/* $NetBSD: lfs_alloc.c,v 1.137.8.1 2020/08/17 10:30:22 martin Exp $ */
/*-
* Copyright (c) 1999, 2000, 2001, 2002, 2003, 2007 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Konrad E. Schroder <perseant@hhhh.org>.
*
* 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.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. 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 FOUNDATION 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.
*/
/*
* Copyright (c) 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. 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.
*
* @(#)lfs_alloc.c 8.4 (Berkeley) 1/4/94
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: lfs_alloc.c,v 1.137.8.1 2020/08/17 10:30:22 martin Exp $");
#if defined(_KERNEL_OPT)
#include "opt_quota.h"
#endif
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/buf.h>
#include <sys/lock.h>
#include <sys/vnode.h>
#include <sys/syslog.h>
#include <sys/mount.h>
#include <sys/malloc.h>
#include <sys/pool.h>
#include <sys/proc.h>
#include <sys/kauth.h>
#include <ufs/lfs/ulfs_quotacommon.h>
#include <ufs/lfs/ulfs_inode.h>
#include <ufs/lfs/ulfsmount.h>
#include <ufs/lfs/ulfs_extern.h>
#include <ufs/lfs/lfs.h>
#include <ufs/lfs/lfs_accessors.h>
#include <ufs/lfs/lfs_extern.h>
#include <ufs/lfs/lfs_kernel.h>
/* Constants for inode free bitmap */
#define BMSHIFT 5 /* 2 ** 5 = 32 */
#define BMMASK ((1 << BMSHIFT) - 1)
#define SET_BITMAP_FREE(F, I) do { \
DLOG((DLOG_ALLOC, "lfs: ino %d wrd %d bit %d set\n", (int)(I), \
(int)((I) >> BMSHIFT), (int)((I) & BMMASK))); \
(F)->lfs_ino_bitmap[(I) >> BMSHIFT] |= (1 << ((I) & BMMASK)); \
} while (0)
#define CLR_BITMAP_FREE(F, I) do { \
DLOG((DLOG_ALLOC, "lfs: ino %d wrd %d bit %d clr\n", (int)(I), \
(int)((I) >> BMSHIFT), (int)((I) & BMMASK))); \
(F)->lfs_ino_bitmap[(I) >> BMSHIFT] &= ~(1 << ((I) & BMMASK)); \
} while(0)
#define ISSET_BITMAP_FREE(F, I) \
((F)->lfs_ino_bitmap[(I) >> BMSHIFT] & (1 << ((I) & BMMASK)))
/*
* Add a new block to the Ifile, to accommodate future file creations.
* Called with the segment lock held.
*/
int
lfs_extend_ifile(struct lfs *fs, kauth_cred_t cred)
{
struct vnode *vp;
struct inode *ip;
IFILE64 *ifp64;
IFILE32 *ifp32;
IFILE_V1 *ifp_v1;
struct buf *bp, *cbp;
int error;
daddr_t i, blkno, xmax;
ino_t oldlast, maxino;
CLEANERINFO *cip;
ASSERT_SEGLOCK(fs);
/* XXX should check or assert that we aren't readonly. */
/*
* Get a block and extend the ifile inode. Leave the buffer for
* the block in bp.
*/
vp = fs->lfs_ivnode;
ip = VTOI(vp);
blkno = lfs_lblkno(fs, ip->i_size);
if ((error = lfs_balloc(vp, ip->i_size, lfs_sb_getbsize(fs), cred, 0,
&bp)) != 0) {
return (error);
}
ip->i_size += lfs_sb_getbsize(fs);
lfs_dino_setsize(fs, ip->i_din, ip->i_size);
uvm_vnp_setsize(vp, ip->i_size);
/*
* Compute the new number of inodes, and reallocate the in-memory
* inode freemap.
