-08-09 17:43:27 +00:00
|
|
|
*/
|
2013-08-09 11:11:11 +00:00
|
|
|
#define VPO_UNUSED01 0x01 /* --available-- */
|
|
|
|
#define VPO_SWAPSLEEP 0x02 /* waiting for swap to finish */
|
Replace the page hold queue, PQ_HOLD, by a new page flag, PG_UNHOLDFREE,
because the queue itself serves no purpose. When a held page is freed,
inserting the page into the hold queue has the side effect of setting the
page's "queue" field to PQ_HOLD. Later, when the page is unheld, it will
be freed because the "queue" field is PQ_HOLD. In other words, PQ_HOLD is
used as a flag, not a queue. So, this change replaces it with a flag.
To accomodate the new page flag, make the page's "flags" field wider and
"oflags" field narrower.
Reviewed by: kib
2012-10-29 06:15:04 +00:00
|
|
|
#define VPO_UNMANAGED 0x04 /* no PV management for page */
|
|
|
|
#define VPO_SWAPINPROG 0x08 /* swap I/O in progress on page */
|
|
|
|
#define VPO_NOSYNC 0x10 /* do not collect for syncer */
|
2006-08-09 17:43:27 +00:00
|
|
|
|
2013-08-09 11:11:11 +00:00
|
|
|
/*
|
|
|
|
* Busy page implementation details.
|
|
|
|
* The algorithm is taken mostly by rwlock(9) and sx(9) locks implementation,
|
|
|
|
* even if the support for owner identity is removed because of size
|
|
|
|
* constraints. Checks on lock recursion are then not possible, while the
|
|
|
|
* lock assertions effectiveness is someway reduced.
|
|
|
|
*/
|
|
|
|
#define VPB_BIT_SHARED 0x01
|
|
|
|
#define VPB_BIT_EXCLUSIVE 0x02
|
|
|
|
#define VPB_BIT_WAITERS 0x04
|
|
|
|
#define VPB_BIT_FLAGMASK \
|
|
|
|
(VPB_BIT_SHARED | VPB_BIT_EXCLUSIVE | VPB_BIT_WAITERS)
|
|
|
|
|
|
|
|
#define VPB_SHARERS_SHIFT 3
|
|
|
|
#define VPB_SHARERS(x) \
|
|
|
|
(((x) & ~VPB_BIT_FLAGMASK) >> VPB_SHARERS_SHIFT)
|
|
|
|
#define VPB_SHARERS_WORD(x) ((x) << VPB_SHARERS_SHIFT | VPB_BIT_SHARED)
|
|
|
|
#define VPB_ONE_SHARER (1 << VPB_SHARERS_SHIFT)
|
|
|
|
|
|
|
|
#define VPB_SINGLE_EXCLUSIVER VPB_BIT_EXCLUSIVE
|
|
|
|
|
|
|
|
#define VPB_UNBUSIED VPB_SHARERS_WORD(0)
|
|
|
|
|
2011-01-17 19:17:26 +00:00
|
|
|
#define PQ_NONE 255
|
|
|
|
#define PQ_INACTIVE 0
|
|
|
|
#define PQ_ACTIVE 1
|
Replace the page hold queue, PQ_HOLD, by a new page flag, PG_UNHOLDFREE,
because the queue itself serves no purpose. When a held page is freed,
inserting the page into the hold queue has the side effect of setting the
page's "queue" field to PQ_HOLD. Later, when the page is unheld, it will
be freed because the "queue" field is PQ_HOLD. In other words, PQ_HOLD is
used as a flag, not a queue. So, this change replaces it with a flag.
To accomodate the new page flag, make the page's "flags" field wider and
"oflags" field narrower.
Reviewed by: kib
2012-10-29 06:15:04 +00:00
|
|
|
#define PQ_COUNT 2
|
2002-07-04 22:07:37 +00:00
|
|
|
|
2012-11-13 02:50:39 +00:00
|
|
|
TAILQ_HEAD(pglist, vm_page);
|
2013-08-10 17:36:42 +00:00
|
|
|
SLIST_HEAD(spglist, vm_page);
|
2009-10-04 18:53:10 +00:00
|
|
|
|
2012-11-13 02:50:39 +00:00
|
|
|
struct vm_pagequeue {
|
|
|
|
struct mtx pq_mutex;
|
|
|
|
struct pglist pq_pl;
|
Split the pagequeues per NUMA domains, and split pageademon process
into threads each processing queue in a single domain. The structure
of the pagedaemons and queues is kept intact, most of the changes come
from the need for code to find an owning page queue for given page,
calculated from the segment containing the page.
The tie between NUMA domain and pagedaemon thread/pagequeue split is
rather arbitrary, the multithreaded daemon could be allowed for the
single-domain machines, or one domain might be split into several page
domains, to further increase concurrency.
Right now, each pagedaemon thread tries to reach the global target,
precalculated at the start of the pass. This is not optimal, since it
could cause excessive page deactivation and freeing. The code should
be changed to re-check the global page deficit state in the loop after
some number of iterations.
The pagedaemons reach the quorum before starting the OOM, since one
thread inability to meet the target is normal for split queues. Only
when all pagedaemons fail to produce enough reusable pages, OOM is
started by single selected thread.
Launder is modified to take into account the segments layout with
regard to the region for which cleaning is performed.
Based on the preliminary patch by jeff, sponsored by EMC / Isilon
Storage Division.
Reviewed by: alc
Tested by: pho
Sponsored by: The FreeBSD Foundation
2013-08-07 16:36:38 +00:00
|
|
|
int pq_cnt;
|
|
|
|
int * const pq_vcnt;
|
|
|
|
const char * const pq_name;
|
2009-10-04 18:53:10 +00:00
|
|
|
} __aligned(CACHE_LINE_SIZE);
|
|
|
|
|
Split the pagequeues per NUMA domains, and split pageademon process
into threads each processing queue in a single domain. The structure
of the pagedaemons and queues is kept intact, most of the changes come
from the need for code to find an owning page queue for given page,
calculated from the segment containing the page.
The tie between NUMA domain and pagedaemon thread/pagequeue split is
rather arbitrary, the multithreaded daemon could be allowed for the
single-domain machines, or one domain might be split into several page
domains, to further increase concurrency.
Right now, each pagedaemon thread tries to reach the global target,
precalculated at the start of the pass. This is not optimal, since it
could cause excessive page deactivation and freeing. The code should
be changed to re-check the global page deficit state in the loop after
some number of iterations.
The pagedaemons reach the quorum before starting the OOM, since one
thread inability to meet the target is normal for split queues. Only
when all pagedaemons fail to produce enough reusable pages, OOM is
started by single selected thread.
Launder is modified to take into account the segments layout with
regard to the region for which cleaning is performed.
Based on the preliminary patch by jeff, sponsored by EMC / Isilon
Storage Division.
Reviewed by: alc
Tested by: pho
Sponsored by: The FreeBSD Foundation
2013-08-07 16:36:38 +00:00
|
|
|
|
|
|
|
struct vm_domain {
|
|
|
|
struct vm_pagequeue vmd_pagequeues[PQ_COUNT];
|
|
|
|
u_int vmd_page_count;
|
|
|
|
u_int vmd_free_count;
|
|
|
|
long vmd_segs; /* bitmask of the segments */
|
|
|
|
boolean_t vmd_oom;
|
|
|
|
int vmd_pass; /* local pagedaemon pass */
|
|
|
|
struct vm_page vmd_marker; /* marker for pagedaemon private use */
|
|
|
|
};
|
|
|
|
|
|
|
|
extern struct vm_domain vm_dom[MAXMEMDOM];
|
2012-11-13 02:50:39 +00:00
|
|
|
|
|
|
|
#define vm_pagequeue_assert_locked(pq) mtx_assert(&(pq)->pq_mutex, MA_OWNED)
|
|
|
|
#define vm_pagequeue_lock(pq) mtx_lock(&(pq)->pq_mutex)
|
|
|
|
#define vm_pagequeue_unlock(pq) mtx_unlock(&(pq)->pq_mutex)
|
2009-10-04 18:53:10 +00:00
|
|
|
|
Split the pagequeues per NUMA domains, and split pageademon process
into threads each processing queue in a single domain. The structure
of the pagedaemons and queues is kept intact, most of the changes come
from the need for code to find an owning page queue for given page,
calculated from the segment containing the page.
The tie between NUMA domain and pagedaemon thread/pagequeue split is
rather arbitrary, the multithreaded daemon could be allowed for the
single-domain machines, or one domain might be split into several page
domains, to further increase concurrency.
Right now, each pagedaemon thread tries to reach the global target,
precalculated at the start of the pass. This is not optimal, since it
could cause excessive page deactivation and freeing. The code should
be changed to re-check the global page deficit state in the loop after
some number of iterations.
The pagedaemons reach the quorum before starting the OOM, since one
thread inability to meet the target is normal for split queues. Only
when all pagedaemons fail to produce enough reusable pages, OOM is
started by single selected thread.
Launder is modified to take into account the segments layout with
regard to the region for which cleaning is performed.
Based on the preliminary patch by jeff, sponsored by EMC / Isilon
Storage Division.
