conversion of ino_t from u_short to u_long (making dir entries 32 bytes each)
[unix-history] / usr / src / sys / ufs / ffs / ufs_inode.c
/* ufs_inode.c 4.7 81/11/08 */
#include "../h/param.h"
#include "../h/systm.h"
#include "../h/mount.h"
#include "../h/dir.h"
#include "../h/user.h"
#include "../h/inode.h"
#include "../h/ino.h"
#include "../h/filsys.h"
#include "../h/conf.h"
#include "../h/buf.h"
#include "../h/inline.h"
#define INOHSZ 63
#define INOHASH(dev,ino) (((dev)+(ino))%INOHSZ)
short inohash[INOHSZ];
short ifreel;
/*
* Initialize hash links for inodes
* and build inode free list.
*/
ihinit()
{
register int i;
register struct inode *ip = inode;
ifreel = 0;
for (i = 0; i < ninode-1; i++, ip++)
ip->i_hlink = i+1;
ip->i_hlink = -1;
for (i = 0; i < INOHSZ; i++)
inohash[i] = -1;
}
/*
* Find an inode if it is incore.
* This is the equivalent, for inodes,
* of ``incore'' in bio.c or ``pfind'' in subr.c.
*/
struct inode *
ifind(dev, ino)
dev_t dev;
ino_t ino;
{
register struct inode *ip;
for (ip = &inode[inohash[INOHASH(dev,ino)]]; ip != &inode[-1];
ip = &inode[ip->i_hlink])
if (ino==ip->i_number && dev==ip->i_dev)
return (ip);
return ((struct inode *)0);
}
/*
* Look up an inode by device,inumber.
* If it is in core (in the inode structure),
* honor the locking protocol.
* If it is not in core, read it in from the
* specified device.
* If the inode is mounted on, perform
* the indicated indirection.
* In all cases, a pointer to a locked
* inode structure is returned.
*
* panic: no imt -- if the mounted file
* system is not in the mount table.
* "cannot happen"
*/
struct inode *
iget(dev, ino)
dev_t dev;
ino_t ino;
{
register struct inode *ip;
register struct mount *mp;
register struct buf *bp;
register struct dinode *dp;
register int slot;
loop:
slot = INOHASH(dev, ino);
ip = &inode[inohash[slot]];
while (ip != &inode[-1]) {
if (ino == ip->i_number && dev == ip->i_dev) {
if ((ip->i_flag&ILOCK) != 0) {
ip->i_flag |= IWANT;
sleep((caddr_t)ip, PINOD);
goto loop;
}
if ((ip->i_flag&IMOUNT) != 0) {
for(mp = &mount[0]; mp < &mount[NMOUNT]; mp++)
if (mp->m_inodp == ip) {
dev = mp->m_dev;
ino = ROOTINO;
goto loop;
}
panic("no imt");
}
ip->i_count++;
ip->i_flag |= ILOCK;
return(ip);
}
ip = &inode[ip->i_hlink];
}
if (ifreel < 0) {
tablefull("inode");
u.u_error = ENFILE;
return(NULL);
}
ip = &inode[ifreel];
ifreel = ip->i_hlink;
ip->i_hlink = inohash[slot];
inohash[slot] = ip - inode;
ip->i_dev = dev;
ip->i_number = ino;
ip->i_flag = ILOCK;
ip->i_count++;
ip->i_un.i_lastr = 0;
bp = bread(dev, itod(ino));
/*
* Check I/O errors
*/
if ((bp->b_flags&B_ERROR) != 0) {
brelse(bp);
iput(ip);
return(NULL);
}
dp = bp->b_un.b_dino;
dp += itoo(ino);
iexpand(ip, dp);
brelse(bp);
return(ip);
}
iexpand(ip, dp)
register struct inode *ip;
register struct dinode *dp;
{
register char *p1, *p2;
register int i;
ip->i_mode = dp->di_mode;
ip->i_nlink = dp->di_nlink;
ip->i_uid = dp->di_uid;
ip->i_gid = dp->di_gid;
ip->i_size = dp->di_size;
p1 = (char *)ip->i_un.i_addr;
p2 = (char *)dp->di_addr;
for(i=0; i<NADDR; i++) {
*p1++ = *p2++;
*p1++ = *p2++;
*p1++ = *p2++;
*p1++ = 0;
}
}
/*
* Decrement reference count of
* an inode structure.
* On the last reference,
* write the inode out and if necessary,
* truncate and deallocate the file.
*/
iput(ip)
register struct inode *ip;
{
register int i, x;
register struct inode *jp;
if (ip->i_count == 1) {
ip->i_flag |= ILOCK;
if (ip->i_nlink <= 0) {
itrunc(ip);
ip->i_mode = 0;
ip->i_flag |= IUPD|ICHG;
ifree(ip->i_dev, ip->i_number);
}
IUPDAT(ip, &time, &time, 0);
irele(ip);
i = INOHASH(ip->i_dev, ip->i_number);
x = ip - inode;
if (inohash[i] == x) {
inohash[i] = ip->i_hlink;
} else {
for (jp = &inode[inohash[i]]; jp != &inode[-1];
jp = &inode[jp->i_hlink])
if (jp->i_hlink == x) {
jp->i_hlink = ip->i_hlink;
goto done;
}
panic("iput");
}
done:
ip->i_hlink = ifreel;
ifreel = x;
ip->i_flag = 0;
ip->i_number = 0;
} else
irele(ip);
ip->i_count--;
}
/*
* Check accessed and update flags on
* an inode structure.
* If any is on, update the inode
* with the current time.
* If waitfor is given, then must insure
* i/o order so wait for write to complete.