*/
maxino = ((ip->i_size >> lfs_sb_getbshift(fs)) - lfs_sb_getcleansz(fs) -
lfs_sb_getsegtabsz(fs)) * lfs_sb_getifpb(fs);
fs->lfs_ino_bitmap = (lfs_bm_t *)
realloc(fs->lfs_ino_bitmap, ((maxino + BMMASK) >> BMSHIFT) *
sizeof(lfs_bm_t), M_SEGMENT, M_WAITOK);
KASSERT(fs->lfs_ino_bitmap != NULL);
/* first new inode number */
i = (blkno - lfs_sb_getsegtabsz(fs) - lfs_sb_getcleansz(fs)) *
lfs_sb_getifpb(fs);
/*
* We insert the new inodes at the head of the free list.
* Under normal circumstances, the free list is empty here,
* so we are also incidentally placing them at the end (which
* we must do if we are to keep them in order).
*/
LFS_GET_HEADFREE(fs, cip, cbp, &oldlast);
LFS_PUT_HEADFREE(fs, cip, cbp, i);
KASSERTMSG((lfs_sb_getfreehd(fs) != LFS_UNUSED_INUM),
"inode 0 allocated [2]");
/* inode number to stop at (XXX: why *x*max?) */
xmax = i + lfs_sb_getifpb(fs);
/*
* Initialize the ifile block.
*
* XXX: these loops should be restructured to use the accessor
* functions instead of using cutpaste polymorphism.
*/
if (fs->lfs_is64) {
for (ifp64 = (IFILE64 *)bp->b_data; i < xmax; ++ifp64) {
SET_BITMAP_FREE(fs, i);
ifp64->if_version = 1;
ifp64->if_daddr = LFS_UNUSED_DADDR;
ifp64->if_nextfree = ++i;
}
ifp64--;
ifp64->if_nextfree = oldlast;
} else if (lfs_sb_getversion(fs) > 1) {
for (ifp32 = (IFILE32 *)bp->b_data; i < xmax; ++ifp32) {
SET_BITMAP_FREE(fs, i);
ifp32->if_version = 1;
ifp32->if_daddr = LFS_UNUSED_DADDR;
ifp32->if_nextfree = ++i;
}
ifp32--;
ifp32->if_nextfree = oldlast;
} else {
for (ifp_v1 = (IFILE_V1 *)bp->b_data; i < xmax; ++ifp_v1) {
SET_BITMAP_FREE(fs, i);
ifp_v1->if_version = 1;
ifp_v1->if_daddr = LFS_UNUSED_DADDR;
ifp_v1->if_nextfree = ++i;
}
ifp_v1--;
ifp_v1->if_nextfree = oldlast;
}
LFS_PUT_TAILFREE(fs, cip, cbp, xmax - 1);
/*
* Write out the new block.
*/
(void) LFS_BWRITE_LOG(bp); /* Ifile */
return 0;
}
/*
* Allocate an inode for a new file.
*
* Takes the segment lock. Also (while holding it) takes lfs_lock
* to frob fs->lfs_fmod.
*
* XXX: the mode argument is unused; should just get rid of it.
*/
/* ARGSUSED */
/* VOP_BWRITE 2i times */
int
lfs_valloc(struct vnode *pvp, int mode, kauth_cred_t cred,
ino_t *ino, int *gen)
{
struct lfs *fs;
struct buf *bp, *cbp;
IFILE *ifp;
int error;
CLEANERINFO *cip;
fs = VTOI(pvp)->i_lfs;
if (fs->lfs_ronly)
return EROFS;
ASSERT_NO_SEGLOCK(fs);
lfs_seglock(fs, SEGM_PROT);
/* Get the head of the freelist. */
LFS_GET_HEADFREE(fs, cip, cbp, ino);
/* paranoia */
KASSERT(*ino != LFS_UNUSED_INUM && *ino != LFS_IFILE_INUM);
DLOG((DLOG_ALLOC, "lfs_valloc: allocate inode %" PRId64 "\n",
*ino));
/* Update the in-memory inode freemap */
CLR_BITMAP_FREE(fs, *ino);
/*
* Fetch the ifile entry and make sure the inode is really
* free.