Reviewed by: alc
Tested by: pho
Sponsored by: The FreeBSD Foundation
2013-08-07 16:36:38 +00:00
|
|
|
#ifdef _KERNEL
|
|
|
|
static __inline void
|
|
|
|
vm_pagequeue_cnt_add(struct vm_pagequeue *pq, int addend)
|
|
|
|
{
|
|
|
|
|
|
|
|
#ifdef notyet
|
|
|
|
vm_pagequeue_assert_locked(pq);
|
|
|
|
#endif
|
|
|
|
pq->pq_cnt += addend;
|
|
|
|
atomic_add_int(pq->pq_vcnt, addend);
|
|
|
|
}
|
|
|
|
#define vm_pagequeue_cnt_inc(pq) vm_pagequeue_cnt_add((pq), 1)
|
|
|
|
#define vm_pagequeue_cnt_dec(pq) vm_pagequeue_cnt_add((pq), -1)
|
|
|
|
#endif /* _KERNEL */
|
|
|
|
|
2012-10-31 18:07:18 +00:00
|
|
|
extern struct mtx_padalign vm_page_queue_free_mtx;
|
|
|
|
extern struct mtx_padalign pa_lock[];
|
2009-10-04 18:53:10 +00:00
|
|
|
|
2010-04-30 00:46:43 +00:00
|
|
|
#if defined(__arm__)
|
|
|
|
#define PDRSHIFT PDR_SHIFT
|
|
|
|
#elif !defined(PDRSHIFT)
|
|
|
|
#define PDRSHIFT 21
|
|
|
|
#endif
|
1999-03-14 20:40:15 +00:00
|
|
|
|
2010-04-30 00:46:43 +00:00
|
|
|
#define pa_index(pa) ((pa) >> PDRSHIFT)
|
2012-10-31 18:07:18 +00:00
|
|
|
#define PA_LOCKPTR(pa) ((struct mtx *)(&pa_lock[pa_index(pa) % PA_LOCK_COUNT]))
|
2010-04-30 00:46:43 +00:00
|
|
|
#define PA_LOCKOBJPTR(pa) ((struct lock_object *)PA_LOCKPTR((pa)))
|
|
|
|
#define PA_LOCK(pa) mtx_lock(PA_LOCKPTR(pa))
|
|
|
|
#define PA_TRYLOCK(pa) mtx_trylock(PA_LOCKPTR(pa))
|
|
|
|
#define PA_UNLOCK(pa) mtx_unlock(PA_LOCKPTR(pa))
|
|
|
|
#define PA_UNLOCK_COND(pa) \
|
|
|
|
do { \
|
Roughly half of a typical pmap_mincore() implementation is machine-
independent code. Move this code into mincore(), and eliminate the
page queues lock from pmap_mincore().
Push down the page queues lock into pmap_clear_modify(),
pmap_clear_reference(), and pmap_is_modified(). Assert that these
functions are never passed an unmanaged page.
Eliminate an inaccurate comment from powerpc/powerpc/mmu_if.m:
Contrary to what the comment says, pmap_mincore() is not simply an
optimization. Without a complete pmap_mincore() implementation,
mincore() cannot return either MINCORE_MODIFIED or MINCORE_REFERENCED
because only the pmap can provide this information.
Eliminate the page queues lock from vfs_setdirty_locked_object(),
vm_pageout_clean(), vm_object_page_collect_flush(), and
vm_object_page_clean(). Generally speaking, these are all accesses
to the page's dirty field, which are synchronized by the containing
vm object's lock.
Reduce the scope of the page queues lock in vm_object_madvise() and
vm_page_dontneed().
Reviewed by: kib (an earlier version)
2010-05-24 14:26:57 +00:00
|
|
|
if ((pa) != 0) { \
|
|
|
|
PA_UNLOCK((pa)); \
|
|
|
|
(pa) = 0; \
|
|
|
|
} \
|
2010-04-30 00:46:43 +00:00
|
|
|
} while (0)
|
|
|
|
|
|
|
|
#define PA_LOCK_ASSERT(pa, a) mtx_assert(PA_LOCKPTR(pa), (a))
|
|
|
|
|
2011-11-29 13:07:32 +00:00
|
|
|
#ifdef KLD_MODULE
|
|
|
|
#define vm_page_lock(m) vm_page_lock_KBI((m), LOCK_FILE, LOCK_LINE)
|
|
|
|
#define vm_page_unlock(m) vm_page_unlock_KBI((m), LOCK_FILE, LOCK_LINE)
|
|
|
|
#define vm_page_trylock(m) vm_page_trylock_KBI((m), LOCK_FILE, LOCK_LINE)
|
|
|
|
#else /* !KLD_MODULE */
|
2010-04-30 00:46:43 +00:00
|
|
|
#define vm_page_lockptr(m) (PA_LOCKPTR(VM_PAGE_TO_PHYS((m))))
|
|
|
|
#define vm_page_lock(m) mtx_lock(vm_page_lockptr((m)))
|
|
|
|
#define vm_page_unlock(m) mtx_unlock(vm_page_lockptr((m)))
|
|
|
|
#define vm_page_trylock(m) mtx_trylock(vm_page_lockptr((m)))
|
2013-05-31 16:00:42 +00:00
|
|
|
#endif
|
|
|
|
#if defined(INVARIANTS)
|
2013-06-03 01:22:54 +00:00
|
|
|
#define vm_page_assert_locked(m) \
|
|
|
|
vm_page_assert_locked_KBI((m), __FILE__, __LINE__)
|
2013-05-31 16:00:42 +00:00
|
|
|
#define vm_page_lock_assert(m, a) \
|
|
|
|
vm_page_lock_assert_KBI((m), (a), __FILE__, __LINE__)
|
|
|
|
#else
|
2013-06-03 01:22:54 +00:00
|
|
|
#define vm_page_assert_locked(m)
|
2013-05-31 16:00:42 +00:00
|
|
|
#define vm_page_lock_assert(m, a)
|
2011-11-29 13:07:32 +00:00
|
|
|
#endif
|
2010-04-30 00:46:43 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
2012-08-03 01:48:15 +00:00
|
|
|
* The vm_page's aflags are updated using atomic operations. To set or clear
|
|
|
|
* these flags, the functions vm_page_aflag_set() and vm_page_aflag_clear()
|
|
|
|
* must be used. Neither these flags nor these functions are part of the KBI.
|
2011-09-06 10:30:11 +00:00
|
|
|
*
|
2013-06-03 01:22:54 +00:00
|
|
|
* PGA_REFERENCED may be cleared only if the page is locked. It is set by
|
|
|
|
* both the MI and MD VM layers. However, kernel loadable modules should not
|
|
|
|
* directly set this flag. They should call vm_page_reference() instead.
|
2010-06-10 16:56:35 +00:00
|
|
|
*
|
2011-09-06 10:30:11 +00:00
|
|
|
* PGA_WRITEABLE is set exclusively on managed pages by pmap_enter(). When it
|
2013-08-09 11:11:11 +00:00
|
|
|
* does so, the page must be exclusive busied. The MI VM layer must never
|
|
|
|
* access this flag directly. Instead, it should call
|
|
|
|
* pmap_page_is_write_mapped().
|
2012-04-06 19:59:37 +00:00
|
|
|
*
|
|
|
|
* PGA_EXECUTABLE may be set by pmap routines, and indicates that a page has
|
2012-06-16 18:56:19 +00:00
|
|
|
* at least one executable mapping. It is not consumed by the MI VM layer.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2011-09-06 10:30:11 +00:00
|
|
|
#define PGA_WRITEABLE 0x01 /* page may be mapped writeable */
|
|
|
|
#define PGA_REFERENCED 0x02 /* page has been referenced */
|
2012-04-06 19:59:37 +00:00
|
|
|
#define PGA_EXECUTABLE 0x04 /* page may be mapped executable */
|
2011-09-06 10:30:11 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Page flags. If changed at any other time than page allocation or
|
|
|
|
* freeing, the modification must be protected by the vm_page lock.
|
|
|
|
*/
|
Replace the page hold queue, PQ_HOLD, by a new page flag, PG_UNHOLDFREE,
because the queue itself serves no purpose. When a held page is freed,
inserting the page into the hold queue has the side effect of setting the
page's "queue" field to PQ_HOLD. Later, when the page is unheld, it will
be freed because the "queue" field is PQ_HOLD. In other words, PQ_HOLD is
used as a flag, not a queue. So, this change replaces it with a flag.
To accomodate the new page flag, make the page's "flags" field wider and
"oflags" field narrower.
Reviewed by: kib
2012-10-29 06:15:04 +00:00
|
|
|
#define PG_CACHED 0x0001 /* page is cached */
|
|
|
|
#define PG_FREE 0x0002 /* page is free */
|
|
|
|
#define PG_FICTITIOUS 0x0004 /* physical page doesn't exist */
|
|
|
|
#define PG_ZERO 0x0008 /* page is zeroed */
|
|
|
|
#define PG_MARKER 0x0010 /* special queue marker page */
|
|
|
|
#define PG_WINATCFLS 0x0040 /* flush dirty page on inactive q */
|
|
|
|
#define PG_NODUMP 0x0080 /* don't include this page in a dump */
|
|
|
|
#define PG_UNHOLDFREE 0x0100 /* delayed free of a held page */
|
1994-05-24 10:09:53 +00:00
|
|
|
|
NOTE: libkvm, w, ps, 'top', and any other utility which depends on struct
proc or any VM system structure will have to be rebuilt!!!
Much needed overhaul of the VM system. Included in this first round of
changes:
1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages,
haspage, and sync operations are supported. The haspage interface now
provides information about clusterability. All pager routines now take
struct vm_object's instead of "pagers".
