freebsd-dev/sys/kern/kern_malloc.c
David Greenman 0d94caffca 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

387 lines
10 KiB
C

/*
* Copyright (c) 1987, 1991, 1993
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 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.
*
* @(#)kern_malloc.c 8.3 (Berkeley) 1/4/94
* $Id: kern_malloc.c,v 1.6 1994/12/17 04:04:42 davidg Exp $
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/proc.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <vm/vm.h>
#include <vm/vm_kern.h>
struct kmembuckets bucket[MINBUCKET + 16];
struct kmemstats kmemstats[M_LAST];
struct kmemusage *kmemusage;
char *kmembase, *kmemlimit;
char *memname[] = INITKMEMNAMES;
#ifdef DIAGNOSTIC
/*
* This structure provides a set of masks to catch unaligned frees.
*/
long addrmask[] = { 0,
0x00000001, 0x00000003, 0x00000007, 0x0000000f,
0x0000001f, 0x0000003f, 0x0000007f, 0x000000ff,
0x000001ff, 0x000003ff, 0x000007ff, 0x00000fff,
0x00001fff, 0x00003fff, 0x00007fff, 0x0000ffff,
};
/*
* The WEIRD_ADDR is used as known text to copy into free objects so
* that modifications after frees can be detected.
*/
#define WEIRD_ADDR 0xdeadbeef
#define MAX_COPY 32
/*
* Normally the first word of the structure is used to hold the list
* pointer for free objects. However, when running with diagnostics,
* we use the third and fourth fields, so as to catch modifications
* in the most commonly trashed first two words.
*/
struct freelist {
long spare0;
short type;
long spare1;
caddr_t next;
};
#else /* !DIAGNOSTIC */
struct freelist {
caddr_t next;
};
#endif /* DIAGNOSTIC */
/*
* Allocate a block of memory
*/
void *
malloc(size, type, flags)
unsigned long size;
int type, flags;
{
register struct kmembuckets *kbp;
register struct kmemusage *kup;
register struct freelist *freep;
long indx, npg, allocsize;
int s;
caddr_t va, cp, savedlist;
#ifdef DIAGNOSTIC
long *end, *lp;
int copysize;
char *savedtype;
#endif
#ifdef KMEMSTATS
register struct kmemstats *ksp = &kmemstats[type];
if (((unsigned long)type) > M_LAST)
panic("malloc - bogus type");
#endif
indx = BUCKETINDX(size);
kbp = &bucket[indx];
s = splhigh();
#ifdef KMEMSTATS
while (ksp->ks_memuse >= ksp->ks_limit) {
if (flags & M_NOWAIT) {
splx(s);
return ((void *) NULL);
}
if (ksp->ks_limblocks < 65535)
ksp->ks_limblocks++;
tsleep((caddr_t)ksp, PSWP+2, memname[type], 0);
}
ksp->ks_size |= 1 << indx;
#endif
#ifdef DIAGNOSTIC
copysize = 1 << indx < MAX_COPY ? 1 << indx : MAX_COPY;
#endif
if (kbp->kb_next == NULL) {
kbp->kb_last = NULL;
if (size > MAXALLOCSAVE)
allocsize = roundup(size, CLBYTES);
else
allocsize = 1 << indx;
npg = clrnd(btoc(allocsize));
va = (caddr_t) kmem_malloc(kmem_map, (vm_size_t)ctob(npg),
!(flags & M_NOWAIT));
if (va == NULL) {
splx(s);
return ((void *) NULL);
}
#ifdef KMEMSTATS
kbp->kb_total += kbp->kb_elmpercl;
#endif
kup = btokup(va);
kup->ku_indx = indx;
if (allocsize > MAXALLOCSAVE) {
if (npg > 65535)
panic("malloc: allocation too large");
kup->ku_pagecnt = npg;
#ifdef KMEMSTATS
ksp->ks_memuse += allocsize;
#endif
goto out;
}
#ifdef KMEMSTATS
kup->ku_freecnt = kbp->kb_elmpercl;
kbp->kb_totalfree += kbp->kb_elmpercl;
#endif
/*
* Just in case we blocked while allocating memory,
* and someone else also allocated memory for this
* bucket, don't assume the list is still empty.
