freebsd-dev/sys/kern/kern_malloc.c

426 lines
12 KiB
C
Raw Normal View History

1994-05-24 10:09:53 +00:00
/*
* 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
1998-02-09 06:11:36 +00:00
* $Id: kern_malloc.c,v 1.42 1998/02/06 12:13:23 eivind Exp $
1994-05-24 10:09:53 +00:00
*/
#include <sys/param.h>
#include <sys/systm.h>
1994-05-24 10:09:53 +00:00
#include <sys/kernel.h>
#define MALLOC_INSTANTIATE
1994-05-24 10:09:53 +00:00
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/vmmeter.h>
#include <sys/lock.h>
1994-05-24 10:09:53 +00:00
#include <vm/vm.h>
#include <vm/vm_param.h>
1994-05-24 10:09:53 +00:00
#include <vm/vm_kern.h>
#include <vm/vm_extern.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
1994-05-24 10:09:53 +00:00
static void kmeminit __P((void *));
static void malloc_init __P((struct malloc_type *));
SYSINIT(kmem, SI_SUB_KMEM, SI_ORDER_FIRST, kmeminit, NULL)
static MALLOC_DEFINE(M_FREE, "free", "should be on free list");
1998-02-09 06:11:36 +00:00
static struct malloc_type *kmemstatistics = M_FREE;
static struct kmembuckets bucket[MINBUCKET + 16];
static struct kmemusage *kmemusage;
static char *kmembase;
static char *kmemlimit;
1994-05-24 10:09:53 +00:00
#ifdef DIAGNOSTIC
/*
* This structure provides a set of masks to catch unaligned frees.
*/
static long addrmask[] = { 0,
1994-05-24 10:09:53 +00:00
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 0xdeadc0de
#define MAX_COPY 64
1994-05-24 10:09:53 +00:00
/*
* 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;
struct malloc_type *type;
1994-05-24 10:09:53 +00:00
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;
struct malloc_type *type;
int flags;
1994-05-24 10:09:53 +00:00
{
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
register struct malloc_type *ksp = type;
if (!type->ks_next)
malloc_init(type);
1994-05-24 10:09:53 +00:00
indx = BUCKETINDX(size);
kbp = &bucket[indx];
s = splhigh();
1994-05-24 10:09:53 +00:00
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, type->ks_shortdesc, 0);
1994-05-24 10:09:53 +00:00
}
ksp->ks_size |= 1 << indx;
#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, PAGE_SIZE);
1994-05-24 10:09:53 +00:00
else
allocsize = 1 << indx;
npg = btoc(allocsize);
va = (caddr_t) kmem_malloc(kmem_map, (vm_size_t)ctob(npg), flags);
1994-05-24 10:09:53 +00:00
if (va == NULL) {
splx(s);
return ((void *) NULL);
}
kbp->kb_total += kbp->kb_elmpercl;
kup = btokup(va);
kup->ku_indx = indx;
if (allocsize > MAXALLOCSAVE) {
if (npg > 65535)
panic("malloc: allocation too large");
kup->ku_pagecnt = npg;
ksp->ks_memuse += allocsize;
goto out;
}
kup->ku_freecnt = kbp->kb_elmpercl;
kbp->kb_totalfree += kbp->kb_elmpercl;
/*
* 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 * PAGE_SIZE) - allocsize;
1994-05-24 10:09:53 +00:00
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 = (char *) type->ks_shortdesc;
1994-05-24 10:09:53 +00:00
#if BYTE_ORDER == BIG_ENDIAN
freep->type = (struct malloc_type *)WEIRD_ADDR >> 16;
1994-05-24 10:09:53 +00:00
#endif
#if BYTE_ORDER == LITTLE_ENDIAN
freep->type = (struct malloc_type *)WEIRD_ADDR;
1994-05-24 10:09:53 +00:00
#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",
1994-05-24 10:09:53 +00:00
"Data modified on freelist: word", lp - (long *)va,
va, size, "previous type", savedtype, *lp, WEIRD_ADDR);
break;
}
freep->spare0 = 0;
#endif /* DIAGNOSTIC */
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;
splx(s);
return ((void *) va);
}
/*
* Free a block of memory allocated by malloc.
*/
void
free(addr, type)
void *addr;
struct malloc_type *type;
1994-05-24 10:09:53 +00:00
{
register struct kmembuckets *kbp;
register struct kmemusage *kup;
register struct freelist *freep;
long size;
int s;
#ifdef DIAGNOSTIC
struct freelist *fp;
1994-05-24 10:09:53 +00:00
long *end, *lp, alloc, copysize;
#endif
register struct malloc_type *ksp = type;
if (!type->ks_next)
panic("freeing with unknown type (%s)", type->ks_shortdesc);
1994-05-24 10:09:53 +00:00
#ifdef DIAGNOSTIC
if ((char *)addr < kmembase || (char *)addr >= kmemlimit) {
panic("free: address 0x%x out of range", addr);
}
#endif
1994-05-24 10:09:53 +00:00
kup = btokup(addr);
size = 1 << kup->ku_indx;
kbp = &bucket[kup->ku_indx];
s = splhigh();
1994-05-24 10:09:53 +00:00
#ifdef DIAGNOSTIC
/*
* Check for returns of data that do not point to the
* beginning of the allocation.
