X-Git-Url: https://git.subgeniuskitty.com/unix-history/.git/blobdiff_plain/b9ccf5146beb58e8bd314ef1b64611ee3eaea45e..fd88f5c5678c80ff5e338adc372d28a52ad20530:/usr/src/sys/ufs/lfs/lfs_alloc.c diff --git a/usr/src/sys/ufs/lfs/lfs_alloc.c b/usr/src/sys/ufs/lfs/lfs_alloc.c index bd93b59d35..837d2ea496 100644 --- a/usr/src/sys/ufs/lfs/lfs_alloc.c +++ b/usr/src/sys/ufs/lfs/lfs_alloc.c @@ -1,1106 +1,251 @@ /* - * Copyright (c) 1982, 1986, 1989 Regents of the University of California. - * All rights reserved. + * Copyright (c) 1991, 1993, 1995 + * The Regents of the University of California. All rights reserved. * - * %sccs.include.redist.c% + * 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. * - * @(#)lfs_alloc.c 7.25 (Berkeley) %G% + * 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. + * + * @(#)lfs_alloc.c 8.7 (Berkeley) 5/14/95 */ -#include "param.h" -#include "systm.h" -#include "buf.h" -#include "proc.h" -#include "vnode.h" -#include "kernel.h" -#include "syslog.h" - -#include "quota.h" -#include "inode.h" -#include "fs.h" - -extern u_long hashalloc(); -extern ino_t ialloccg(); -extern daddr_t alloccg(); -extern daddr_t alloccgblk(); -extern daddr_t fragextend(); -extern daddr_t blkpref(); -extern daddr_t mapsearch(); -extern int inside[], around[]; -extern unsigned char *fragtbl[]; - -/* - * Allocate a block in the file system. - * - * The size of the requested block is given, which must be some - * multiple of fs_fsize and <= fs_bsize. - * A preference may be optionally specified. If a preference is given - * the following hierarchy is used to allocate a block: - * 1) allocate the requested block. - * 2) allocate a rotationally optimal block in the same cylinder. - * 3) allocate a block in the same cylinder group. - * 4) quadradically rehash into other cylinder groups, until an - * available block is located. - * If no block preference is given the following heirarchy is used - * to allocate a block: - * 1) allocate a block in the cylinder group that contains the - * inode for the file. - * 2) quadradically rehash into other cylinder groups, until an - * available block is located. - */ -alloc(ip, lbn, bpref, size, bnp) - register struct inode *ip; - daddr_t lbn, bpref; - int size; - daddr_t *bnp; +#include +#include +#include +#include +#include +#include +#include + +#include + +#include +#include +#include +#include + +#include +#include + +extern u_long nextgennumber; + +/* Allocate a new inode. */ +/* ARGSUSED */ +int +lfs_valloc(ap) + struct vop_valloc_args /* { + struct vnode *a_pvp; + int a_mode; + struct ucred *a_cred; + struct vnode **a_vpp; + } */ *ap; { - daddr_t bno; - register struct fs *fs; - register struct buf *bp; - int cg, error; - struct ucred *cred = curproc->p_ucred; /* XXX */ - - *bnp = 0; - fs = ip->i_fs; - if ((unsigned)size > fs->fs_bsize || fragoff(fs, size) != 0) { - printf("dev = 0x%x, bsize = %d, size = %d, fs = %s\n", - ip->i_dev, fs->fs_bsize, size, fs->fs_fsmnt); - panic("alloc: bad size"); - } - if (size == fs->fs_bsize && fs->fs_cstotal.cs_nbfree == 0) - goto nospace; - if (cred->cr_uid != 0 && freespace(fs, fs->fs_minfree) <= 0) - goto nospace; -#ifdef QUOTA - if (error = chkdq(ip, (long)btodb(size), cred, 0)) - return (error); -#endif - if (bpref >= fs->fs_size) - bpref = 0; - if (bpref == 0) - cg = itog(fs, ip->i_number); - else - cg = dtog(fs, bpref); - bno = (daddr_t)hashalloc(ip, cg, (long)bpref, size, - (u_long (*)())alloccg); - if (bno > 0) { - ip->i_blocks += btodb(size); - ip->i_flag |= IUPD|ICHG; - *bnp = bno; - return (0); - } -#ifdef QUOTA - /* - * Restore user's disk quota because allocation failed. - */ - (void) chkdq(ip, (long)-btodb(size), cred, FORCE); + struct lfs *fs; + struct buf *bp; + struct ifile *ifp; + struct inode *ip; + struct vnode *vp; + ufs_daddr_t blkno; + ino_t new_ino; + u_long i, max; + int error; + + /* Get the head of the freelist. */ + fs = VTOI(ap->a_pvp)->i_lfs; + new_ino = fs->lfs_free; +#ifdef ALLOCPRINT + printf("lfs_ialloc: allocate inode %d\n", new_ino); #endif -nospace: - fserr(fs, cred->cr_uid, "file system full"); - uprintf("\n%s: write failed, file system is full\n", fs->fs_fsmnt); - return (ENOSPC); -} -/* - * Reallocate a fragment to a bigger size - * - * The number and size of the old block is given, and a preference - * and new size is also specified. The allocator attempts to extend - * the original