*/
iupdat(ip, ta, tm, waitfor)
register struct inode *ip;
time_t *ta, *tm;
int waitfor;
{
register struct buf *bp;
struct dinode *dp;
register char *p1, *p2;
register int i;
if ((ip->i_flag&(IUPD|IACC|ICHG)) != 0) {
if (getfs(ip->i_dev)->s_ronly)
return;
bp = bread(ip->i_dev, itod(ip->i_number));
if (bp->b_flags & B_ERROR) {
brelse(bp);
return;
}
dp = bp->b_un.b_dino;
dp += itoo(ip->i_number);
dp->di_mode = ip->i_mode;
dp->di_nlink = ip->i_nlink;
dp->di_uid = ip->i_uid;
dp->di_gid = ip->i_gid;
dp->di_size = ip->i_size;
p1 = (char *)dp->di_addr;
p2 = (char *)ip->i_un.i_addr;
for(i=0; i<NADDR; i++) {
*p1++ = *p2++;
*p1++ = *p2++;
*p1++ = *p2++;
if (*p2++ != 0)
printf("iaddress > 2^24\n");
}
if (ip->i_flag&IACC)
dp->di_atime = *ta;
if (ip->i_flag&IUPD)
dp->di_mtime = *tm;
if (ip->i_flag&ICHG)
dp->di_ctime = time;
ip->i_flag &= ~(IUPD|IACC|ICHG);
if (waitfor)
bwrite(bp);
else
bdwrite(bp);
}
}
/*
* Free all the disk blocks associated
* with the specified inode structure.
* The blocks of the file are removed
* in reverse order. This FILO
* algorithm will tend to maintain
* a contiguous free list much longer
* than FIFO.
*/
itrunc(ip)
register struct inode *ip;
{
register i;
dev_t dev;
daddr_t bn;
struct inode itmp;
i = ip->i_mode & IFMT;
if (i!=IFREG && i!=IFDIR)
return;
/*
* Clean inode on disk before freeing blocks
* to insure no duplicates if system crashes.
*/
itmp = *ip;
itmp.i_size = 0;
for (i = 0; i < NADDR; i++)
itmp.i_un.i_addr[i] = 0;
itmp.i_flag |= ICHG|IUPD;
iupdat(&itmp, &time, &time, 1);
ip->i_flag &= ~(IUPD|IACC|ICHG);
/*
* Now return blocks to free list... if machine
* crashes, they will be harmless MISSING blocks.
*/
dev = ip->i_dev;
for(i=NADDR-1; i>=0; i--) {
bn = ip->i_un.i_addr[i];
if (bn == (daddr_t)0)
continue;
ip->i_un.i_addr[i] = (daddr_t)0;
switch(i) {
default:
free(dev, bn);
break;
case NADDR-3:
tloop(dev, bn, 0, 0);
break;
case NADDR-2:
tloop(dev, bn, 1, 0);
break;
case NADDR-1:
tloop(dev, bn, 1, 1);
}
}
ip->i_size = 0;
/*
* Inode was written and flags updated above.
* No need to modify flags here.
*/
}
tloop(dev, bn, f1, f2)
dev_t dev;
daddr_t bn;
{
register i;
register struct buf *bp;
register daddr_t *bap;
daddr_t nb;
bp = NULL;
for(i=NINDIR-1; i>=0; i--) {
if (bp == NULL) {
bp = bread(dev, bn);
if (bp->b_flags & B_ERROR) {
brelse(bp);
return;
}
bap = bp->b_un.b_daddr;
}
nb = bap[i];
if (nb == (daddr_t)0)
continue;
if (f1) {
brelse(bp);
bp = NULL;
tloop(dev, nb, f2, 0);
} else
free(dev, nb);
}
if (bp != NULL)
brelse(bp);
free(dev, bn);
}
/*
* Make a new file.
*/
struct inode *
maknode(mode)
{
register struct inode *ip;
ip = ialloc(u.u_pdir->i_dev);
if (ip == NULL) {
iput(u.u_pdir);
return(NULL);
}
ip->i_flag |= IACC|IUPD|ICHG;
if ((mode&IFMT) == 0)
mode |= IFREG;
ip->i_mode = mode & ~u.u_cmask;
ip->i_nlink = 1;
ip->i_uid = u.u_uid;
ip->i_gid = u.u_gid;
/*
* Make sure inode goes to disk before directory entry.
*/
iupdat(ip, &time, &time, 1);
wdir(ip);
return(ip);
}
/*
* Write a directory entry with
* parameters left as side effects
* to a call to namei.
*/
wdir(ip)
struct inode *ip;
{
u.u_dent.d_ino = ip->i_number;
bcopy((caddr_t)u.u_dbuf, (caddr_t)u.u_dent.d_name, DIRSIZ);
u.u_count = sizeof(struct direct);
u.u_segflg = 1;
u.u_base = (caddr_t)&u.u_dent;
writei(u.u_pdir);
iput(u.u_pdir);
}
#ifdef ilock
#undef ilock
#endif
#ifdef irele
#undef irele
#endif
/*
* Lock an inode. If its already locked, set the WANT bit and sleep.
*/
ilock(ip)
register struct inode *ip;
{
while (ip->i_flag&ILOCK) {
ip->i_flag |= IWANT;
sleep((caddr_t)ip, PINOD);
}
ip->i_flag |= ILOCK;
}
/*
* Unlock an inode. If WANT bit is on, wakeup.
*/
irele(ip)
register struct inode *ip;
{
ip->i_flag &= ~ILOCK;
if (ip->i_flag&IWANT) {
ip->i_flag &= ~IWANT;
wakeup((caddr_t)ip);
}
}