*/
LFS_IENTRY(ifp, fs, *ino, bp);
if (lfs_if_getdaddr(fs, ifp) != LFS_UNUSED_DADDR)
panic("lfs_valloc: inuse inode %" PRId64 " on the free list",
*ino);
/* Update the inode freelist head in the superblock. */
LFS_PUT_HEADFREE(fs, cip, cbp, lfs_if_getnextfree(fs, ifp));
DLOG((DLOG_ALLOC, "lfs_valloc: headfree %" PRId64 " -> %ju\n",
*ino, (uintmax_t)lfs_if_getnextfree(fs, ifp)));
/*
* Retrieve the version number from the ifile entry. It was
* bumped by vfree, so don't bump it again.
*/
*gen = lfs_if_getversion(fs, ifp);
/* Done with ifile entry */
brelse(bp, 0);
if (lfs_sb_getfreehd(fs) == LFS_UNUSED_INUM) {
/*
* No more inodes; extend the ifile so that the next
* lfs_valloc will succeed.
*/
if ((error = lfs_extend_ifile(fs, cred)) != 0) {
/* restore the freelist */
LFS_PUT_HEADFREE(fs, cip, cbp, *ino);
/* unlock and return */
lfs_segunlock(fs);
return error;
}
}
KASSERTMSG((lfs_sb_getfreehd(fs) != LFS_UNUSED_INUM),
"inode 0 allocated [3]");
/* Set superblock modified bit */
mutex_enter(&lfs_lock);
fs->lfs_fmod = 1;
mutex_exit(&lfs_lock);
/* increment file count */
lfs_sb_addnfiles(fs, 1);
/* done */
lfs_segunlock(fs);
return 0;
}
/*
* Allocate an inode for a new file, with given inode number and
* version.
*
* Called in the same context as lfs_valloc and therefore shares the
* same locking assumptions.
*/
int
lfs_valloc_fixed(struct lfs *fs, ino_t ino, int vers)
{
IFILE *ifp;
struct buf *bp, *cbp;
ino_t headino, thisino, oldnext;
CLEANERINFO *cip;
if (fs->lfs_ronly)
return EROFS;
ASSERT_NO_SEGLOCK(fs);
lfs_seglock(fs, SEGM_PROT);
/*
* If the ifile is too short to contain this inum, extend it.
*
* XXX: lfs_extend_ifile should take a size instead of always
* doing just one block at time.
*/
while (VTOI(fs->lfs_ivnode)->i_size <= (ino /
lfs_sb_getifpb(fs) + lfs_sb_getcleansz(fs) + lfs_sb_getsegtabsz(fs))
<< lfs_sb_getbshift(fs)) {
lfs_extend_ifile(fs, NOCRED);
}
/*
* fetch the ifile entry; get the inode freelist next pointer,
* and set the version as directed.
*/
LFS_IENTRY(ifp, fs, ino, bp);
oldnext = lfs_if_getnextfree(fs, ifp);
lfs_if_setversion(fs, ifp, vers);
brelse(bp, 0);
/* Get head of inode freelist */
LFS_GET_HEADFREE(fs, cip, cbp, &headino);
if (headino == ino) {
/* Easy case: the inode we wanted was at the head */
LFS_PUT_HEADFREE(fs, cip, cbp, oldnext);
} else {
ino_t nextfree;
/* Have to find the desired inode in the freelist... */
thisino = headino;
while (1) {
/* read this ifile entry */
LFS_IENTRY(ifp, fs, thisino, bp);
nextfree = lfs_if_getnextfree(fs, ifp);
/* stop if we find it or we hit the end */
if (nextfree == ino ||
nextfree == LFS_UNUSED_INUM)
break;
/* nope, keep going... */
thisino = nextfree;
brelse(bp, 0);
}
if (nextfree == LFS_UNUSED_INUM) {
/* hit the end -- this inode is not available */
brelse(bp, 0);
lfs_segunlock(fs);
return ENOENT;
}
/* found it; update the next pointer */
lfs_if_setnextfree(fs, ifp, oldnext);
/* write the ifile block */
LFS_BWRITE_LOG(bp);
}
/* done */
lfs_segunlock(fs);
return 0;
}
#if 0
/*
* Find the highest-numbered allocated inode.