2) Improved data structures. In the previous paradigm, there is constant
confusion caused by pagers being both a data structure ("allocate a
pager") and a collection of routines. The idea of a pager structure has
escentially been eliminated. Objects now have types, and this type is
used to index the appropriate pager. In most cases, items in the pager
structure were duplicated in the object data structure and thus were
unnecessary. In the few cases that remained, a un_pager structure union
was created in the object to contain these items.
3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now
be removed. For instance, vm_object_enter(), vm_object_lookup(),
vm_object_remove(), and the associated object hash list were some of the
things that were removed.
4) simple_lock's removed. Discussion with several people reveals that the
SMP locking primitives used in the VM system aren't likely the mechanism
that we'll be adopting. Even if it were, the locking that was in the code
was very inadequate and would have to be mostly re-done anyway. The
locking in a uni-processor kernel was a no-op but went a long way toward
making the code difficult to read and debug.
5) Places that attempted to kludge-up the fact that we don't have kernel
thread support have been fixed to reflect the reality that we are really
dealing with processes, not threads. The VM system didn't have complete
thread support, so the comments and mis-named routines were just wrong.
We now use tsleep and wakeup directly in the lock routines, for instance.
6) Where appropriate, the pagers have been improved, especially in the
pager_alloc routines. Most of the pager_allocs have been rewritten and
are now faster and easier to maintain.
7) The pagedaemon pageout clustering algorithm has been rewritten and
now tries harder to output an even number of pages before and after
the requested page. This is sort of the reverse of the ideal pagein
algorithm and should provide better overall performance.
8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup
have been removed. Some other unnecessary casts have also been removed.
9) Some almost useless debugging code removed.
10) Terminology of shadow objects vs. backing objects straightened out.
The fact that the vm_object data structure escentially had this
backwards really confused things. The use of "shadow" and "backing
object" throughout the code is now internally consistent and correct
in the Mach terminology.
11) Several minor bug fixes, including one in the vm daemon that caused
0 RSS objects to not get purged as intended.
12) A "default pager" has now been created which cleans up the transition
of objects to the "swap" type. The previous checks throughout the code
for swp->pg_data != NULL were really ugly. This change also provides
the rudiments for future backing of "anonymous" memory by something
other than the swap pager (via the vnode pager, for example), and it
allows the decision about which of these pagers to use to be made
dynamically (although will need some additional decision code to do
this, of course).
13) (dyson) MAP_COPY has been deprecated and the corresponding "copy
object" code has been removed. MAP_COPY was undocumented and non-
standard. It was furthermore broken in several ways which caused its
behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will
continue to work correctly, but via the slightly different semantics
of MAP_PRIVATE.
14) (dyson) Sharing maps have been removed. It's marginal usefulness in a
threads design can be worked around in other ways. Both #12 and #13
were done to simplify the code and improve readability and maintain-
ability. (As were most all of these changes)
TODO:
1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing
this will reduce the vnode pager to a mere fraction of its current size.
2) Rewrite vm_fault and the swap/vnode pagers to use the clustering
information provided by the new haspage pager interface. This will
substantially reduce the overhead by eliminating a large number of
VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be
improved to provide both a "behind" and "ahead" indication of
contiguousness.
3) Implement the extended features of pager_haspage in swap_pager_haspage().
It currently just says 0 pages ahead/behind.
4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps
via a much more general mechanism that could also be used for disk
striping of regular filesystems.
5) Do something to improve the architecture of vm_object_collapse(). The
fact that it makes calls into the swap pager and knows too much about
how the swap pager operates really bothers me. It also doesn't allow
for collapsing of non-swap pager objects ("unnamed" objects backed by
other pagers).
1995-07-13 08:48:48 +00:00
|
|
|
/*
|
|
|
|
* Misc constants.
|
|
|
|
*/
|
|
|
|
#define ACT_DECLINE 1
|
|
|
|
#define ACT_ADVANCE 3
|
This commit does a couple of things:
Re-enables the RSS limiting, and the routine is now tail-recursive,
making it much more safe (eliminates the possiblity of kernel stack
overflow.) Also, the RSS limiting is a little more intelligent about
finding the likely objects that are pushing the process over the limit.
Added some sysctls that help with VM system tuning.
New sysctl features:
1) Enable/disable lru pageout algorithm.
vm.pageout_algorithm = 0, default algorithm that works
well, especially using X windows and heavy
memory loading. Can have adverse effects,
sometimes slowing down program loading.
vm.pageout_algorithm = 1, close to true LRU. Works much
better than clock, etc. Does not work as well as
the default algorithm in general. Certain memory
"malloc" type benchmarks work a little better with
this setting.
Please give me feedback on the performance results
associated with these.
2) Enable/disable swapping.
vm.swapping_enabled = 1, default.
vm.swapping_enabled = 0, useful for cases where swapping
degrades performance.
The config option "NO_SWAPPING" is still operative, and
takes precedence over the sysctl. If "NO_SWAPPING" is
specified, the sysctl still exists, but "vm.swapping_enabled"
is hard-wired to "0".
Each of these can be changed "on the fly."
1996-06-26 05:39:27 +00:00
|
|
|
#define ACT_INIT 5
|
1996-09-08 20:44:49 +00:00
|
|
|
#define ACT_MAX 64
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1999-12-29 05:07:58 +00:00
|
|
|
#ifdef _KERNEL
|
2007-05-05 19:50:28 +00:00
|
|
|
|
2012-08-03 01:48:15 +00:00
|
|
|
#include <sys/systm.h>
|
|
|
|
|
|
|
|
#include <machine/atomic.h>
|
2007-05-05 19:50:28 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
2012-11-17 04:00:19 +00:00
|
|
|
* Each pageable resident page falls into one of four lists:
|
1994-05-24 10:09:53 +00:00
|
|
|
*
|
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
|
|
|
* free
|
1994-05-24 10:09:53 +00:00
|
|
|
* Available for allocation now.
|
NOTE: libkvm, w, ps, 'top', and any other utility which depends on struct
proc or any VM system structure will have to be rebuilt!!!
Much needed overhaul of the VM system. Included in this first round of
changes:
1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages,
haspage, and sync operations are supported. The haspage interface now
provides information about clusterability. All pager routines now take
struct vm_object's instead of "pagers".
2) Improved data structures. In the previous paradigm, there is constant
confusion caused by pagers being both a data structure ("allocate a
pager") and a collection of routines. The idea of a pager structure has
escentially been eliminated. Objects now have types, and this type is
used to index the appropriate pager. In most cases, items in the pager
structure were duplicated in the object data structure and thus were
unnecessary. In the few cases that remained, a un_pager structure union
was created in the object to contain these items.
3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now
be removed. For instance, vm_object_enter(), vm_object_lookup(),
vm_object_remove(), and the associated object hash list were some of the
things that were removed.
4) simple_lock's removed. Discussion with several people reveals that the
SMP locking primitives used in the VM system aren't likely the mechanism
that we'll be adopting. Even if it were, the locking that was in the code
was very inadequate and would have to be mostly re-done anyway. The
locking in a uni-processor kernel was a no-op but went a long way toward
making the code difficult to read and debug.
5) Places that attempted to kludge-up the fact that we don't have kernel
thread support have been fixed to reflect the reality that we are really
dealing with processes, not threads. The VM system didn't have complete
thread support, so the comments and mis-named routines were just wrong.
We now use tsleep and wakeup directly in the lock routines, for instance.
6) Where appropriate, the pagers have been improved, especially in the
pager_alloc routines. Most of the pager_allocs have been rewritten and
are now faster and easier to maintain.
7) The pagedaemon pageout clustering algorithm has been rewritten and
now tries harder to output an even number of pages before and after
the requested page. This is sort of the reverse of the ideal pagein
algorithm and should provide better overall performance.
8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup
have been removed. Some other unnecessary casts have also been removed.
9) Some almost useless debugging code removed.
10) Terminology of shadow objects vs. backing objects straightened out.
The fact that the vm_object data structure escentially had this
backwards really confused things. The use of "shadow" and "backing
object" throughout the code is now internally consistent and correct
in the Mach terminology.
11) Several minor bug fixes, including one in the vm daemon that caused
0 RSS objects to not get purged as intended.
12) A "default pager" has now been created which cleans up the transition
of objects to the "swap" type. The previous checks throughout the code
for swp->pg_data != NULL were really ugly. This change also provides
the rudiments for future backing of "anonymous" memory by something
other than the swap pager (via the vnode pager, for example), and it
allows the decision about which of these pagers to use to be made
dynamically (although will need some additional decision code to do
this, of course).
13) (dyson) MAP_COPY has been deprecated and the corresponding "copy
object" code has been removed. MAP_COPY was undocumented and non-
standard. It was furthermore broken in several ways which caused its
behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will
continue to work correctly, but via the slightly different semantics
of MAP_PRIVATE.
14) (dyson) Sharing maps have been removed. It's marginal usefulness in a
threads design can be worked around in other ways. Both #12 and #13
were done to simplify the code and improve readability and maintain-
ability. (As were most all of these changes)
TODO:
1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing
this will reduce the vnode pager to a mere fraction of its current size.
2) Rewrite vm_fault and the swap/vnode pagers to use the clustering
information provided by the new haspage pager interface. This will
substantially reduce the overhead by eliminating a large number of
VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be
improved to provide both a "behind" and "ahead" indication of
contiguousness.
3) Implement the extended features of pager_haspage in swap_pager_haspage().