*/
savedlist = kbp->kb_next;
kbp->kb_next = cp = va + (npg * NBPG) - allocsize;
for (;;) {
freep = (struct freelist *)cp;
#ifdef DIAGNOSTIC
/*
* Copy in known text to detect modification
* after freeing.
*/
end = (long *)&cp[copysize];
for (lp = (long *)cp; lp < end; lp++)
*lp = WEIRD_ADDR;
freep->type = M_FREE;
#endif /* DIAGNOSTIC */
if (cp <= va)
break;
cp -= allocsize;
freep->next = cp;
}
freep->next = savedlist;
if (kbp->kb_last == NULL)
kbp->kb_last = (caddr_t)freep;
}
va = kbp->kb_next;
kbp->kb_next = ((struct freelist *)va)->next;
#ifdef DIAGNOSTIC
freep = (struct freelist *)va;
savedtype = (unsigned)freep->type < M_LAST ?
memname[freep->type] : "???";
if (kbp->kb_next &&
!kernacc(kbp->kb_next, sizeof(struct freelist), 0)) {
printf("%s of object %p size %ld %s %s (invalid addr %p)\n",
"Data modified on freelist: word 2.5", va, size,
"previous type", savedtype, kbp->kb_next);
kbp->kb_next = NULL;
}
#if BYTE_ORDER == BIG_ENDIAN
freep->type = WEIRD_ADDR >> 16;
#endif
#if BYTE_ORDER == LITTLE_ENDIAN
freep->type = (short)WEIRD_ADDR;
#endif
if (((long)(&freep->next)) & 0x2)
freep->next = (caddr_t)((WEIRD_ADDR >> 16)|(WEIRD_ADDR << 16));
else
freep->next = (caddr_t)WEIRD_ADDR;
end = (long *)&va[copysize];
for (lp = (long *)va; lp < end; lp++) {
if (*lp == WEIRD_ADDR)
continue;
printf("%s %d of object %p size %ld %s %s (0x%lx != 0x%x)\n",
"Data modified on freelist: word", lp - (long *)va,
va, size, "previous type", savedtype, *lp, WEIRD_ADDR);
break;
}
freep->spare0 = 0;
#endif /* DIAGNOSTIC */
#ifdef KMEMSTATS
kup = btokup(va);
if (kup->ku_indx != indx)
panic("malloc: wrong bucket");
if (kup->ku_freecnt == 0)
panic("malloc: lost data");
kup->ku_freecnt--;
kbp->kb_totalfree--;
ksp->ks_memuse += 1 << indx;
out:
kbp->kb_calls++;
ksp->ks_inuse++;
ksp->ks_calls++;
if (ksp->ks_memuse > ksp->ks_maxused)
ksp->ks_maxused = ksp->ks_memuse;
#else
out:
#endif
splx(s);
return ((void *) va);
}
/*
* Free a block of memory allocated by malloc.
*/
void
free(addr, type)
void *addr;
int type;
{
register struct kmembuckets *kbp;
register struct kmemusage *kup;
register struct freelist *freep;
long size;
int s;
#ifdef DIAGNOSTIC
caddr_t cp;
long *end, *lp, alloc, copysize;
#endif
#ifdef KMEMSTATS
register struct kmemstats *ksp = &kmemstats[type];
#endif
kup = btokup(addr);
size = 1 << kup->ku_indx;
kbp = &bucket[kup->ku_indx];
s = splhigh();
#ifdef DIAGNOSTIC
/*
* Check for returns of data that do not point to the
* beginning of the allocation.