* This will be used to shrink the Ifile.
*/
static inline ino_t
lfs_last_alloc_ino(struct lfs *fs)
{
ino_t ino, maxino;
maxino = ((fs->lfs_ivnode->v_size >> lfs_sb_getbshift(fs)) -
lfs_sb_getcleansz(fs) - lfs_sb_getsegtabsz(fs)) *
lfs_sb_getifpb(fs);
for (ino = maxino - 1; ino > LFS_UNUSED_INUM; --ino) {
if (ISSET_BITMAP_FREE(fs, ino) == 0)
break;
}
return ino;
}
#endif
/*
* Find the previous (next lowest numbered) free inode, if any.
* If there is none, return LFS_UNUSED_INUM.
*
* XXX: locking?
*/
static inline ino_t
lfs_freelist_prev(struct lfs *fs, ino_t ino)
{
ino_t tino, bound, bb, freehdbb;
if (lfs_sb_getfreehd(fs) == LFS_UNUSED_INUM) {
/* No free inodes at all */
return LFS_UNUSED_INUM;
}
/* Search our own word first */
bound = ino & ~BMMASK;
for (tino = ino - 1; tino >= bound && tino > LFS_UNUSED_INUM; tino--)
if (ISSET_BITMAP_FREE(fs, tino))
return tino;
/* If there are no lower words to search, just return */
if (ino >> BMSHIFT == 0)
return LFS_UNUSED_INUM;
/*
* Find a word with a free inode in it. We have to be a bit
* careful here since ino_t is unsigned.
*/
freehdbb = (lfs_sb_getfreehd(fs) >> BMSHIFT);
for (bb = (ino >> BMSHIFT) - 1; bb >= freehdbb && bb > 0; --bb)
if (fs->lfs_ino_bitmap[bb])
break;
if (fs->lfs_ino_bitmap[bb] == 0)
return LFS_UNUSED_INUM;
/* Search the word we found */
for (tino = (bb << BMSHIFT) | BMMASK; tino >= (bb << BMSHIFT) &&
tino > LFS_UNUSED_INUM; tino--)
if (ISSET_BITMAP_FREE(fs, tino))
break;
/* Avoid returning reserved inode numbers */
if (tino <= LFS_IFILE_INUM)
tino = LFS_UNUSED_INUM;
return tino;
}
/*
* Free an inode.
*
* Takes lfs_seglock. Also (independently) takes vp->v_interlock.
*/
/* ARGUSED */
/* VOP_BWRITE 2i times */
int
lfs_vfree(struct vnode *vp, ino_t ino, int mode)
{
SEGUSE *sup;
CLEANERINFO *cip;
struct buf *cbp, *bp;
IFILE *ifp;
struct inode *ip;
struct lfs *fs;
daddr_t old_iaddr;
ino_t otail;
/* Get the inode number and file system. */
ip = VTOI(vp);
fs = ip->i_lfs;
ino = ip->i_number;
/* XXX: assert not readonly */
ASSERT_NO_SEGLOCK(fs);
DLOG((DLOG_ALLOC, "lfs_vfree: free ino %lld\n", (long long)ino));
/* Drain of pending writes */
mutex_enter(vp->v_interlock);
while (lfs_sb_getversion(fs) > 1 && WRITEINPROG(vp)) {
cv_wait(&vp->v_cv, vp->v_interlock);
}
mutex_exit(vp->v_interlock);
lfs_seglock(fs, SEGM_PROT);
/*
* If the inode was in a dirop, it isn't now.
*
* XXX: why are (v_uflag & VU_DIROP) and (ip->i_state & IN_ADIROP)
* not updated together in one function? (and why do both exist,
* anyway?)