It currently just says 0 pages ahead/behind.
4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps
via a much more general mechanism that could also be used for disk
striping of regular filesystems.
5) Do something to improve the architecture of vm_object_collapse(). The
fact that it makes calls into the swap pager and knows too much about
how the swap pager operates really bothers me. It also doesn't allow
for collapsing of non-swap pager objects ("unnamed" objects backed by
other pagers).
1995-07-13 08:48:48 +00:00
|
|
|
*
|
|
|
|
* cache
|
Change the management of cached pages (PQ_CACHE) in two fundamental
ways:
(1) Cached pages are no longer kept in the object's resident page
splay tree and memq. Instead, they are kept in a separate per-object
splay tree of cached pages. However, access to this new per-object
splay tree is synchronized by the _free_ page queues lock, not to be
confused with the heavily contended page queues lock. Consequently, a
cached page can be reclaimed by vm_page_alloc(9) without acquiring the
object's lock or the page queues lock.
This solves a problem independently reported by tegge@ and Isilon.
Specifically, they observed the page daemon consuming a great deal of
CPU time because of pages bouncing back and forth between the cache
queue (PQ_CACHE) and the inactive queue (PQ_INACTIVE). The source of
this problem turned out to be a deadlock avoidance strategy employed
when selecting a cached page to reclaim in vm_page_select_cache().
However, the root cause was really that reclaiming a cached page
required the acquisition of an object lock while the page queues lock
was already held. Thus, this change addresses the problem at its
root, by eliminating the need to acquire the object's lock.
Moreover, keeping cached pages in the object's primary splay tree and
memq was, in effect, optimizing for the uncommon case. Cached pages
are reclaimed far, far more often than they are reactivated. Instead,
this change makes reclamation cheaper, especially in terms of
synchronization overhead, and reactivation more expensive, because
reactivated pages will have to be reentered into the object's primary
splay tree and memq.
(2) Cached pages are now stored alongside free pages in the physical
memory allocator's buddy queues, increasing the likelihood that large
allocations of contiguous physical memory (i.e., superpages) will
succeed.
Finally, as a result of this change long-standing restrictions on when
and where a cached page can be reclaimed and returned by
vm_page_alloc(9) are eliminated. Specifically, calls to
vm_page_alloc(9) specifying VM_ALLOC_INTERRUPT can now reclaim and
return a formerly cached page. Consequently, a call to malloc(9)
specifying M_NOWAIT is less likely to fail.
Discussed with: many over the course of the summer, including jeff@,
Justin Husted @ Isilon, peter@, tegge@
Tested by: an earlier version by kris@
Approved by: re (kensmith)
2007-09-25 06:25:06 +00:00
|
|
|
* Almost available for allocation. Still associated with
|
|
|
|
* an object, but clean and immediately freeable.
|
NOTE: libkvm, w, ps, 'top', and any other utility which depends on struct
proc or any VM system structure will have to be rebuilt!!!
Much needed overhaul of the VM system. Included in this first round of
changes:
1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages,
haspage, and sync operations are supported. The haspage interface now
provides information about clusterability. All pager routines now take
struct vm_object's instead of "pagers".
2) Improved data structures. In the previous paradigm, there is constant
confusion caused by pagers being both a data structure ("allocate a
pager") and a collection of routines. The idea of a pager structure has
escentially been eliminated. Objects now have types, and this type is
used to index the appropriate pager. In most cases, items in the pager
structure were duplicated in the object data structure and thus were
unnecessary. In the few cases that remained, a un_pager structure union
was created in the object to contain these items.
3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now
be removed. For instance, vm_object_enter(), vm_object_lookup(),
vm_object_remove(), and the associated object hash list were some of the
things that were removed.
4) simple_lock's removed. Discussion with several people reveals that the
SMP locking primitives used in the VM system aren't likely the mechanism
that we'll be adopting. Even if it were, the locking that was in the code
was very inadequate and would have to be mostly re-done anyway. The
locking in a uni-processor kernel was a no-op but went a long way toward
making the code difficult to read and debug.
5) Places that attempted to kludge-up the fact that we don't have kernel
thread support have been fixed to reflect the reality that we are really
dealing with processes, not threads. The VM system didn't have complete
thread support, so the comments and mis-named routines were just wrong.
We now use tsleep and wakeup directly in the lock routines, for instance.
6) Where appropriate, the pagers have been improved, especially in the
pager_alloc routines. Most of the pager_allocs have been rewritten and
are now faster and easier to maintain.
7) The pagedaemon pageout clustering algorithm has been rewritten and
now tries harder to output an even number of pages before and after
the requested page. This is sort of the reverse of the ideal pagein
algorithm and should provide better overall performance.
8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup
have been removed. Some other unnecessary casts have also been removed.
9) Some almost useless debugging code removed.
10) Terminology of shadow objects vs. backing objects straightened out.
The fact that the vm_object data structure escentially had this
backwards really confused things. The use of "shadow" and "backing
object" throughout the code is now internally consistent and correct
in the Mach terminology.
11) Several minor bug fixes, including one in the vm daemon that caused
0 RSS objects to not get purged as intended.
12) A "default pager" has now been created which cleans up the transition
of objects to the "swap" type. The previous checks throughout the code
for swp->pg_data != NULL were really ugly. This change also provides
the rudiments for future backing of "anonymous" memory by something
other than the swap pager (via the vnode pager, for example), and it
allows the decision about which of these pagers to use to be made
dynamically (although will need some additional decision code to do
this, of course).
13) (dyson) MAP_COPY has been deprecated and the corresponding "copy
object" code has been removed. MAP_COPY was undocumented and non-
standard. It was furthermore broken in several ways which caused its
behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will
continue to work correctly, but via the slightly different semantics
of MAP_PRIVATE.
14) (dyson) Sharing maps have been removed. It's marginal usefulness in a
threads design can be worked around in other ways. Both #12 and #13
were done to simplify the code and improve readability and maintain-
ability. (As were most all of these changes)
TODO:
1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing
this will reduce the vnode pager to a mere fraction of its current size.
2) Rewrite vm_fault and the swap/vnode pagers to use the clustering
information provided by the new haspage pager interface. This will
substantially reduce the overhead by eliminating a large number of
VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be
improved to provide both a "behind" and "ahead" indication of
contiguousness.
3) Implement the extended features of pager_haspage in swap_pager_haspage().
It currently just says 0 pages ahead/behind.
4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps
via a much more general mechanism that could also be used for disk
striping of regular filesystems.
5) Do something to improve the architecture of vm_object_collapse(). The
fact that it makes calls into the swap pager and knows too much about
how the swap pager operates really bothers me. It also doesn't allow
for collapsing of non-swap pager objects ("unnamed" objects backed by
other pagers).
1995-07-13 08:48:48 +00:00
|
|
|
*
|
2007-07-13 04:42:20 +00:00
|
|
|
* The following lists are LRU sorted:
|
|
|
|
*
|
1994-05-24 10:09:53 +00:00
|
|
|
* inactive
|
1996-01-30 23:02:38 +00:00
|
|
|
* Low activity, candidates for reclamation.
|
1994-05-24 10:09:53 +00:00
|
|
|
* This is the list of pages that should be
|
|
|
|
* paged out next.
|
NOTE: libkvm, w, ps, 'top', and any other utility which depends on struct
proc or any VM system structure will have to be rebuilt!!!
Much needed overhaul of the VM system. Included in this first round of
changes:
1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages,
haspage, and sync operations are supported. The haspage interface now
provides information about clusterability. All pager routines now take
struct vm_object's instead of "pagers".
2) Improved data structures. In the previous paradigm, there is constant
confusion caused by pagers being both a data structure ("allocate a
pager") and a collection of routines. The idea of a pager structure has
escentially been eliminated. Objects now have types, and this type is
used to index the appropriate pager. In most cases, items in the pager
structure were duplicated in the object data structure and thus were
unnecessary. In the few cases that remained, a un_pager structure union
was created in the object to contain these items.
3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now
be removed. For instance, vm_object_enter(), vm_object_lookup(),
vm_object_remove(), and the associated object hash list were some of the
things that were removed.
4) simple_lock's removed. Discussion with several people reveals that the
SMP locking primitives used in the VM system aren't likely the mechanism
that we'll be adopting. Even if it were, the locking that was in the code
was very inadequate and would have to be mostly re-done anyway. The
locking in a uni-processor kernel was a no-op but went a long way toward
making the code difficult to read and debug.
5) Places that attempted to kludge-up the fact that we don't have kernel
thread support have been fixed to reflect the reality that we are really
dealing with processes, not threads. The VM system didn't have complete
thread support, so the comments and mis-named routines were just wrong.
We now use tsleep and wakeup directly in the lock routines, for instance.
6) Where appropriate, the pagers have been improved, especially in the
pager_alloc routines. Most of the pager_allocs have been rewritten and
are now faster and easier to maintain.
7) The pagedaemon pageout clustering algorithm has been rewritten and
now tries harder to output an even number of pages before and after
the requested page. This is sort of the reverse of the ideal pagein
algorithm and should provide better overall performance.
8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup
have been removed. Some other unnecessary casts have also been removed.
9) Some almost useless debugging code removed.
10) Terminology of shadow objects vs. backing objects straightened out.
The fact that the vm_object data structure escentially had this
backwards really confused things. The use of "shadow" and "backing
object" throughout the code is now internally consistent and correct
in the Mach terminology.
