ref: 25139465363b2230b69789013e3e0b5f82d76564
dir: /sys/src/9/port/swap.c/
#include "u.h"
#include "../port/lib.h"
#include "mem.h"
#include "dat.h"
#include "fns.h"
#include "../port/error.h"
static int canflush(Proc*, Segment*);
static void executeio(void);
static void pageout(Proc*, Segment*);
static void pagepte(int, Page**);
static void pager(void*);
Image swapimage;
static int swopen;
static Page **iolist;
static int ioptr;
static ushort ageclock;
void
swapinit(void)
{
swapalloc.swmap = xalloc(conf.nswap);
swapalloc.top = &swapalloc.swmap[conf.nswap];
swapalloc.alloc = swapalloc.swmap;
swapalloc.last = swapalloc.swmap;
swapalloc.free = conf.nswap;
swapalloc.xref = 0;
iolist = xalloc(conf.nswppo*sizeof(Page*));
if(swapalloc.swmap == 0 || iolist == 0)
panic("swapinit: not enough memory");
swapimage.notext = 1;
}
static uintptr
newswap(void)
{
uchar *look;
lock(&swapalloc);
if(swapalloc.free == 0) {
unlock(&swapalloc);
return ~0;
}
look = memchr(swapalloc.last, 0, swapalloc.top-swapalloc.last);
if(look == nil)
look = memchr(swapalloc.swmap, 0, swapalloc.last-swapalloc.swmap);
*look = 2; /* ref for pte + io transaction */
swapalloc.last = look;
swapalloc.free--;
unlock(&swapalloc);
return (look-swapalloc.swmap) * BY2PG;
}
void
putswap(Page *p)
{
uchar *idx;
lock(&swapalloc);
idx = &swapalloc.swmap[((uintptr)p)/BY2PG];
if(*idx == 0)
panic("putswap %#p ref == 0", p);
if(*idx == 255) {
if(swapalloc.xref == 0)
panic("putswap %#p xref == 0", p);
if(--swapalloc.xref == 0) {
for(idx = swapalloc.swmap; idx < swapalloc.top; idx++) {
if(*idx == 255) {
*idx = 0;
swapalloc.free++;
}
}
}
} else {
if(--(*idx) == 0)
swapalloc.free++;
}
unlock(&swapalloc);
}
void
dupswap(Page *p)
{
uchar *idx;
lock(&swapalloc);
idx = &swapalloc.swmap[((uintptr)p)/BY2PG];
if(*idx == 255)
swapalloc.xref++;
else {
if(++(*idx) == 255)
swapalloc.xref += 255;
}
unlock(&swapalloc);
}
int
swapcount(uintptr daddr)
{
return swapalloc.swmap[daddr/BY2PG];
}
void
kickpager(void)
{
static Ref started;
if(started.ref || incref(&started) != 1)
wakeup(&swapalloc.r);
else
kproc("pager", pager, 0);
}
static int
reclaim(void)
{
ulong np;
for(;;){
if((np = pagereclaim(&fscache, 1000)) > 0) {
if(0) print("reclaim: %lud fscache\n", np);
} else if((np = pagereclaim(&swapimage, 1000)) > 0) {
if(0) print("reclaim: %lud swap\n", np);
} else if((np = imagereclaim(1000)) > 0) {
if(0) print("reclaim: %lud image\n", np);
}
if(!needpages(nil))
return 1; /* have pages, done */
if(np == 0)
return 0; /* didnt reclaim, need to swap */
sched();
}
}
static void
pager(void*)
{
int i;
Segment *s;
Proc *p, *ep;
p = proctab(0);
ep = &p[conf.nproc];
while(waserror())
;
for(;;){
up->psstate = "Reclaim";
if(reclaim()){
up->psstate = "Idle";
wakeup(&palloc.pwait[0]);
wakeup(&palloc.pwait[1]);
sleep(&swapalloc.r, needpages, nil);
continue;
}
if(swapimage.c == nil || swapalloc.free == 0){
Killbig:
killbig("out of memory");
freebroken(); /* can use the memory */
sched();
continue;
}
i = ageclock;
do {
if(++p >= ep){
if(++ageclock == i)
goto Killbig;
p = proctab(0);
}
} while(p->state == Dead || p->noswap || !canqlock(&p->seglock));
up->psstate = "Pageout";
for(i = 0; i < NSEG; i++) {
if((s = p->seg[i]) != nil) {
switch(s->type&SG_TYPE) {
default:
break;
case SG_TEXT:
pageout(p, s);
break;
case SG_DATA:
case SG_BSS:
case SG_STACK:
case SG_SHARED:
pageout(p, s);
break;
}
}
}
qunlock(&p->seglock);
if(ioptr > 0) {
up->psstate = "I/O";
executeio();
}
}
}
static void
pageout(Proc *p, Segment *s)
{
int type, i, size;
short age;
Pte *l;
Page **pg, *entry;
if(!canqlock(s)) /* We cannot afford to wait, we will surely deadlock */
return;
if(!canflush(p, s)) { /* Able to invalidate all tlbs with references */