block. Failing that, the regular block allocator is - * invoked to get an appropriate block. - */ -realloccg(ip, lbprev, bpref, osize, nsize, bpp) - register struct inode *ip; - off_t lbprev; - daddr_t bpref; - int osize, nsize; - struct buf **bpp; -{ - register struct fs *fs; - struct buf *bp, *obp; - int cg, request, error; - daddr_t bprev, bno; - struct ucred *cred = curproc->p_ucred; /* XXX */ - - *bpp = 0; - fs = ip->i_fs; - if ((unsigned)osize > fs->fs_bsize || fragoff(fs, osize) != 0 || - (unsigned)nsize > fs->fs_bsize || fragoff(fs, nsize) != 0) { - printf("dev = 0x%x, bsize = %d, osize = %d, nsize = %d, fs = %s\n", - ip->i_dev, fs->fs_bsize, osize, nsize, fs->fs_fsmnt); - panic("realloccg: bad size"); - } - if (cred->cr_uid != 0 && freespace(fs, fs->fs_minfree) <= 0) - goto nospace; - if ((bprev = ip->i_db[lbprev]) == 0) { - printf("dev = 0x%x, bsize = %d, bprev = %d, fs = %s\n", - ip->i_dev, fs->fs_bsize, bprev, fs->fs_fsmnt); - panic("realloccg: bad bprev"); - } /* - * Allocate the extra space in the buffer. + * Remove the inode from the free list and write the new start + * of the free list into the superblock. */ - if (error = bread(ITOV(ip), lbprev, osize, NOCRED, &bp)) { - brelse(bp); - return (error); - } -#ifdef QUOTA - if (error = chkdq(ip, (long)btodb(nsize - osize), cred, 0)) { - brelse(bp); - return (error); - } -#endif - allocbuf(bp, nsize); - bp->b_flags |= B_DONE; - bzero(bp->b_un.b_addr + osize, (unsigned)nsize - osize); - /* - * Check for extension in the existing location. - */ - cg = dtog(fs, bprev); - if (bno = fragextend(ip, cg, (long)bprev, osize, nsize)) { - if (bp->b_blkno != fsbtodb(fs, bno)) - panic("bad blockno"); - ip->i_blocks += btodb(nsize - osize); - ip->i_flag |= IUPD|ICHG; - *bpp = bp; - return (0); - } - /* - * Allocate a new disk location. - */ - if (bpref >= fs->fs_size) - bpref = 0; - switch ((int)fs->fs_optim) { - case FS_OPTSPACE: - /* - * Allocate an exact sized fragment. Although this makes - * best use of space, we will waste time relocating it if - * the file continues to grow. If the fragmentation is - * less than half of the minimum free reserve, we choose - * to begin optimizing for time. - */ - request = nsize; - if (fs->fs_minfree < 5 || - fs->fs_cstotal.cs_nffree > - fs->fs_dsize * fs->fs_minfree / (2 * 100)) - break; - log(LOG_NOTICE, "%s: optimization changed from SPACE to TIME\n", - fs->fs_fsmnt); - fs->fs_optim = FS_OPTTIME; - break; - case FS_OPTTIME: - /* - * At this point we have discovered a file that is trying - * to grow a small fragment to a larger fragment. To save - * time, we allocate a full sized block, then free the - * unused portion. If the file continues to grow, the - * `fragextend' call above will be able to grow it in place - * without further copying. If aberrant programs cause - * disk fragmentation to grow within 2% of the free reserve, - * we choose to begin optimizing for space. - */ - request = fs->fs_bsize; - if (fs->fs_cstotal.cs_nffree < - fs->fs_dsize * (fs->fs_minfree - 2) / 100) - break; - log(LOG_NOTICE, "%s: optimization changed from TIME to SPACE\n", - fs->fs_fsmnt); - fs->fs_optim = FS_OPTSPACE; - break; - default: - printf("dev = 0x%x, optim = %d, fs = %s\n", - ip->i_dev, fs->fs_optim, fs->fs_fsmnt); - panic("realloccg: bad optim"); - /* NOTREACHED */ - } - bno = (daddr_t)hashalloc(ip, cg, (long)bpref, request, - (u_long (*)())alloccg); - if (bno > 0) { -#ifdef SECSIZE - obp = bread(ip->i_dev, fsbtodb(fs, bprev), osize, - fs->fs_dbsize); -#else SECSIZE - count = howmany(osize, CLBYTES); - for (i = 0; i < count; i++) -#ifdef SECSIZE - munhash(ip->i_dev, bn + i * CLBYTES / fs->fs_dbsize); -#else SECSIZE - munhash(ip->i_dev, bn + i * CLBYTES / DEV_BSIZE); -#endif SECSIZE - blkfree(ip, bprev, (off_t)osize); - if (nsize < request) - blkfree(ip, bno + numfrags(fs, nsize), - (off_t)(request - nsize)); - ip->i_blocks += btodb(nsize - osize); - ip->i_flag |= IUPD|ICHG; - *bpp = bp; - return (0); - } -#ifdef QUOTA - /* - * Restore user's disk quota because allocation failed. - */ - (void) chkdq(ip, (long)-btodb(nsize - osize), cred, FORCE); -#endif + LFS_IENTRY(ifp, fs, new_ino, bp); + if (ifp->if_daddr != LFS_UNUSED_DADDR) + panic("lfs_ialloc: inuse inode on the free list"); + fs->lfs_free = ifp->if_nextfree; brelse(bp); -nospace: - /* - * no space available - */ - fserr(fs, cred->cr_uid, "file system full"); - uprintf("\n%s: write failed, file system is full\n", fs->fs_fsmnt); - return (ENOSPC); -} - -/* - * Allocate an inode in the file system. - * - * A preference may be optionally specified. If a preference is given - * the following hierarchy is used to allocate an inode: - * 1) allocate the requested inode. - * 2) allocate an inode in the same cylinder group. - * 3) quadradically rehash into other cylinder groups, until an - * available inode is located. - * If no inode preference is given the following heirarchy is used - * to allocate an inode: - * 1) allocate an inode in cylinder group 0. - * 2) quadradically rehash into other cylinder groups, until an - * available inode is located. - */ -ialloc(pip, ipref, mode, cred, ipp) - register struct inode *pip; - ino_t ipref; - int mode; - struct ucred *cred; - struct inode **ipp; -{ - ino_t ino; - register struct fs *fs; - register struct inode *ip; - int cg, error; - - *ipp = 0; - fs = pip->i_fs; - if (fs->fs_cstotal.cs_nifree == 0) - goto noinodes; - if (ipref >= fs->fs_ncg * fs->fs_ipg) - ipref = 0; - cg = itog(fs, ipref); - ino = (ino_t)hashalloc(pip, cg, (long)ipref, mode, ialloccg); - if (ino == 0) - goto noinodes; - error = iget(pip, ino, ipp); - if (error) { - ifree(pip, ino, mode); - return (error); - } - ip = *ipp; - if (ip->i_mode) { - printf("mode = 0%o, inum = %d, fs = %s\n", - ip->i_mode, ip->i_number, fs->fs_fsmnt); - panic("ialloc: dup alloc"); - } - if (ip->i_blocks) { /* XXX */ - printf("free inode %s/%d had %d blocks\n", - fs->fs_fsmnt, ino, ip->i_blocks); - ip->i_blocks = 0; - } - ip->i_flags = 0; - /* - * Set up a new generation number for this inode. - */ - if (++nextgennumber < (u_long)time.tv_sec) - nextgennumber = time.tv_sec; - ip->i_gen = nextgennumber; - return (0); -noinodes: - fserr(fs, cred->cr_uid, "out of inodes"); - uprintf("\n%s: create/symlink failed, no inodes free\n", fs->fs_fsmnt); - return (ENOSPC); -} - -/* - * Find a cylinder to place a directory. - * - * The policy implemented by this algorithm is to select from - * among those cylinder groups with above the average number of - * free inodes, the one with the smallest number of directories. - */ -ino_t -dirpref(fs) - register struct fs *fs; -{ - int cg, minndir, mincg, avgifree; - avgifree = fs->fs_cstotal.cs_nifree / fs->fs_ncg; - minndir = fs->fs_ipg; - mincg = 0; - for (cg = 0; cg < fs->fs_ncg; cg++) - if (fs->fs_cs(fs, cg).cs_ndir < minndir && - fs->fs_cs(fs, cg).cs_nifree >= avgifree) { - mincg = cg; - minndir = fs->fs_cs(fs, cg).cs_ndir; + /* Extend IFILE so that the next lfs_valloc will succeed. */ + if (fs->lfs_free == LFS_UNUSED_INUM) { + vp = fs->lfs_ivnode; + ip = VTOI(vp); + blkno = lblkno(fs, ip->i_size); + lfs_balloc(vp, 0, fs->lfs_bsize, blkno, &bp); + ip->i_size += fs->lfs_bsize; + vnode_pager_setsize(vp, (u_long)ip->i_size); + vnode_pager_uncache(vp); + + i = (blkno - fs->lfs_segtabsz - fs->lfs_cleansz) * + fs->lfs_ifpb; + fs->lfs_free = i; + max = i + fs->lfs_ifpb; + for (ifp = (struct ifile *)bp->b_data; i < max; ++ifp) { + ifp->if_version = 1; + ifp->if_daddr = LFS_UNUSED_DADDR; + ifp->if_nextfree = ++i; } - return ((ino_t)(fs->fs_ipg * mincg)); -} + ifp--; + ifp->if_nextfree = LFS_UNUSED_INUM; + if (error = VOP_BWRITE(bp)) + return (error); + } -/* - * Select the desired position for the next block in a file. The file is - * logically divided into sections. The first section is composed of the - * direct blocks. Each additional section contains fs_maxbpg blocks. - * - * If no blocks have been allocated in the first section, the policy is to - * request a block in the same cylinder group as the inode that describes - * the file. If no blocks have been allocated in any other section, the - * policy is to place the section in a cylinder group with a greater than - * average number of free blocks. An appropriate cylinder group is found - * by using a rotor that sweeps the cylinder groups. When a new group of - * blocks is needed, the sweep begins in the cylinder group following the - * cylinder group from which the previous allocation was made. The sweep - * continues until a cylinder group with greater than the average number - * of free blocks is found. If the allocation is for the first block in an - * indirect block, the information on the previous allocation is unavailable; - * here a best guess is made based upon the logical block number being - * allocated. - * - * If a section is already partially allocated, the policy is to - * contiguously allocate fs_maxcontig blocks. The end of one of these - * contiguous blocks and the beginning of the next is physically separated - * so that the disk head will be in transit between them for at least - * fs_rotdelay milliseconds. This is to allow time for the processor to - * schedule another I/O transfer. - */ -daddr_t -blkpref(ip, lbn, indx, bap) - struct inode *ip; - daddr_t lbn; - int indx; - daddr_t *bap; -{ - register struct fs *fs; - register int cg; - int avgbfree, startcg; - daddr_t nextblk; + /* Create a vnode to associate with the inode. */ + if (error = lfs_vcreate(ap->a_pvp->v_mount, new_ino, &vp)) + return (error); - fs = ip->i_fs; - if (indx % fs->fs_maxbpg == 0 || bap[indx - 1] == 0) { - if (lbn < NDADDR) { - cg = itog(fs, ip->i_number); - return (fs->fs_fpg * cg + fs->fs_frag); - } - /* - * Find a cylinder with greater than average number of - * unused data blocks. - */ - if (indx == 0 || bap[indx - 1] == 0) - startcg = itog(fs, ip->i_number) + lbn / fs->fs_maxbpg; - else - startcg = dtog(fs, bap[indx - 1]) + 1; - startcg %= fs->fs_ncg; - avgbfree = fs->fs_cstotal.cs_nbfree / fs->fs_ncg; - for (cg = startcg; cg < fs->fs_ncg; cg++) - if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { - fs->fs_cgrotor = cg; - return (fs->fs_fpg * cg + fs->fs_frag); - } - for (cg = 0; cg <= startcg; cg++) - if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { - fs->fs_cgrotor = cg; - return (fs->fs_fpg * cg + fs->fs_frag); - } - return (NULL); - } - /* - * One or more previous blocks have been laid out. If less - * than fs_maxcontig previous blocks are contiguous, the - * next block is requested contiguously, otherwise it is - * requested rotationally delayed by fs_rotdelay milliseconds. - */ - nextblk = bap[indx - 1] + fs->fs_frag; - if (indx > fs->fs_maxcontig && - bap[indx - fs->fs_maxcontig] + blkstofrags(fs, fs->fs_maxcontig) - != nextblk) - return (nextblk); - if (fs->fs_rotdelay != 0) - /* - * Here we convert ms of delay to frags as: - * (frags) = (ms) * (rev/sec) * (sect/rev) / - * ((sect/frag) * (ms/sec)) - * then round up to the next block. - */ - nextblk += roundup(fs->fs_rotdelay * fs->fs_rps * fs->fs_nsect / - (NSPF(fs) * 1000), fs->fs_frag); - return (nextblk); -} -/* - * Implement the cylinder overflow algorithm. - * - * The policy implemented by this algorithm is: - * 1) allocate the block in its requested cylinder group. - * 2) quadradically rehash on the cylinder group number. - * 3) brute force search for a free block. - */ -/*VARARGS5*/ -u_long -hashalloc(ip, cg, pref, size, allocator) - struct inode *ip; - int cg; - long pref; - int size; /* size for data blocks, mode for inodes */ - u_long (*allocator)(); -{ - register struct fs *fs; - long result; - int i, icg = cg; + ip = VTOI(vp); + /* Zero out the direct and indirect block addresses. */ + bzero(&ip->i_din, sizeof(struct dinode)); + ip->i_din.di_inumber = new_ino; - fs = ip->i_fs; - /* - * 1: preferred cylinder group - */ - result = (*allocator)(ip, cg, pref, size); - if (result) - return (result); - /* - * 2: quadratic rehash - */ - for (i = 1; i < fs->fs_ncg; i *= 2) { - cg += i; - if (cg >= fs->fs_ncg) - cg -= fs->fs_ncg; - result = (*allocator)(ip, cg, 0, size); - if (result) - return (result); - } - /* - * 3: brute force search - * Note that we start at i == 2, since 0 was checked initially, - * and 1 is always checked in the quadratic rehash. - */ - cg = (icg + 2) % fs->fs_ncg; - for (i = 2; i < fs->fs_ncg; i++) { - result = (*allocator)(ip, cg, 0, size); - if (result) - return (result); - cg++; - if (cg == fs->fs_ncg) - cg = 0; - } - return (NULL); -} + /* Set a new generation number for this inode. */ + if (++nextgennumber < (u_long)time.tv_sec) + nextgennumber = time.tv_sec; + ip->i_gen = nextgennumber; -/* - * Determine whether a fragment can be extended. - * - * Check to see if the necessary fragments are available, and - * if they are, allocate them. - */ -daddr_t -fragextend(ip, cg, bprev, osize, nsize) - struct inode *ip; - int cg; - long bprev; - int osize, nsize; -{ - register struct fs *fs; - register struct cg *cgp; - struct buf *bp; - long bno; - int frags, bbase; - int i, error; + /* Insert into the inode hash table. */ + ufs_ihashins(ip); - fs = ip->i_fs; - if (fs->fs_cs(fs, cg).cs_nffree < numfrags(fs, nsize - osize)) - return (NULL); - frags = numfrags(fs, nsize); - bbase = fragnum(fs, bprev); - if (bbase > fragnum(fs, (bprev + frags - 1))) { - /* cannot extend across a block boundary */ - return (NULL); - } -#ifdef SECSIZE - bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), (int)fs->fs_cgsize, - fs->fs_dbsize); -#else SECSIZE - error = bread(ip->i_devvp, fsbtodb(fs, cgtod(fs, cg)), - (int)fs->fs_cgsize, NOCRED, &bp); - if (error) { - brelse(bp); - return (NULL); - } -#endif SECSIZE - cgp = bp->b_un.b_cg; - if (!cg_chkmagic(cgp)) { - brelse(bp); - return (NULL); - } - cgp->cg_time = time.tv_sec; - bno = dtogd(fs, bprev); - for (i = numfrags(fs, osize); i < frags; i++) - if (isclr(cg_blksfree(cgp), bno + i)) { - brelse(bp); - return (NULL); - } - /* - * the current fragment can be extended - * deduct the count on fragment being extended into - * increase the count on the remaining fragment (if any) - * allocate the extended piece - */ - for (i = frags; i < fs->fs_frag - bbase; i++) - if (isclr(cg_blksfree(cgp), bno + i)) - break; - cgp->cg_frsum[i - numfrags(fs, osize)]--; - if (i != frags) - cgp->cg_frsum[i - frags]++; - for (i = numfrags(fs, osize); i < frags; i++) { - clrbit(cg_blksfree(cgp), bno + i); - cgp->cg_cs.cs_nffree--; - fs->fs_cstotal.cs_nffree--; - fs->fs_cs(fs, cg).cs_nffree--; + if (error = ufs_vinit(vp->v_mount, lfs_specop_p, LFS_FIFOOPS, &vp)) { + vput(vp); + *ap->a_vpp = NULL; + return (error); } - fs->fs_fmod++; - bdwrite(bp); - return (bprev); -} -/* - * Determine whether a block can be allocated. - * - * Check to see if a block of the apprpriate size is available, - * and if it is, allocate it. - */ -daddr_t -alloccg(ip, cg, bpref, size) - struct inode *ip; - int cg; - daddr_t bpref; - int size; -{ - register struct fs *fs; - register struct cg *cgp; - struct buf *bp; - register int i; - int error, bno, frags, allocsiz; + *ap->a_vpp = vp; + vp->v_flag |= VDIROP; + VREF(ip->i_devvp); - fs = ip->i_fs; - if (fs->fs_cs(fs, cg).cs_nbfree == 0 && size == fs->fs_bsize) - return (NULL); -#ifdef SECSIZE - bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), (int)fs->fs_cgsize, - fs->fs_dbsize); -#else SECSIZE - error = bread(ip->i_devvp, fsbtodb(fs, cgtod(fs, cg)), - (int)fs->fs_cgsize, NOCRED, &bp); - if (error) { - brelse(bp); - return (NULL); - } -#endif SECSIZE - cgp = bp->b_un.b_cg; - if (!cg_chkmagic(cgp) || - (cgp->cg_cs.cs_nbfree == 0 && size == fs->fs_bsize)) { - brelse(bp); - return (NULL); - } - cgp->cg_time = time.tv_sec; - if (size == fs->fs_bsize) { - bno = alloccgblk(fs, cgp, bpref); - bdwrite(bp); - return (bno); - } - /* - * check to see if any fragments are already available - * allocsiz is the size which will be allocated, hacking - * it down to a smaller size if necessary - */ - frags = numfrags(fs, size); - for (allocsiz = frags; allocsiz < fs->fs_frag; allocsiz++) - if (cgp->cg_frsum[allocsiz] != 0) - break; - if (allocsiz == fs->fs_frag) { - /* - * no fragments were available, so a block will be - * allocated, and hacked up - */ - if (cgp->cg_cs.cs_nbfree == 0) { - brelse(bp); - return (NULL); - } - bno = alloccgblk(fs, cgp, bpref); - bpref = dtogd(fs, bno); - for (i = frags; i < fs->fs_frag; i++) - setbit(cg_blksfree(cgp), bpref + i); - i = fs->fs_frag - frags; - cgp->cg_cs.cs_nffree += i; - fs->fs_cstotal.cs_nffree += i; - fs->fs_cs(fs, cg).cs_nffree += i; - fs->fs_fmod++; - cgp->cg_frsum[i]++; - bdwrite(bp); - return (bno); - } - bno = mapsearch(fs, cgp, bpref, allocsiz); - if (bno < 0) { - brelse(bp); - return (NULL); - } - for (i = 0; i < frags; i++) - clrbit(cg_blksfree(cgp), bno + i); - cgp->cg_cs.cs_nffree -= frags; - fs->fs_cstotal.cs_nffree -= frags; - fs->fs_cs(fs, cg).cs_nffree -= frags; - fs->fs_fmod++; - cgp->cg_frsum[allocsiz]--; - if (frags != allocsiz) - cgp->cg_frsum[allocsiz - frags]++; - bdwrite(bp); - return (cg * fs->fs_fpg + bno); + /* Set superblock modified bit and increment file count. */ + fs->lfs_fmod = 1; + ++fs->lfs_nfiles; + return (0); } -/* - * Allocate a block in a cylinder group. - * - * This algorithm implements the following policy: - * 1) allocate the requested block. - * 2) allocate a rotationally optimal block in the same cylinder. - * 3) allocate the next available block on the block rotor for the - * specified cylinder group. - * Note that this routine only allocates fs_bsize blocks; these - * blocks may be fragmented by the