11) Several minor bug fixes, including one in the vm daemon that caused
0 RSS objects to not get purged as intended.
12) A "default pager" has now been created which cleans up the transition
of objects to the "swap" type. The previous checks throughout the code
for swp->pg_data != NULL were really ugly. This change also provides
the rudiments for future backing of "anonymous" memory by something
other than the swap pager (via the vnode pager, for example), and it
allows the decision about which of these pagers to use to be made
dynamically (although will need some additional decision code to do
this, of course).
13) (dyson) MAP_COPY has been deprecated and the corresponding "copy
object" code has been removed. MAP_COPY was undocumented and non-
standard. It was furthermore broken in several ways which caused its
behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will
continue to work correctly, but via the slightly different semantics
of MAP_PRIVATE.
14) (dyson) Sharing maps have been removed. It's marginal usefulness in a
threads design can be worked around in other ways. Both #12 and #13
were done to simplify the code and improve readability and maintain-
ability. (As were most all of these changes)
TODO:
1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing
this will reduce the vnode pager to a mere fraction of its current size.
2) Rewrite vm_fault and the swap/vnode pagers to use the clustering
information provided by the new haspage pager interface. This will
substantially reduce the overhead by eliminating a large number of
VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be
improved to provide both a "behind" and "ahead" indication of
contiguousness.
3) Implement the extended features of pager_haspage in swap_pager_haspage().
It currently just says 0 pages ahead/behind.
4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps
via a much more general mechanism that could also be used for disk
striping of regular filesystems.
5) Do something to improve the architecture of vm_object_collapse(). The
fact that it makes calls into the swap pager and knows too much about
how the swap pager operates really bothers me. It also doesn't allow
for collapsing of non-swap pager objects ("unnamed" objects backed by
other pagers).
1995-07-13 08:48:48 +00:00
|
|
|
*
|
1994-05-24 10:09:53 +00:00
|
|
|
* active
|
NOTE: libkvm, w, ps, 'top', and any other utility which depends on struct
proc or any VM system structure will have to be rebuilt!!!
Much needed overhaul of the VM system. Included in this first round of
changes:
1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages,
haspage, and sync operations are supported. The haspage interface now
provides information about clusterability. All pager routines now take
struct vm_object's instead of "pagers".
2) Improved data structures. In the previous paradigm, there is constant
confusion caused by pagers being both a data structure ("allocate a
pager") and a collection of routines. The idea of a pager structure has
escentially been eliminated. Objects now have types, and this type is
used to index the appropriate pager. In most cases, items in the pager
structure were duplicated in the object data structure and thus were
unnecessary. In the few cases that remained, a un_pager structure union
was created in the object to contain these items.
3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now
be removed. For instance, vm_object_enter(), vm_object_lookup(),
vm_object_remove(), and the associated object hash list were some of the
things that were removed.
4) simple_lock's removed. Discussion with several people reveals that the
SMP locking primitives used in the VM system aren't likely the mechanism
that we'll be adopting. Even if it were, the locking that was in the code
was very inadequate and would have to be mostly re-done anyway. The
locking in a uni-processor kernel was a no-op but went a long way toward
making the code difficult to read and debug.
5) Places that attempted to kludge-up the fact that we don't have kernel
thread support have been fixed to reflect the reality that we are really
dealing with processes, not threads. The VM system didn't have complete
thread support, so the comments and mis-named routines were just wrong.
We now use tsleep and wakeup directly in the lock routines, for instance.
6) Where appropriate, the pagers have been improved, especially in the
pager_alloc routines. Most of the pager_allocs have been rewritten and
are now faster and easier to maintain.
7) The pagedaemon pageout clustering algorithm has been rewritten and
now tries harder to output an even number of pages before and after
the requested page. This is sort of the reverse of the ideal pagein
algorithm and should provide better overall performance.
8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup
have been removed. Some other unnecessary casts have also been removed.
9) Some almost useless debugging code removed.
10) Terminology of shadow objects vs. backing objects straightened out.
The fact that the vm_object data structure escentially had this
backwards really confused things. The use of "shadow" and "backing
object" throughout the code is now internally consistent and correct
in the Mach terminology.
11) Several minor bug fixes, including one in the vm daemon that caused
0 RSS objects to not get purged as intended.
12) A "default pager" has now been created which cleans up the transition
of objects to the "swap" type. The previous checks throughout the code
for swp->pg_data != NULL were really ugly. This change also provides
the rudiments for future backing of "anonymous" memory by something
other than the swap pager (via the vnode pager, for example), and it
allows the decision about which of these pagers to use to be made
dynamically (although will need some additional decision code to do
this, of course).
13) (dyson) MAP_COPY has been deprecated and the corresponding "copy
object" code has been removed. MAP_COPY was undocumented and non-
standard. It was furthermore broken in several ways which caused its
behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will
continue to work correctly, but via the slightly different semantics
of MAP_PRIVATE.
14) (dyson) Sharing maps have been removed. It's marginal usefulness in a
threads design can be worked around in other ways. Both #12 and #13
were done to simplify the code and improve readability and maintain-
ability. (As were most all of these changes)
TODO:
1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing
this will reduce the vnode pager to a mere fraction of its current size.
2) Rewrite vm_fault and the swap/vnode pagers to use the clustering
information provided by the new haspage pager interface. This will
substantially reduce the overhead by eliminating a large number of
VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be
improved to provide both a "behind" and "ahead" indication of
contiguousness.
3) Implement the extended features of pager_haspage in swap_pager_haspage().
It currently just says 0 pages ahead/behind.
4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps
via a much more general mechanism that could also be used for disk
striping of regular filesystems.
5) Do something to improve the architecture of vm_object_collapse(). The
fact that it makes calls into the swap pager and knows too much about
how the swap pager operates really bothers me. It also doesn't allow
for collapsing of non-swap pager objects ("unnamed" objects backed by
other pagers).
1995-07-13 08:48:48 +00:00
|
|
|
* Pages that are "active" i.e. they have been
|
|
|
|
* recently referenced.
|
1995-09-03 20:11:26 +00:00
|
|
|
*
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
|
1995-12-11 04:58:34 +00:00
|
|
|
extern int vm_page_zero_count;
|
|
|
|
|
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
|
|
|
extern vm_page_t vm_page_array; /* First resident page in table */
|
2012-05-12 20:10:18 +00:00
|
|
|
extern long vm_page_array_size; /* number of vm_page_t's */
|
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
|
|
|
extern long first_page; /* first physical page number */
|
|
|
|
|
Enable the new physical memory allocator.
This allocator uses a binary buddy system with a twist. First and
foremost, this allocator is required to support the implementation of
superpages. As a side effect, it enables a more robust implementation
of contigmalloc(9). Moreover, this reimplementation of
contigmalloc(9) eliminates the acquisition of Giant by
contigmalloc(..., M_NOWAIT, ...).
The twist is that this allocator tries to reduce the number of TLB
misses incurred by accesses through a direct map to small, UMA-managed
objects and page table pages. Roughly speaking, the physical pages
that are allocated for such purposes are clustered together in the
physical address space. The performance benefits vary. In the most
extreme case, a uniprocessor kernel running on an Opteron, I measured
an 18% reduction in system time during a buildworld.
This allocator does not implement page coloring. The reason is that
superpages have much the same effect. The contiguous physical memory
allocation necessary for a superpage is inherently colored.
Finally, the one caveat is that this allocator does not effectively
support prezeroed pages. I hope this is temporary. On i386, this is
a slight pessimization. However, on amd64, the beneficial effects of
the direct-map optimization outweigh the ill effects. I speculate
that this is true in general of machines with a direct map.
Approved by: re
2007-06-16 04:57:06 +00:00
|
|
|
#define VM_PAGE_IS_FREE(m) (((m)->flags & PG_FREE) != 0)
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#define VM_PAGE_TO_PHYS(entry) ((entry)->phys_addr)
|
|
|
|
|
2012-05-12 20:42:56 +00:00
|
|
|
vm_page_t PHYS_TO_VM_PAGE(vm_paddr_t pa);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2002-07-18 04:08:10 +00:00
|
|
|
/* page allocation classes: */
|
1998-02-05 03:32:49 +00:00
|
|
|
#define VM_ALLOC_NORMAL 0
|
|
|
|
#define VM_ALLOC_INTERRUPT 1
|
|
|
|
#define VM_ALLOC_SYSTEM 2
|
2002-07-18 04:08:10 +00:00
|
|
|
#define VM_ALLOC_CLASS_MASK 3
|
|
|
|
/* page allocation flags: */
|
2002-11-01 00:59:03 +00:00
|
|
|
#define VM_ALLOC_WIRED 0x0020 /* non pageable */
|
|
|
|
#define VM_ALLOC_ZERO 0x0040 /* Try to obtain a zeroed page */
|
|
|
|
#define VM_ALLOC_NOOBJ 0x0100 /* No associated object */
|
2004-10-24 06:15:36 +00:00
|
|
|
#define VM_ALLOC_NOBUSY 0x0200 /* Do not busy the page */
|
Change the management of cached pages (PQ_CACHE) in two fundamental
ways:
(1) Cached pages are no longer kept in the object's resident page
splay tree and memq. Instead, they are kept in a separate per-object
splay tree of cached pages. However, access to this new per-object
splay tree is synchronized by the _free_ page queues lock, not to be
confused with the heavily contended page queues lock. Consequently, a
cached page can be reclaimed by vm_page_alloc(9) without acquiring the
object's lock or the page queues lock.