*/
UNMARK_VNODE(vp);
mutex_enter(&lfs_lock);
if (vp->v_uflag & VU_DIROP) {
vp->v_uflag &= ~VU_DIROP;
--lfs_dirvcount;
--fs->lfs_dirvcount;
TAILQ_REMOVE(&fs->lfs_dchainhd, ip, i_lfs_dchain);
wakeup(&fs->lfs_dirvcount);
wakeup(&lfs_dirvcount);
mutex_exit(&lfs_lock);
vrele(vp);
/*
* If this inode is not going to be written any more, any
* segment accounting left over from its truncation needs
* to occur at the end of the next dirops flush. Attach
* them to the fs-wide list for that purpose.
*/
if (LIST_FIRST(&ip->i_lfs_segdhd) != NULL) {
struct segdelta *sd;
while((sd = LIST_FIRST(&ip->i_lfs_segdhd)) != NULL) {
LIST_REMOVE(sd, list);
LIST_INSERT_HEAD(&fs->lfs_segdhd, sd, list);
}
}
} else {
/*
* If it's not a dirop, we can finalize right away.
*/
mutex_exit(&lfs_lock);
lfs_finalize_ino_seguse(fs, ip);
}
/* it is no longer an unwritten inode, so update the counts */
mutex_enter(&lfs_lock);
LFS_CLR_UINO(ip, IN_ACCESSED|IN_CLEANING|IN_MODIFIED);
mutex_exit(&lfs_lock);
/* Turn off all inode modification flags */
ip->i_state &= ~IN_ALLMOD;
/* Mark it deleted */
ip->i_lfs_iflags |= LFSI_DELETED;
/* Mark it free in the in-memory inode freemap */
SET_BITMAP_FREE(fs, ino);
/*
* Set the ifile's inode entry to unused, increment its version number
* and link it onto the free chain.
*/
/* fetch the ifile entry */
LFS_IENTRY(ifp, fs, ino, bp);
/* update the on-disk address (to "nowhere") */
old_iaddr = lfs_if_getdaddr(fs, ifp);
lfs_if_setdaddr(fs, ifp, LFS_UNUSED_DADDR);
/* bump the version */
lfs_if_setversion(fs, ifp, lfs_if_getversion(fs, ifp) + 1);
if (lfs_sb_getversion(fs) == 1) {
ino_t nextfree;
/* insert on freelist */
LFS_GET_HEADFREE(fs, cip, cbp, &nextfree);
lfs_if_setnextfree(fs, ifp, nextfree);
LFS_PUT_HEADFREE(fs, cip, cbp, ino);
/* write the ifile block */
(void) LFS_BWRITE_LOG(bp); /* Ifile */
} else {
ino_t tino, onf;
/*
* Clear the freelist next pointer and write the ifile
* block. XXX: why? I'm sure there must be a reason but
* it seems both silly and dangerous.
*/
lfs_if_setnextfree(fs, ifp, LFS_UNUSED_INUM);
(void) LFS_BWRITE_LOG(bp); /* Ifile */
/*
* Insert on freelist in order.
*/
/* Find the next lower (by number) free inode */
tino = lfs_freelist_prev(fs, ino);
if (tino == LFS_UNUSED_INUM) {
ino_t nextfree;
/*
* There isn't one; put us on the freelist head.
*/
/* reload the ifile block */
LFS_IENTRY(ifp, fs, ino, bp);
/* update the list */
LFS_GET_HEADFREE(fs, cip, cbp, &nextfree);
lfs_if_setnextfree(fs, ifp, nextfree);
LFS_PUT_HEADFREE(fs, cip, cbp, ino);
DLOG((DLOG_ALLOC, "lfs_vfree: headfree %lld -> %lld\n",
(long long)nextfree, (long long)ino));
/* write the ifile block */
LFS_BWRITE_LOG(bp); /* Ifile */
/* If the list was empty, set tail too */
LFS_GET_TAILFREE(fs, cip, cbp, &otail);
if (otail == LFS_UNUSED_INUM) {
LFS_PUT_TAILFREE(fs, cip, cbp, ino);
DLOG((DLOG_ALLOC, "lfs_vfree: tailfree %lld "
"-> %lld\n", (long long)otail,
(long long)ino));
}
} else {
/*
* Insert this inode into the list after tino.