routine that allocates them. - */ -daddr_t -alloccgblk(fs, cgp, bpref) - register struct fs *fs; - register struct cg *cgp; - daddr_t bpref; +/* Create a new vnode/inode pair and initialize what fields we can. */ +int +lfs_vcreate(mp, ino, vpp) + struct mount *mp; + ino_t ino; + struct vnode **vpp; { - daddr_t bno; - int cylno, pos, delta; - short *cylbp; - register int i; + extern int (**lfs_vnodeop_p)(); + struct inode *ip; + struct ufsmount *ump; + int error, i; - if (bpref == 0) { - bpref = cgp->cg_rotor; - goto norot; - } - bpref = blknum(fs, bpref); - bpref = dtogd(fs, bpref); - /* - * if the requested block is available, use it - */ - if (isblock(fs, cg_blksfree(cgp), fragstoblks(fs, bpref))) { - bno = bpref; - goto gotit; - } - /* - * check for a block available on the same cylinder - */ - cylno = cbtocylno(fs, bpref); - if (cg_blktot(cgp)[cylno] == 0) - goto norot; - if (fs->fs_cpc == 0) { - /* - * block layout info is not available, so just have - * to take any block in this cylinder. - */ - bpref = howmany(fs->fs_spc * cylno, NSPF(fs)); - goto norot; - } - /* - * check the summary information to see if a block is - * available in the requested cylinder starting at the - * requested rotational position and proceeding around. - */ - cylbp = cg_blks(fs, cgp, cylno); - pos = cbtorpos(fs, bpref); - for (i = pos; i < fs->fs_nrpos; i++) - if (cylbp[i] > 0) - break; - if (i == fs->fs_nrpos) - for (i = 0; i < pos; i++) - if (cylbp[i] > 0) - break; - if (cylbp[i] > 0) { - /* - * found a rotational position, now find the actual - * block. A panic if none is actually there. - */ - pos = cylno % fs->fs_cpc; - bno = (cylno - pos) * fs->fs_spc / NSPB(fs); - if (fs_postbl(fs, pos)[i] == -1) { - printf("pos = %d, i = %d, fs = %s\n", - pos, i, fs->fs_fsmnt); - panic("alloccgblk: cyl groups corrupted"); - } - for (i = fs_postbl(fs, pos)[i];; ) { - if (isblock(fs, cg_blksfree(cgp), bno + i)) { - bno = blkstofrags(fs, (bno + i)); - goto gotit; - } - delta = fs_rotbl(fs)[i]; - if (delta <= 0 || - delta + i > fragstoblks(fs, fs->fs_fpg)) - break; - i += delta; - } - printf("pos = %d, i = %d, fs = %s\n", pos, i, fs->fs_fsmnt); - panic("alloccgblk: can't find blk in cyl"); + /* Create the vnode. */ + if (error = getnewvnode(VT_LFS, mp, lfs_vnodeop_p, vpp)) { + *vpp = NULL; + return (error); } -norot: - /* - * no blocks in the requested cylinder, so take next - * available one in this cylinder group. - */ - bno = mapsearch(fs, cgp, bpref, (int)fs->fs_frag); - if (bno < 0) - return (NULL); - cgp->cg_rotor = bno; -gotit: - clrblock(fs, cg_blksfree(cgp), (long)fragstoblks(fs, bno)); - cgp->cg_cs.cs_nbfree--; - fs->fs_cstotal.cs_nbfree--; - fs->fs_cs(fs, cgp->cg_cgx).cs_nbfree--; - cylno = cbtocylno(fs, bno); - cg_blks(fs, cgp, cylno)[cbtorpos(fs, bno)]--; - cg_blktot(cgp)[cylno]--; - fs->fs_fmod++; - return (cgp->cg_cgx * fs->fs_fpg + bno); -} -/* - * Determine whether an inode can be allocated. - * - * Check to see if an inode is available, and if it is, - * allocate it using the following policy: - * 1) allocate the requested inode. - * 2) allocate the next available inode after the requested - * inode in the specified cylinder group. - */ -ino_t -ialloccg(ip, cg, ipref, mode) - struct inode *ip; - int cg; - daddr_t ipref; - int mode; -{ - register struct fs *fs; - register struct cg *cgp; - struct buf *bp; - int error, start, len, loc, map, i; + /* Get a pointer to the private mount structure. */ + ump = VFSTOUFS(mp); - fs = ip->i_fs; - if (fs->fs_cs(fs, cg).cs_nifree == 0) - return (NULL); -#ifdef SECSIZE - bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), (int)fs->fs_cgsize, - fs->fs_dbsize); -#else SECSIZE - error = bread(ip->i_devvp, fsbtodb(fs, cgtod(fs, cg)), - (int)fs->fs_cgsize, NOCRED, &bp); - if (error) { - brelse(bp); - return (NULL); - } -#endif SECSIZE - cgp = bp->b_un.b_cg; - if (!cg_chkmagic(cgp) || cgp->cg_cs.cs_nifree == 0) { - brelse(bp); - return (NULL); - } - cgp->cg_time = time.tv_sec; - if (ipref) { - ipref %= fs->fs_ipg; - if (isclr(cg_inosused(cgp), ipref)) - goto gotit; - } - start = cgp->cg_irotor / NBBY; - len = howmany(fs->fs_ipg - cgp->cg_irotor, NBBY); - loc = skpc(0xff, len, &cg_inosused(cgp)[start]); - if (loc == 0) { - len = start + 1; - start = 0; - loc = skpc(0xff, len, &cg_inosused(cgp)[0]); - if (loc == 0) { - printf("cg = %s, irotor = %d, fs = %s\n", - cg, cgp->cg_irotor, fs->fs_fsmnt); - panic("ialloccg: map corrupted"); - /* NOTREACHED */ - } - } - i = start + len - loc; - map = cg_inosused(cgp)[i]; - ipref = i * NBBY; - for (i = 1; i < (1 << NBBY); i <<= 1, ipref++) { - if ((map & i) == 0) { - cgp->cg_irotor = ipref; - goto gotit; - } - } - printf("fs = %s\n", fs->fs_fsmnt); - panic("ialloccg: block not in map"); - /* NOTREACHED */ -gotit: - setbit(cg_inosused(cgp), ipref); - cgp->cg_cs.cs_nifree--; - fs->fs_cstotal.cs_nifree--; - fs->fs_cs(fs, cg).cs_nifree--; - fs->fs_fmod++; - if ((mode & IFMT) == IFDIR) { - cgp->cg_cs.cs_ndir++; - fs->fs_cstotal.cs_ndir++; - fs->fs_cs(fs, cg).cs_ndir++; - } - bdwrite(bp); - return (cg * fs->fs_ipg + ipref); + /* Initialize the inode. */ + MALLOC(ip, struct inode *, sizeof(struct inode), M_LFSNODE, M_WAITOK); + lockinit(&ip->i_lock, PINOD, "lfsinode", 0, 0); + (*vpp)->v_data = ip; + ip->i_vnode = *vpp; + ip->i_devvp = ump->um_devvp; + ip->i_flag = IN_MODIFIED; + ip->i_dev = ump->um_dev; + ip->i_number = ip->i_din.di_inumber = ino; + ip->i_lfs = ump->um_lfs; +#ifdef QUOTA + for (i = 0; i < MAXQUOTAS; i++) + ip->i_dquot[i] = NODQUOT; +#endif + ip->i_lockf = 0; + ip->i_diroff = 0; + ip->i_mode = 0; + ip->i_size = 0; + ip->i_blocks = 0; + ++ump->um_lfs->lfs_uinodes; + return (0); } -/* - * Free a block or fragment. - * - * The specified block or fragment is placed back in the - * free map. If a fragment is deallocated, a possible - * block reassembly is checked. - */ -blkfree(ip, bno, size) - register struct inode *ip; - daddr_t bno; - off_t size; +/* Free an inode. */ +/* ARGUSED */ +int +lfs_vfree(ap) + struct vop_vfree_args /* { + struct vnode *a_pvp; + ino_t a_ino; + int a_mode; + } */ *ap; { - register struct fs *fs; - register struct cg *cgp; + SEGUSE *sup; struct buf *bp; - int error, cg, blk, frags, bbase; - register int i; - struct ucred *cred = curproc->p_ucred; /* XXX */ - - fs = ip->i_fs; - if ((unsigned)size > fs->fs_bsize || fragoff(fs, size) != 0) { - printf("dev = 0x%x, bsize = %d, size = %d, fs = %s\n", - ip->i_dev, fs->fs_bsize, size, fs->fs_fsmnt); - panic("blkfree: bad size"); - } - cg = dtog(fs, bno); - if ((unsigned)bno >= fs->fs_size) { - printf("bad block %d, ino %d\n", bno, ip->i_number); - fserr(fs, cred->cr_uid, "bad block"); - return; - } -#ifdef SECSIZE - bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), (int)fs->fs_cgsize, - fs->fs_dbsize); -#else SECSIZE - error = bread(ip->i_devvp, fsbtodb(fs, cgtod(fs, cg)), - (int)fs->fs_cgsize, NOCRED, &bp); - if (error) { - brelse(bp); - return; - } -#endif SECSIZE - cgp = bp->b_un.b_cg; - if (!cg_chkmagic(cgp)) { - brelse(bp); - return; - } - cgp->cg_time = time.tv_sec; - bno = dtogd(fs, bno); - if (size == fs->fs_bsize) { - if (isblock(fs, cg_blksfree(cgp), fragstoblks(fs, bno))) { - printf("dev = 0x%x, block = %d, fs = %s\n", - ip->i_dev, bno, fs->fs_fsmnt); - panic("blkfree: freeing free block"); - } - setblock(fs, cg_blksfree(cgp), fragstoblks(fs, bno)); - cgp->cg_cs.cs_nbfree++; - fs->fs_cstotal.cs_nbfree++; - fs->fs_cs(fs, cg).cs_nbfree++; - i = cbtocylno(fs, bno); - cg_blks(fs, cgp, i)[cbtorpos(fs, bno)]++; - cg_blktot(cgp)[i]++; - } else { - bbase = bno - fragnum(fs, bno); - /* - * decrement the counts associated with the old frags - */ - blk = blkmap(fs, cg_blksfree(cgp), bbase); - fragacct(fs, blk, cgp->cg_frsum, -1); - /* - * deallocate the fragment - */ - frags = numfrags(fs, size); - for (i = 0; i < frags; i++) { - if (isset(cg_blksfree(cgp), bno + i)) { - printf("dev = 0x%x, block = %d, fs = %s\n", - ip->i_dev, bno + i, fs->fs_fsmnt); - panic("blkfree: freeing free frag"); - } - setbit(cg_blksfree(cgp), bno + i); - } - cgp->cg_cs.cs_nffree += i; - fs->fs_cstotal.cs_nffree += i; - fs->fs_cs(fs, cg).cs_nffree += i; - /* - * add back in counts associated with the new frags - */ - blk = blkmap(fs, cg_blksfree(cgp), bbase); - fragacct(fs, blk, cgp->cg_frsum, 1); - /* - * if a complete block has been reassembled, account for it - */ - if (isblock(fs, cg_blksfree(cgp), - (daddr_t)fragstoblks(fs, bbase))) { - cgp->cg_cs.cs_nffree -= fs->fs_frag; - fs->fs_cstotal.cs_nffree -= fs->fs_frag; - fs->fs_cs(fs, cg).cs_nffree -= fs->fs_frag; - cgp->cg_cs.cs_nbfree++; - fs->fs_cstotal.cs_nbfree++; - fs->fs_cs(fs, cg).cs_nbfree++; - i = cbtocylno(fs, bbase); - cg_blks(fs, cgp, i)[cbtorpos(fs, bbase)]++; - cg_blktot(cgp)[i]++; - } - } - fs->fs_fmod++; - bdwrite(bp); -} - -/* - * Free