This solves a problem independently reported by tegge@ and Isilon.
Specifically, they observed the page daemon consuming a great deal of
CPU time because of pages bouncing back and forth between the cache
queue (PQ_CACHE) and the inactive queue (PQ_INACTIVE). The source of
this problem turned out to be a deadlock avoidance strategy employed
when selecting a cached page to reclaim in vm_page_select_cache().
However, the root cause was really that reclaiming a cached page
required the acquisition of an object lock while the page queues lock
was already held. Thus, this change addresses the problem at its
root, by eliminating the need to acquire the object's lock.
Moreover, keeping cached pages in the object's primary splay tree and
memq was, in effect, optimizing for the uncommon case. Cached pages
are reclaimed far, far more often than they are reactivated. Instead,
this change makes reclamation cheaper, especially in terms of
synchronization overhead, and reactivation more expensive, because
reactivated pages will have to be reentered into the object's primary
splay tree and memq.
(2) Cached pages are now stored alongside free pages in the physical
memory allocator's buddy queues, increasing the likelihood that large
allocations of contiguous physical memory (i.e., superpages) will
succeed.
Finally, as a result of this change long-standing restrictions on when
and where a cached page can be reclaimed and returned by
vm_page_alloc(9) are eliminated. Specifically, calls to
vm_page_alloc(9) specifying VM_ALLOC_INTERRUPT can now reclaim and
return a formerly cached page. Consequently, a call to malloc(9)
specifying M_NOWAIT is less likely to fail.
Discussed with: many over the course of the summer, including jeff@,
Justin Husted @ Isilon, peter@, tegge@
Tested by: an earlier version by kris@
Approved by: re (kensmith)
2007-09-25 06:25:06 +00:00
|
|
|
#define VM_ALLOC_IFCACHED 0x0400 /* Fail if the page is not cached */
|
|
|
|
#define VM_ALLOC_IFNOTCACHED 0x0800 /* Fail if the page is cached */
|
2010-07-05 21:13:32 +00:00
|
|
|
#define VM_ALLOC_IGN_SBUSY 0x1000 /* vm_page_grab() only */
|
2012-01-27 20:18:31 +00:00
|
|
|
#define VM_ALLOC_NODUMP 0x2000 /* don't include in dump */
|
2013-08-09 11:11:11 +00:00
|
|
|
#define VM_ALLOC_SBUSY 0x4000 /* Shared busy the page */
|
2010-07-05 21:13:32 +00:00
|
|
|
|
|
|
|
#define VM_ALLOC_COUNT_SHIFT 16
|
|
|
|
#define VM_ALLOC_COUNT(count) ((count) << VM_ALLOC_COUNT_SHIFT)
|
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
|
|
|
|
2012-11-14 20:01:40 +00:00
|
|
|
#ifdef M_NOWAIT
|
|
|
|
static inline int
|
|
|
|
malloc2vm_flags(int malloc_flags)
|
|
|
|
{
|
|
|
|
int pflags;
|
|
|
|
|
2012-11-16 05:49:56 +00:00
|
|
|
KASSERT((malloc_flags & M_USE_RESERVE) == 0 ||
|
|
|
|
(malloc_flags & M_NOWAIT) != 0,
|
|
|
|
("M_USE_RESERVE requires M_NOWAIT"));
|
2012-11-14 20:01:40 +00:00
|
|
|
pflags = (malloc_flags & M_USE_RESERVE) != 0 ? VM_ALLOC_INTERRUPT :
|
|
|
|
VM_ALLOC_SYSTEM;
|
|
|
|
if ((malloc_flags & M_ZERO) != 0)
|
|
|
|
pflags |= VM_ALLOC_ZERO;
|
|
|
|
if ((malloc_flags & M_NODUMP) != 0)
|
|
|
|
pflags |= VM_ALLOC_NODUMP;
|
|
|
|
return (pflags);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2013-08-09 11:11:11 +00:00
|
|
|
void vm_page_busy_downgrade(vm_page_t m);
|
|
|
|
void vm_page_busy_sleep(vm_page_t m, const char *msg);
|
2001-07-04 20:15:18 +00:00
|
|
|
void vm_page_flash(vm_page_t m);
|
|
|
|
void vm_page_hold(vm_page_t mem);
|
|
|
|
void vm_page_unhold(vm_page_t mem);
|
|
|
|
void vm_page_free(vm_page_t m);
|
|
|
|
void vm_page_free_zero(vm_page_t m);
|
|
|
|
|
|
|
|
void vm_page_activate (vm_page_t);
|
2013-06-10 01:48:21 +00:00
|
|
|
void vm_page_advise(vm_page_t m, int advice);
|
2001-07-04 20:15:18 +00:00
|
|
|
vm_page_t vm_page_alloc (vm_object_t, vm_pindex_t, int);
|
Refactor the code that performs physically contiguous memory allocation,
yielding a new public interface, vm_page_alloc_contig(). This new function
addresses some of the limitations of the current interfaces, contigmalloc()
and kmem_alloc_contig(). For example, the physically contiguous memory that
is allocated with those interfaces can only be allocated to the kernel vm
object and must be mapped into the kernel virtual address space. It also
provides functionality that vm_phys_alloc_contig() doesn't, such as wiring
the returned pages. Moreover, unlike that function, it respects the low
water marks on the paging queues and wakes up the page daemon when
necessary. That said, at present, this new function can't be applied to all
types of vm objects. However, that restriction will be eliminated in the
coming weeks.
From a design standpoint, this change also addresses an inconsistency
between vm_phys_alloc_contig() and the other vm_phys_alloc*() functions.
Specifically, vm_phys_alloc_contig() manipulated vm_page fields that other
functions in vm/vm_phys.c didn't. Moreover, vm_phys_alloc_contig() knew
about vnodes and reservations. Now, vm_page_alloc_contig() is responsible
for these things.
Reviewed by: kib
Discussed with: jhb
2011-11-16 16:46:09 +00:00
|
|
|
vm_page_t vm_page_alloc_contig(vm_object_t object, vm_pindex_t pindex, int req,
|
|
|
|
u_long npages, vm_paddr_t low, vm_paddr_t high, u_long alignment,
|
|
|
|
vm_paddr_t boundary, vm_memattr_t memattr);
|
2010-11-28 05:51:31 +00:00
|
|
|
vm_page_t vm_page_alloc_freelist(int, int);
|
2001-07-04 20:15:18 +00:00
|
|
|
vm_page_t vm_page_grab (vm_object_t, vm_pindex_t, int);
|
2009-05-30 22:15:55 +00:00
|
|
|
void vm_page_cache(vm_page_t);
|
2013-02-06 14:29:05 +00:00
|
|
|
void vm_page_cache_free(vm_object_t, vm_pindex_t, vm_pindex_t);
|
2013-02-05 00:09:33 +00:00
|
|
|
void vm_page_cache_transfer(vm_object_t, vm_pindex_t, vm_object_t);
|
2001-07-04 20:15:18 +00:00
|
|
|
int vm_page_try_to_cache (vm_page_t);
|
|
|
|
int vm_page_try_to_free (vm_page_t);
|
|
|
|
void vm_page_deactivate (vm_page_t);
|
2012-11-13 02:50:39 +00:00
|
|
|
void vm_page_dequeue(vm_page_t m);
|
|
|
|
void vm_page_dequeue_locked(vm_page_t m);
|
2010-07-04 11:13:33 +00:00
|
|
|
vm_page_t vm_page_find_least(vm_object_t, vm_pindex_t);
|
2011-03-11 07:07:48 +00:00
|
|
|
vm_page_t vm_page_getfake(vm_paddr_t paddr, vm_memattr_t memattr);
|
2012-05-12 20:34:22 +00:00
|
|
|
void vm_page_initfake(vm_page_t m, vm_paddr_t paddr, vm_memattr_t memattr);
|
2013-08-09 11:28:55 +00:00
|
|
|
int vm_page_insert (vm_page_t, vm_object_t, vm_pindex_t);
|
2013-02-06 14:50:34 +00:00
|
|
|
boolean_t vm_page_is_cached(vm_object_t object, vm_pindex_t pindex);
|
2001-07-04 20:15:18 +00:00
|
|
|
vm_page_t vm_page_lookup (vm_object_t, vm_pindex_t);
|
2010-06-21 23:27:24 +00:00
|
|
|
vm_page_t vm_page_next(vm_page_t m);
|
2010-04-30 00:46:43 +00:00
|
|
|
int vm_page_pa_tryrelock(pmap_t, vm_paddr_t, vm_paddr_t *);
|
Split the pagequeues per NUMA domains, and split pageademon process
into threads each processing queue in a single domain. The structure
of the pagedaemons and queues is kept intact, most of the changes come
from the need for code to find an owning page queue for given page,
calculated from the segment containing the page.
The tie between NUMA domain and pagedaemon thread/pagequeue split is
rather arbitrary, the multithreaded daemon could be allowed for the
single-domain machines, or one domain might be split into several page
domains, to further increase concurrency.
Right now, each pagedaemon thread tries to reach the global target,
precalculated at the start of the pass. This is not optimal, since it
could cause excessive page deactivation and freeing. The code should
be changed to re-check the global page deficit state in the loop after
some number of iterations.
The pagedaemons reach the quorum before starting the OOM, since one
thread inability to meet the target is normal for split queues. Only
when all pagedaemons fail to produce enough reusable pages, OOM is
started by single selected thread.