*/
if (size > NBPG * CLSIZE)
alloc = addrmask[BUCKETINDX(NBPG * CLSIZE)];
else
alloc = addrmask[kup->ku_indx];
if (((u_long)addr & alloc) != 0)
panic("free: unaligned addr 0x%x, size %d, type %s, mask %d\n",
addr, size, memname[type], alloc);
#endif /* DIAGNOSTIC */
if (size > MAXALLOCSAVE) {
kmem_free(kmem_map, (vm_offset_t)addr, ctob(kup->ku_pagecnt));
#ifdef KMEMSTATS
size = kup->ku_pagecnt << PGSHIFT;
ksp->ks_memuse -= size;
kup->ku_indx = 0;
kup->ku_pagecnt = 0;
if (ksp->ks_memuse + size >= ksp->ks_limit &&
ksp->ks_memuse < ksp->ks_limit)
wakeup((caddr_t)ksp);
ksp->ks_inuse--;
kbp->kb_total -= 1;
#endif
splx(s);
return;
}
freep = (struct freelist *)addr;
#ifdef DIAGNOSTIC
/*
* Check for multiple frees. Use a quick check to see if
* it looks free before laboriously searching the freelist.
*/
if (freep->spare0 == WEIRD_ADDR) {
for (cp = kbp->kb_next; cp; cp = *(caddr_t *)cp) {
if (addr != cp)
continue;
printf("multiply freed item %p\n", addr);
panic("free: duplicated free");
}
}
/*
* Copy in known text to detect modification after freeing
* and to make it look free. Also, save the type being freed
* so we can list likely culprit if modification is detected
* when the object is reallocated.
*/
copysize = size < MAX_COPY ? size : MAX_COPY;
end = (long *)&((caddr_t)addr)[copysize];
for (lp = (long *)addr; lp < end; lp++)
*lp = WEIRD_ADDR;
freep->type = type;
#endif /* DIAGNOSTIC */
#ifdef KMEMSTATS
kup->ku_freecnt++;
if (kup->ku_freecnt >= kbp->kb_elmpercl)
if (kup->ku_freecnt > kbp->kb_elmpercl)
panic("free: multiple frees");
else if (kbp->kb_totalfree > kbp->kb_highwat)
kbp->kb_couldfree++;
kbp->kb_totalfree++;
ksp->ks_memuse -= size;
if (ksp->ks_memuse + size >= ksp->ks_limit &&
ksp->ks_memuse < ksp->ks_limit)
wakeup((caddr_t)ksp);
ksp->ks_inuse--;
#endif
if (kbp->kb_next == NULL)
kbp->kb_next = addr;
else
((struct freelist *)kbp->kb_last)->next = addr;
freep->next = NULL;
kbp->kb_last = addr;
splx(s);
}
/*
* Initialize the kernel memory allocator
*/
void
kmeminit()
{
register long indx;
int npg;
#if ((MAXALLOCSAVE & (MAXALLOCSAVE - 1)) != 0)
ERROR!_kmeminit:_MAXALLOCSAVE_not_power_of_2
#endif
#if (MAXALLOCSAVE > MINALLOCSIZE * 32768)
ERROR!_kmeminit:_MAXALLOCSAVE_too_big
#endif
#if (MAXALLOCSAVE < CLBYTES)
ERROR!_kmeminit:_MAXALLOCSAVE_too_small
#endif
npg = VM_KMEM_SIZE/ NBPG;
if( npg > cnt.v_page_count)
npg = cnt.v_page_count;
kmemusage = (struct kmemusage *) kmem_alloc(kernel_map,
(vm_size_t)(npg * sizeof(struct kmemusage)));
kmem_map = kmem_suballoc(kernel_map, (vm_offset_t *)&kmembase,
(vm_offset_t *)&kmemlimit, (vm_size_t)(npg * NBPG), FALSE);
#ifdef KMEMSTATS
for (indx = 0; indx < MINBUCKET + 16; indx++) {
if (1 << indx >= CLBYTES)
bucket[indx].kb_elmpercl = 1;
else
bucket[indx].kb_elmpercl = CLBYTES / (1 << indx);
bucket[indx].kb_highwat = 5 * bucket[indx].kb_elmpercl;
}
for (indx = 0; indx < M_LAST; indx++)
kmemstats[indx].ks_limit = npg * NBPG * 6 / 10;
#endif
}