an inode. - * - * The specified inode is placed back in the free map. - */ -ifree(ip, ino, mode) + struct ifile *ifp; struct inode *ip; + struct lfs *fs; + ufs_daddr_t old_iaddr; ino_t ino; - int mode; -{ - register struct fs *fs; - register struct cg *cgp; - struct buf *bp; - int error, cg; - fs = ip->i_fs; - if ((unsigned)ino >= fs->fs_ipg*fs->fs_ncg) { - printf("dev = 0x%x, ino = %d, fs = %s\n", - ip->i_dev, ino, fs->fs_fsmnt); - panic("ifree: range"); - } - cg = itog(fs, ino); -#ifdef SECSIZE - bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), (int)fs->fs_cgsize, - fs->fs_dbsize); -#else SECSIZE - error = bread(ip->i_devvp, fsbtodb(fs, cgtod(fs, cg)), - (int)fs->fs_cgsize, NOCRED, &bp); - if (error) { - brelse(bp); - return; - } -#endif SECSIZE - cgp = bp->b_un.b_cg; - if (!cg_chkmagic(cgp)) { - brelse(bp); - return; - } - cgp->cg_time = time.tv_sec; - ino %= fs->fs_ipg; - if (isclr(cg_inosused(cgp), ino)) { - printf("dev = 0x%x, ino = %d, fs = %s\n", - ip->i_dev, ino, fs->fs_fsmnt); - if (fs->fs_ronly == 0) - panic("ifree: freeing free inode"); - } - clrbit(cg_inosused(cgp), ino); - if (ino < cgp->cg_irotor) - cgp->cg_irotor = ino; - cgp->cg_cs.cs_nifree++; - fs->fs_cstotal.cs_nifree++; - fs->fs_cs(fs, cg).cs_nifree++; - if ((mode & IFMT) == IFDIR) { - cgp->cg_cs.cs_ndir--; - fs->fs_cstotal.cs_ndir--; - fs->fs_cs(fs, cg).cs_ndir--; - } - fs->fs_fmod++; - bdwrite(bp); -} - -/* - * Find a block of the specified size in the specified cylinder group. - * - * It is a panic if a request is made to find a block if none are - * available. - */ -daddr_t -mapsearch(fs, cgp, bpref, allocsiz) - register struct fs *fs; - register struct cg *cgp; - daddr_t bpref; - int allocsiz; -{ - daddr_t bno; - int start, len, loc, i; - int blk, field, subfield, pos; - - /* - * find the fragment by searching through the free block - * map for an appropriate bit pattern - */ - if (bpref) - start = dtogd(fs, bpref) / NBBY; - else - start = cgp->cg_frotor / NBBY; - len = howmany(fs->fs_fpg, NBBY) - start; - loc = scanc((unsigned)len, (u_char *)&cg_blksfree(cgp)[start], - (u_char *)fragtbl[fs->fs_frag], - (u_char)(1 << (allocsiz - 1 + (fs->fs_frag % NBBY)))); - if (loc == 0) { - len = start + 1; - start = 0; - loc = scanc((unsigned)len, (u_char *)&cg_blksfree(cgp)[0], - (u_char *)fragtbl[fs->fs_frag], - (u_char)(1 << (allocsiz - 1 + (fs->fs_frag % NBBY)))); - if (loc == 0) { - printf("start = %d, len = %d, fs = %s\n", - start, len, fs->fs_fsmnt); - panic("alloccg: map corrupted"); - /* NOTREACHED */ - } - } - bno = (start + len - loc) * NBBY; - cgp->cg_frotor = bno; - /* - * found the byte in the map - * sift through the bits to find the selected frag - */ - for (i = bno + NBBY; bno < i; bno += fs->fs_frag) { - blk = blkmap(fs, cg_blksfree(cgp), bno); - blk <<= 1; - field = around[allocsiz]; - subfield = inside[allocsiz]; - for (pos = 0; pos <= fs->fs_frag - allocsiz; pos++) { - if ((blk & field) == subfield) - return (bno + pos); - field <<= 1; - subfield <<= 1; - } + /* Get the inode number and file system. */ + ip = VTOI(ap->a_pvp); + fs = ip->i_lfs; + ino = ip->i_number; + if (ip->i_flag & IN_MODIFIED) { + --fs->lfs_uinodes; + ip->i_flag &= + ~(IN_ACCESS | IN_CHANGE | IN_MODIFIED | IN_UPDATE); + } + /* + * Set the ifile's inode entry to unused, increment its version number + * and link it into the free chain. + */ + LFS_IENTRY(ifp, fs, ino, bp); + old_iaddr = ifp->if_daddr; + ifp->if_daddr = LFS_UNUSED_DADDR; + ++ifp->if_version; + ifp->if_nextfree = fs->lfs_free; + fs->lfs_free = ino; + (void) VOP_BWRITE(bp); + + if (old_iaddr != LFS_UNUSED_DADDR) { + LFS_SEGENTRY(sup, fs, datosn(fs, old_iaddr), bp); +#ifdef DIAGNOSTIC + if (sup->su_nbytes < sizeof(struct dinode)) + panic("lfs_vfree: negative byte count (segment %d)\n", + datosn(fs, old_iaddr)); +#endif + sup->su_nbytes -= sizeof(struct dinode); + (void) VOP_BWRITE(bp); } - printf("bno = %d, fs = %s\n", bno, fs->fs_fsmnt); - panic("alloccg: block not in map"); - return (-1); -} - -/* - * Fserr prints the name of a file system with an error diagnostic. - * - * The form of the error message is: - * fs: error message - */ -fserr(fs, uid, cp) - struct fs *fs; - uid_t uid; - char *cp; -{ - log(LOG_ERR, "uid %d on %s: %s\n", uid, fs->fs_fsmnt, cp); + /* Set superblock modified bit and decrement file count. */ + fs->lfs_fmod = 1; + --fs->lfs_nfiles; + return (0); }