Launder is modified to take into account the segments layout with
regard to the region for which cleaning is performed.
Based on the preliminary patch by jeff, sponsored by EMC / Isilon
Storage Division.
Reviewed by: alc
Tested by: pho
Sponsored by: The FreeBSD Foundation
2013-08-07 16:36:38 +00:00
|
|
|
struct vm_pagequeue *vm_page_pagequeue(vm_page_t m);
|
2010-06-21 23:27:24 +00:00
|
|
|
vm_page_t vm_page_prev(vm_page_t m);
|
2011-03-11 07:07:48 +00:00
|
|
|
void vm_page_putfake(vm_page_t m);
|
2012-08-14 11:45:47 +00:00
|
|
|
void vm_page_readahead_finish(vm_page_t m);
|
2011-09-06 10:30:11 +00:00
|
|
|
void vm_page_reference(vm_page_t m);
|
2001-07-04 20:15:18 +00:00
|
|
|
void vm_page_remove (vm_page_t);
|
2013-08-09 11:28:55 +00:00
|
|
|
int vm_page_rename (vm_page_t, vm_object_t, vm_pindex_t);
|
|
|
|
vm_page_t vm_page_replace(vm_page_t mnew, vm_object_t object,
|
|
|
|
vm_pindex_t pindex);
|
2008-03-19 20:24:35 +00:00
|
|
|
void vm_page_requeue(vm_page_t m);
|
2012-11-13 02:50:39 +00:00
|
|
|
void vm_page_requeue_locked(vm_page_t m);
|
2013-08-09 11:11:11 +00:00
|
|
|
int vm_page_sbusied(vm_page_t m);
|
2011-11-30 17:39:00 +00:00
|
|
|
void vm_page_set_valid_range(vm_page_t m, int base, int size);
|
2013-08-09 11:11:11 +00:00
|
|
|
int vm_page_sleep_if_busy(vm_page_t m, const char *msg);
|
2004-04-04 23:33:36 +00:00
|
|
|
vm_offset_t vm_page_startup(vm_offset_t vaddr);
|
2013-08-09 11:11:11 +00:00
|
|
|
void vm_page_sunbusy(vm_page_t m);
|
|
|
|
int vm_page_trysbusy(vm_page_t m);
|
2010-12-17 22:41:22 +00:00
|
|
|
void vm_page_unhold_pages(vm_page_t *ma, int count);
|
2001-07-04 20:15:18 +00:00
|
|
|
void vm_page_unwire (vm_page_t, int);
|
2011-03-11 07:07:48 +00:00
|
|
|
void vm_page_updatefake(vm_page_t m, vm_paddr_t paddr, vm_memattr_t memattr);
|
2001-07-04 20:15:18 +00:00
|
|
|
void vm_page_wire (vm_page_t);
|
2013-08-09 11:11:11 +00:00
|
|
|
void vm_page_xunbusy_hard(vm_page_t m);
|
2001-07-04 20:15:18 +00:00
|
|
|
void vm_page_set_validclean (vm_page_t, int, int);
|
|
|
|
void vm_page_clear_dirty (vm_page_t, int, int);
|
|
|
|
void vm_page_set_invalid (vm_page_t, int, int);
|
|
|
|
int vm_page_is_valid (vm_page_t, int, int);
|
|
|
|
void vm_page_test_dirty (vm_page_t);
|
2011-11-08 11:08:40 +00:00
|
|
|
vm_page_bits_t vm_page_bits(int base, int size);
|
1999-04-05 19:38:30 +00:00
|
|
|
void vm_page_zero_invalid(vm_page_t m, boolean_t setvalid);
|
1999-02-08 00:37:36 +00:00
|
|
|
void vm_page_free_toq(vm_page_t m);
|
2001-08-25 05:00:44 +00:00
|
|
|
void vm_page_zero_idle_wakeup(void);
|
At long last, commit the zero copy sockets code.
MAKEDEV: Add MAKEDEV glue for the ti(4) device nodes.
ti.4: Update the ti(4) man page to include information on the
TI_JUMBO_HDRSPLIT and TI_PRIVATE_JUMBOS kernel options,
and also include information about the new character
device interface and the associated ioctls.
man9/Makefile: Add jumbo.9 and zero_copy.9 man pages and associated
links.
jumbo.9: New man page describing the jumbo buffer allocator
interface and operation.
zero_copy.9: New man page describing the general characteristics of
the zero copy send and receive code, and what an
application author should do to take advantage of the
zero copy functionality.
NOTES: Add entries for ZERO_COPY_SOCKETS, TI_PRIVATE_JUMBOS,
TI_JUMBO_HDRSPLIT, MSIZE, and MCLSHIFT.
conf/files: Add uipc_jumbo.c and uipc_cow.c.
conf/options: Add the 5 options mentioned above.
kern_subr.c: Receive side zero copy implementation. This takes
"disposable" pages attached to an mbuf, gives them to
a user process, and then recycles the user's page.
This is only active when ZERO_COPY_SOCKETS is turned on
and the kern.ipc.zero_copy.receive sysctl variable is
set to 1.
uipc_cow.c: Send side zero copy functions. Takes a page written
by the user and maps it copy on write and assigns it
kernel virtual address space. Removes copy on write
mapping once the buffer has been freed by the network
stack.
uipc_jumbo.c: Jumbo disposable page allocator code. This allocates
(optionally) disposable pages for network drivers that
want to give the user the option of doing zero copy
receive.
uipc_socket.c: Add kern.ipc.zero_copy.{send,receive} sysctls that are
enabled if ZERO_COPY_SOCKETS is turned on.
Add zero copy send support to sosend() -- pages get
mapped into the kernel instead of getting copied if
they meet size and alignment restrictions.
uipc_syscalls.c:Un-staticize some of the sf* functions so that they
can be used elsewhere. (uipc_cow.c)
if_media.c: In the SIOCGIFMEDIA ioctl in ifmedia_ioctl(), avoid
calling malloc() with M_WAITOK. Return an error if
the M_NOWAIT malloc fails.
The ti(4) driver and the wi(4) driver, at least, call
this with a mutex held. This causes witness warnings
for 'ifconfig -a' with a wi(4) or ti(4) board in the
system. (I've only verified for ti(4)).
ip_output.c: Fragment large datagrams so that each segment contains
a multiple of PAGE_SIZE amount of data plus headers.
This allows the receiver to potentially do page
flipping on receives.
if_ti.c: Add zero copy receive support to the ti(4) driver. If
TI_PRIVATE_JUMBOS is not defined, it now uses the
jumbo(9) buffer allocator for jumbo receive buffers.
Add a new character device interface for the ti(4)
driver for the new debugging interface. This allows
(a patched version of) gdb to talk to the Tigon board
and debug the firmware. There are also a few additional
debugging ioctls available through this interface.
Add header splitting support to the ti(4) driver.
Tweak some of the default interrupt coalescing
parameters to more useful defaults.
Add hooks for supporting transmit flow control, but
leave it turned off with a comment describing why it
is turned off.
if_tireg.h: Change the firmware rev to 12.4.11, since we're really
at 12.4.11 plus fixes from 12.4.13.
Add defines needed for debugging.
Remove the ti_stats structure, it is now defined in
sys/tiio.h.
ti_fw.h: 12.4.11 firmware.
ti_fw2.h: 12.4.11 firmware, plus selected fixes from 12.4.13,
and my header splitting patches. Revision 12.4.13
doesn't handle 10/100 negotiation properly. (This
firmware is the same as what was in the tree previously,
with the addition of header splitting support.)
sys/jumbo.h: Jumbo buffer allocator interface.
sys/mbuf.h: Add a new external mbuf type, EXT_DISPOSABLE, to
indicate that the payload buffer can be thrown away /
flipped to a userland process.
socketvar.h: Add prototype for socow_setup.
tiio.h: ioctl interface to the character portion of the ti(4)
driver, plus associated structure/type definitions.
uio.h: Change prototype for uiomoveco() so that we'll know
whether the source page is disposable.
ufs_readwrite.c:Update for new prototype of uiomoveco().
vm_fault.c: In vm_fault(), check to see whether we need to do a page
based copy on write fault.
vm_object.c: Add a new function, vm_object_allocate_wait(). This
does the same thing that vm_object allocate does, except
that it gives the caller the opportunity to specify whether
it should wait on the uma_zalloc() of the object structre.
This allows vm objects to be allocated while holding a
mutex. (Without generating WITNESS warnings.)
vm_object_allocate() is implemented as a call to
vm_object_allocate_wait() with the malloc flag set to
M_WAITOK.
vm_object.h: Add prototype for vm_object_allocate_wait().
vm_page.c: Add page-based copy on write setup, clear and fault
routines.
vm_page.h: Add page based COW function prototypes and variable in
the vm_page structure.
Many thanks to Drew Gallatin, who wrote the zero copy send and receive
code, and to all the other folks who have tested and reviewed this code
over the years.