* We hold the segment lock so we don't have to
* worry about blocks being written out of order.
*/
DLOG((DLOG_ALLOC, "lfs_vfree: insert ino %lld "
" after %lld\n", ino, tino));
/* load the previous inode's ifile block */
LFS_IENTRY(ifp, fs, tino, bp);
/* update the list pointer */
onf = lfs_if_getnextfree(fs, ifp);
lfs_if_setnextfree(fs, ifp, ino);
/* write the block */
LFS_BWRITE_LOG(bp); /* Ifile */
/* load this inode's ifile block */
LFS_IENTRY(ifp, fs, ino, bp);
/* update the list pointer */
lfs_if_setnextfree(fs, ifp, onf);
/* write the block */
LFS_BWRITE_LOG(bp); /* Ifile */
/* If we're last, put us on the tail */
if (onf == LFS_UNUSED_INUM) {
LFS_GET_TAILFREE(fs, cip, cbp, &otail);
LFS_PUT_TAILFREE(fs, cip, cbp, ino);
DLOG((DLOG_ALLOC, "lfs_vfree: tailfree %lld "
"-> %lld\n", (long long)otail,
(long long)ino));
}
}
}
/* XXX: shouldn't this check be further up *before* we trash the fs? */
KASSERTMSG((ino != LFS_UNUSED_INUM), "inode 0 freed");
/*
* Update the segment summary for the segment where the on-disk
* copy used to be.
*/
if (old_iaddr != LFS_UNUSED_DADDR) {
/* load it */
LFS_SEGENTRY(sup, fs, lfs_dtosn(fs, old_iaddr), bp);
/* the number of bytes in the segment should not become < 0 */
KASSERTMSG((sup->su_nbytes >= DINOSIZE(fs)),
"lfs_vfree: negative byte count"
" (segment %" PRIu32 " short by %d)\n",
lfs_dtosn(fs, old_iaddr),
(int)DINOSIZE(fs) - sup->su_nbytes);
/* update the number of bytes in the segment */
sup->su_nbytes -= DINOSIZE(fs);
/* write the segment entry */
LFS_WRITESEGENTRY(sup, fs, lfs_dtosn(fs, old_iaddr), bp); /* Ifile */
}
/* Set superblock modified bit. */
mutex_enter(&lfs_lock);
fs->lfs_fmod = 1;
mutex_exit(&lfs_lock);
/* Decrement file count. */
lfs_sb_subnfiles(fs, 1);
lfs_segunlock(fs);
return (0);
}
/*
* Sort the freelist and set up the free-inode bitmap.
* To be called by lfs_mountfs().
*
* Takes the segmenet lock.
*/
void
lfs_order_freelist(struct lfs *fs, ino_t **orphanp, size_t *norphanp)
{
CLEANERINFO *cip;
IFILE *ifp = NULL;
struct buf *bp;
ino_t ino, firstino, lastino, maxino;
ino_t *orphan = NULL;
size_t norphan = 0;
size_t norphan_alloc = 0;
ASSERT_NO_SEGLOCK(fs);
lfs_seglock(fs, SEGM_PROT);
/* largest inode on fs */
maxino = ((fs->lfs_ivnode->v_size >> lfs_sb_getbshift(fs)) -
lfs_sb_getcleansz(fs) - lfs_sb_getsegtabsz(fs)) * lfs_sb_getifpb(fs);
/* allocate the in-memory inode freemap */
/* XXX: assert that fs->lfs_ino_bitmap is null here */
fs->lfs_ino_bitmap =
malloc(((maxino + BMMASK) >> BMSHIFT) * sizeof(lfs_bm_t),
M_SEGMENT, M_WAITOK | M_ZERO);
KASSERT(fs->lfs_ino_bitmap != NULL);
/*
* Scan the ifile.