qunlock(s);
putseg(s);
return;
}
if(waserror()) {
qunlock(s);
putseg(s);
return;
}
/* Pass through the pte tables looking for memory pages to swap out */
type = s->type&SG_TYPE;
size = s->mapsize;
for(i = 0; i < size; i++) {
l = s->map[i];
if(l == nil)
continue;
for(pg = l->first; pg <= l->last; pg++) {
entry = *pg;
if(pagedout(entry))
continue;
if(entry->modref & PG_REF) {
entry->modref &= ~PG_REF;
entry->refage = ageclock;
continue;
}
age = (short)(ageclock - entry->refage);
if(age < 16)
continue;
pagepte(type, pg);
}
}
poperror();
qunlock(s);
putseg(s);
}
static int
canflush(Proc *p, Segment *s)
{
int i;
Proc *ep;
if(incref(s) == 2) /* Easy if we are the only user */
return canpage(p);
/* Now we must do hardwork to ensure all processes which have tlb
* entries for this segment will be flushed if we succeed in paging it out
*/
p = proctab(0);
ep = &p[conf.nproc];
while(p < ep) {
if(p->state != Dead) {
for(i = 0; i < NSEG; i++)
if(p->seg[i] == s)
if(!canpage(p))
return 0;
}
p++;
}
return 1;
}
static void
pagepte(int type, Page **pg)
{
uintptr daddr;
Page *outp;
outp = *pg;
switch(type) {
case SG_TEXT: /* Revert to demand load */
putpage(outp);
*pg = nil;
break;
case SG_DATA:
case SG_BSS:
case SG_STACK:
case SG_SHARED:
if(ioptr >= conf.nswppo)
break;
/*
* get a new swap address with swapcount 2, one for the pte
* and one extra ref for us while we write the page to disk
*/
daddr = newswap();
if(daddr == ~0)
break;
/* clear any pages referring to it from the cache */
cachedel(&swapimage, daddr);
/* forget anything that it used to cache */
uncachepage(outp);
/*
* enter it into the cache so that a fault happening
* during the write will grab the page from the cache
* rather than one partially written to the disk
*/
outp->daddr = daddr;
cachepage(outp, &swapimage);
*pg = (Page*)(daddr|PG_ONSWAP);
/* Add page to IO transaction list */
iolist[ioptr++] = outp;
break;
}
}
void
pagersummary(void)
{
print("%lud/%lud memory %lud/%lud swap %d iolist\n",
palloc.user-palloc.freecount,
palloc.user, conf.nswap-swapalloc.free, conf.nswap,
ioptr);
}
static void
executeio(void)
{
Page *outp;
int i, n;
Chan *c;
char *kaddr;
KMap *k;
c = swapimage.c;
for(i = 0; i < ioptr; i++) {
if(ioptr > conf.nswppo)
panic("executeio: ioptr %d > %d", ioptr, conf.nswppo);
outp = iolist[i];
assert(outp->ref > 0);
assert(outp->image == &swapimage);
assert(outp->daddr != ~0);
/* only write when swap address still in use */
if(swapcount(outp->daddr) > 1){
k = kmap(outp);
kaddr = (char*)VA(k);
if(waserror())
panic("executeio: page outp I/O error");
n = devtab[c->type]->write(c, kaddr, BY2PG, outp->daddr);
if(n != BY2PG)
nexterror();
kunmap(k);
poperror();
}
/* drop our extra swap reference */
putswap((Page*)outp->daddr);
/* Free up the page after I/O */
putpage(outp);
}
ioptr = 0;
}
int
needpages(void*)
{
return palloc.freecount < swapalloc.headroom;
}
void
setswapchan(Chan *c)
{
uchar dirbuf[sizeof(Dir)+100];
Dir d;
int n;
if(waserror()){
cclose(c);
nexterror();
}
if(swapimage.c != nil) {
if(swapalloc.free != conf.nswap)
error(Einuse);
cclose(swapimage.c);
swapimage.c = nil;
}
/*
* if this isn't a file, set the swap space
* to be at most the size of the partition
*/
if(devtab[c->type]->dc != L'M'){
n = devtab[c->type]->stat(c, dirbuf, sizeof dirbuf);
if(n <= 0 || convM2D(dirbuf, n, &d, nil) == 0)
error("stat failed in setswapchan");
if(d.length < conf.nswppo*BY2PG)
error("swap device too small");
if(d.length < conf.nswap*BY2PG){
conf.nswap = d.length/BY2PG;
swapalloc.top = &swapalloc.swmap[conf.nswap];
swapalloc.free = conf.nswap;
}
}
c->flag &= ~CCACHE;
cclunk(c);
swapimage.c = c;
poperror();
}