*/
if (size > PAGE_SIZE)
alloc = addrmask[BUCKETINDX(PAGE_SIZE)];
1994-05-24 10:09:53 +00:00
else
alloc = addrmask[kup->ku_indx];
if (((u_long)addr & alloc) != 0)
panic("free: unaligned addr 0x%x, size %d, type %s, mask %d",
addr, size, type->ks_shortdesc, alloc);
1994-05-24 10:09:53 +00:00
#endif /* DIAGNOSTIC */
if (size > MAXALLOCSAVE) {
kmem_free(kmem_map, (vm_offset_t)addr, ctob(kup->ku_pagecnt));
size = kup->ku_pagecnt << PAGE_SHIFT;
1994-05-24 10:09:53 +00:00
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;
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) {
fp = (struct freelist *)kbp->kb_next;
while (fp) {
if (fp->spare0 != WEIRD_ADDR) {
printf("trashed free item %p\n", fp);
panic("free: free item modified");
} else if (addr == (caddr_t)fp) {
printf("multiple freed item %p\n", addr);
panic("free: multiple free");
}
fp = (struct freelist *)fp->next;
1994-05-24 10:09:53 +00:00
}
}
/*
* 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 */
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--;
#ifdef OLD_MALLOC_MEMORY_POLICY
1994-05-24 10:09:53 +00:00
if (kbp->kb_next == NULL)
kbp->kb_next = addr;
else
((struct freelist *)kbp->kb_last)->next = addr;
freep->next = NULL;
kbp->kb_last = addr;
#else
/*
* Return memory to the head of the queue for quick reuse. This
* can improve performance by improving the probability of the
* item being in the cache when it is reused.
*/
if (kbp->kb_next == NULL) {
kbp->kb_next = addr;
kbp->kb_last = addr;
freep->next = NULL;
} else {
freep->next = kbp->kb_next;
kbp->kb_next = addr;
}
#endif
1994-05-24 10:09:53 +00:00
splx(s);
}
/*
* Initialize the kernel memory allocator
*/
/* ARGSUSED*/
static void
kmeminit(dummy)
void *dummy;
1994-05-24 10:09:53 +00:00
{
register long indx;
int npg;
#if ((MAXALLOCSAVE & (MAXALLOCSAVE - 1)) != 0)
#error "kmeminit: MAXALLOCSAVE not power of 2"
1994-05-24 10:09:53 +00:00
#endif
#if (MAXALLOCSAVE > MINALLOCSIZE * 32768)
#error "kmeminit: MAXALLOCSAVE too big"
1994-05-24 10:09:53 +00:00
#endif
#if (MAXALLOCSAVE < PAGE_SIZE)
#error "kmeminit: MAXALLOCSAVE too small"
1994-05-24 10:09:53 +00:00
#endif
npg = (nmbufs * MSIZE + nmbclusters * MCLBYTES + VM_KMEM_SIZE)
/ PAGE_SIZE;
1995-05-30 08:16:23 +00:00
1994-05-24 10:09:53 +00:00
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 level code cleanups. 1) Start using TSM. Struct procs continue to point to upages structure, after being freed. Struct vmspace continues to point to pte object and kva space for kstack. u_map is now superfluous. 2) vm_map's don't need to be reference counted. They always exist either in the kernel or in a vmspace. The vmspaces are managed by reference counts. 3) Remove the "wired" vm_map nonsense. 4) No need to keep a cache of kernel stack kva's. 5) Get rid of strange looking ++var, and change to var++. 6) Change more data structures to use our "zone" allocator. Added struct proc, struct vmspace and struct vnode. This saves a significant amount of kva space and physical memory. Additionally, this enables TSM for the zone managed memory. 7) Keep ioopt disabled for now. 8) Remove the now bogus "single use" map concept. 9) Use generation counts or id's for data structures residing in TSM, where it allows us to avoid unneeded restart overhead during traversals, where blocking might occur. 10) Account better for memory deficits, so the pageout daemon will be able to make enough memory available (experimental.) 11) Fix some vnode locking problems. (From Tor, I think.) 12) Add a check in ufs_lookup, to avoid lots of unneeded calls to bcmp. (experimental.) 13) Significantly shrink, cleanup, and make slightly faster the vm_fault.c code. Use generation counts, get rid of unneded collpase operations, and clean up the cluster code. 14) Make vm_zone more suitable for TSM. This commit is partially as a result of discussions and contributions from other people, including DG, Tor Egge, PHK, and probably others that I have forgotten to attribute (so let me know, if I forgot.) This is not the infamous, final cleanup of the vnode stuff, but a necessary step. Vnode mgmt should be correct, but things might still change, and there is still some missing stuff (like ioopt, and physical backing of non-merged cache files, debugging of layering concepts.)
1998-01-22 17:30:44 +00:00
(vm_offset_t *)&kmemlimit, (vm_size_t)(npg * PAGE_SIZE));
kmem_map->system_map = 1;
1994-05-24 10:09:53 +00:00
for (indx = 0; indx < MINBUCKET + 16; indx++) {
if (1 << indx >= PAGE_SIZE)
1994-05-24 10:09:53 +00:00
bucket[indx].kb_elmpercl = 1;
else
bucket[indx].kb_elmpercl = PAGE_SIZE / (1 << indx);
1994-05-24 10:09:53 +00:00
bucket[indx].kb_highwat = 5 * bucket[indx].kb_elmpercl;
}
}
static void
malloc_init(type)
struct malloc_type *type;
{
int npg;
if (type->ks_magic != M_MAGIC)
panic("malloc type lacks magic");
if (cnt.v_page_count == 0)
panic("malloc_init not allowed before vm init");
/*
* Limit maximum memory for each type to 60% of malloc area size or
* 60% of physical memory, whichever is smaller.
*/
npg = (nmbufs * MSIZE + nmbclusters * MCLBYTES + VM_KMEM_SIZE)
/ PAGE_SIZE;
type->ks_limit = min(cnt.v_page_count * PAGE_SIZE,
(npg * PAGE_SIZE - nmbclusters * MCLBYTES
- nmbufs * MSIZE)) * 6 / 10;
type->ks_next = kmemstatistics;
kmemstatistics = type;
1994-05-24 10:09:53 +00:00
}