2002-06-26 03:37:47 +00:00
|
|
|
|
2012-06-20 23:25:47 +00:00
|
|
|
void vm_page_dirty_KBI(vm_page_t m);
|
2011-11-29 13:07:32 +00:00
|
|
|
void vm_page_lock_KBI(vm_page_t m, const char *file, int line);
|
|
|
|
void vm_page_unlock_KBI(vm_page_t m, const char *file, int line);
|
|
|
|
int vm_page_trylock_KBI(vm_page_t m, const char *file, int line);
|
|
|
|
#if defined(INVARIANTS) || defined(INVARIANT_SUPPORT)
|
2013-06-03 01:22:54 +00:00
|
|
|
void vm_page_assert_locked_KBI(vm_page_t m, const char *file, int line);
|
2011-11-29 13:07:32 +00:00
|
|
|
void vm_page_lock_assert_KBI(vm_page_t m, int a, const char *file, int line);
|
|
|
|
#endif
|
|
|
|
|
2013-08-09 11:11:11 +00:00
|
|
|
#define vm_page_assert_sbusied(m) \
|
|
|
|
KASSERT(vm_page_sbusied(m), \
|
|
|
|
("vm_page_assert_sbusied: page %p not shared busy @ %s:%d", \
|
|
|
|
(void *)m, __FILE__, __LINE__));
|
|
|
|
|
|
|
|
#define vm_page_assert_unbusied(m) \
|
|
|
|
KASSERT(!vm_page_busied(m), \
|
|
|
|
("vm_page_assert_unbusied: page %p busy @ %s:%d", \
|
|
|
|
(void *)m, __FILE__, __LINE__));
|
|
|
|
|
|
|
|
#define vm_page_assert_xbusied(m) \
|
|
|
|
KASSERT(vm_page_xbusied(m), \
|
|
|
|
("vm_page_assert_xbusied: page %p not exclusive busy @ %s:%d", \
|
|
|
|
(void *)m, __FILE__, __LINE__));
|
|
|
|
|
|
|
|
#define vm_page_busied(m) \
|
|
|
|
((m)->busy_lock != VPB_UNBUSIED)
|
|
|
|
|
|
|
|
#define vm_page_sbusy(m) do { \
|
|
|
|
if (!vm_page_trysbusy(m)) \
|
|
|
|
panic("%s: page %p failed shared busing", __func__, m); \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
#define vm_page_tryxbusy(m) \
|
|
|
|
(atomic_cmpset_acq_int(&m->busy_lock, VPB_UNBUSIED, \
|
|
|
|
VPB_SINGLE_EXCLUSIVER))
|
|
|
|
|
|
|
|
#define vm_page_xbusied(m) \
|
|
|
|
((m->busy_lock & VPB_SINGLE_EXCLUSIVER) != 0)
|
|
|
|
|
|
|
|
#define vm_page_xbusy(m) do { \
|
|
|
|
if (!vm_page_tryxbusy(m)) \
|
|
|
|
panic("%s: page %p failed exclusive busing", __func__, \
|
|
|
|
m); \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
#define vm_page_xunbusy(m) do { \
|
|
|
|
if (!atomic_cmpset_rel_int(&(m)->busy_lock, \
|
|
|
|
VPB_SINGLE_EXCLUSIVER, VPB_UNBUSIED)) \
|
|
|
|
vm_page_xunbusy_hard(m); \
|
|
|
|
} while (0)
|
|
|
|
|
2011-06-11 20:15:19 +00:00
|
|
|
#ifdef INVARIANTS
|
|
|
|
void vm_page_object_lock_assert(vm_page_t m);
|
|
|
|
#define VM_PAGE_OBJECT_LOCK_ASSERT(m) vm_page_object_lock_assert(m)
|
|
|
|
#else
|
|
|
|
#define VM_PAGE_OBJECT_LOCK_ASSERT(m) (void)0
|
|
|
|
#endif
|
|
|
|
|
2012-08-03 01:48:15 +00:00
|
|
|
/*
|
|
|
|
* We want to use atomic updates for the aflags field, which is 8 bits wide.
|
|
|
|
* However, not all architectures support atomic operations on 8-bit
|
|
|
|
* destinations. In order that we can easily use a 32-bit operation, we
|
|
|
|
* require that the aflags field be 32-bit aligned.
|
|
|
|
*/
|
|
|
|
CTASSERT(offsetof(struct vm_page, aflags) % sizeof(uint32_t) == 0);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Clear the given bits in the specified page.
|
|
|
|
*/
|
|
|
|
static inline void
|
|
|
|
vm_page_aflag_clear(vm_page_t m, uint8_t bits)
|
|
|
|
{
|
|
|
|
uint32_t *addr, val;
|
|
|
|
|
|
|
|
/*
|
2013-06-03 01:22:54 +00:00
|
|
|
* The PGA_REFERENCED flag can only be cleared if the page is locked.
|
2012-08-03 01:48:15 +00:00
|
|
|
*/
|
|
|
|
if ((bits & PGA_REFERENCED) != 0)
|
2013-06-03 01:22:54 +00:00
|
|
|
vm_page_assert_locked(m);
|
2012-08-03 01:48:15 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Access the whole 32-bit word containing the aflags field with an
|
|
|
|
* atomic update. Parallel non-atomic updates to the other fields
|
|
|
|
* within this word are handled properly by the atomic update.
|
|
|
|
*/
|
|
|
|
addr = (void *)&m->aflags;
|
|
|
|
KASSERT(((uintptr_t)addr & (sizeof(uint32_t) - 1)) == 0,
|
|
|
|
("vm_page_aflag_clear: aflags is misaligned"));
|
|
|
|
val = bits;
|
|
|
|
#if BYTE_ORDER == BIG_ENDIAN
|
|
|
|
val <<= 24;
|
|
|
|
#endif
|
|
|
|
atomic_clear_32(addr, val);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Set the given bits in the specified page.
|
|
|
|
*/
|
|
|
|
static inline void
|
|
|
|
vm_page_aflag_set(vm_page_t m, uint8_t bits)
|
|
|
|
{
|
|
|
|
uint32_t *addr, val;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The PGA_WRITEABLE flag can only be set if the page is managed and
|
2013-08-09 11:11:11 +00:00
|
|
|
* exclusive busied. Currently, this flag is only set by pmap_enter().
|
2012-08-03 01:48:15 +00:00
|
|
|
*/
|
|
|
|
KASSERT((bits & PGA_WRITEABLE) == 0 ||
|
2013-08-09 11:11:11 +00:00
|
|
|
((m->oflags & VPO_UNMANAGED) == 0 && vm_page_xbusied(m)),
|
|
|
|
("vm_page_aflag_set: PGA_WRITEABLE and not exclusive busy"));
|
2012-08-03 01:48:15 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Access the whole 32-bit word containing the aflags field with an
|
|
|
|
* atomic update. Parallel non-atomic updates to the other fields
|
|
|
|
* within this word are handled properly by the atomic update.
|
|
|
|
*/
|
|
|
|
addr = (void *)&m->aflags;
|
|
|
|
KASSERT(((uintptr_t)addr & (sizeof(uint32_t) - 1)) == 0,
|
|
|
|
("vm_page_aflag_set: aflags is misaligned"));
|
|
|
|
val = bits;
|
|
|
|
#if BYTE_ORDER == BIG_ENDIAN
|
|
|
|
val <<= 24;
|
|
|
|
#endif
|
|
|
|
atomic_set_32(addr, val);
|
|
|
|
}
|
|
|
|
|
2012-06-20 23:25:47 +00:00
|
|
|
/*
|
|
|
|
* vm_page_dirty:
|
|
|
|
*
|
|
|
|
* Set all bits in the page's dirty field.
|
|
|
|
*
|
|
|
|
* The object containing the specified page must be locked if the
|
|
|
|
* call is made from the machine-independent layer.
|
|
|
|
*
|
|
|
|
* See vm_page_clear_dirty_mask().
|
|
|
|
*/
|
|
|
|
static __inline void
|
|
|
|
vm_page_dirty(vm_page_t m)
|
|
|
|
{
|
|
|
|
|
|
|
|
/* Use vm_page_dirty_KBI() under INVARIANTS to save memory. */
|
|
|
|
#if defined(KLD_MODULE) || defined(INVARIANTS)
|
|
|
|
vm_page_dirty_KBI(m);
|
|
|
|
#else
|
|
|
|
m->dirty = VM_PAGE_BITS_ALL;
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2012-11-13 02:50:39 +00:00
|
|
|
/*
|
|
|
|
* vm_page_remque:
|
|
|
|
*
|
|
|
|
* If the given page is in a page queue, then remove it from that page
|
|
|
|
* queue.
|
|
|
|
*
|
|
|
|
* The page must be locked.
|
|
|
|
*/
|
|
|
|
static inline void
|
|
|
|
vm_page_remque(vm_page_t m)
|
|
|
|
{
|
|
|
|
|
|
|
|
if (m->queue != PQ_NONE)
|
|
|
|
vm_page_dequeue(m);
|
|
|
|
}
|
|
|
|
|
2002-10-20 19:57:55 +00:00
|
|
|
/*
|
|
|
|
* vm_page_undirty:
|
|
|
|
*
|
|
|
|
* Set page to not be dirty. Note: does not clear pmap modify bits
|
|
|
|
*/
|
|
|
|
static __inline void
|
|
|
|
vm_page_undirty(vm_page_t m)
|
|
|
|
{
|
2011-06-11 20:15:19 +00:00
|
|
|
|
|
|
|
VM_PAGE_OBJECT_LOCK_ASSERT(m);
|
2002-10-20 19:57:55 +00:00
|
|
|
m->dirty = 0;
|
|
|
|
}
|
|
|
|
|
1999-12-29 05:07:58 +00:00
|
|
|
#endif /* _KERNEL */
|
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
|
|
|
#endif /* !_VM_PAGE_ */
|