*/
firstino = lastino = LFS_UNUSED_INUM;
for (ino = 0; ino < maxino; ino++) {
/* Load this inode's ifile entry. */
if (ino % lfs_sb_getifpb(fs) == 0)
LFS_IENTRY(ifp, fs, ino, bp);
else
LFS_IENTRY_NEXT(ifp, fs);
/* Don't put zero or ifile on the free list */
if (ino == LFS_UNUSED_INUM || ino == LFS_IFILE_INUM)
continue;
/*
* Address orphaned files.
*
* The idea of this is to free inodes belonging to
* files that were unlinked but not reclaimed, I guess
* because if we're going to scan the whole ifile
* anyway it costs very little to do this. I don't
* immediately see any reason this should be disabled,
* but presumably it doesn't work... not sure what
* happens to such files currently. -- dholland 20160806
*/
if (lfs_if_getnextfree(fs, ifp) == LFS_ORPHAN_NEXTFREE(fs)) {
if (orphan == NULL) {
norphan_alloc = 32; /* XXX pulled from arse */
orphan = kmem_zalloc(sizeof(orphan[0]) *
norphan_alloc, KM_SLEEP);
} else if (norphan == norphan_alloc) {
ino_t *orphan_new;
if (norphan_alloc >= 4096)
norphan_alloc += 4096;
else
norphan_alloc *= 2;
orphan_new = kmem_zalloc(sizeof(orphan[0]) *
norphan_alloc, KM_SLEEP);
memcpy(orphan_new, orphan, sizeof(orphan[0]) *
norphan);
kmem_free(orphan, sizeof(orphan[0]) * norphan);
orphan = orphan_new;
}
orphan[norphan++] = ino;
}
if (lfs_if_getdaddr(fs, ifp) == LFS_UNUSED_DADDR) {
/*
* This inode is free. Put it on the free list.
*/
if (firstino == LFS_UNUSED_INUM) {
/* XXX: assert lastino == LFS_UNUSED_INUM? */
/* remember the first free inode */
firstino = ino;
} else {
/* release this inode's ifile entry */
brelse(bp, 0);
/* XXX: assert lastino != LFS_UNUSED_INUM? */
/* load lastino's ifile entry */
LFS_IENTRY(ifp, fs, lastino, bp);
/* set the list pointer */
lfs_if_setnextfree(fs, ifp, ino);
/* write the block */
LFS_BWRITE_LOG(bp);
/* reload this inode's ifile entry */
LFS_IENTRY(ifp, fs, ino, bp);
}
/* remember the last free inode seen so far */
lastino = ino;
/* Mark this inode free in the in-memory freemap */
SET_BITMAP_FREE(fs, ino);
}
/* If moving to the next ifile block, release the buffer. */
if ((ino + 1) % lfs_sb_getifpb(fs) == 0)
brelse(bp, 0);
}
/* Write the freelist head and tail pointers */
/* XXX: do we need to mark the superblock dirty? */
LFS_PUT_HEADFREE(fs, cip, bp, firstino);
LFS_PUT_TAILFREE(fs, cip, bp, lastino);
/* done */
lfs_segunlock(fs);
/*
* Shrink the array of orphans so we don't have to carry around
* the allocation size.
*/
if (norphan < norphan_alloc) {
ino_t *orphan_new = kmem_alloc(sizeof(orphan[0]) * norphan,
KM_SLEEP);
memcpy(orphan_new, orphan, sizeof(orphan[0]) * norphan);
kmem_free(orphan, sizeof(orphan[0]) * norphan_alloc);
orphan = orphan_new;
norphan_alloc = norphan;
}
*orphanp = orphan;
*norphanp = norphan;
}
/*
* Mark a file orphaned (unlinked but not yet reclaimed) by inode
* number. Do this with a magic freelist next pointer.
*
* XXX: howzabout some locking?
*/
void
lfs_orphan(struct lfs *fs, ino_t ino)
{
IFILE *ifp;
struct buf *bp;
LFS_IENTRY(ifp, fs, ino, bp);
lfs_if_setnextfree(fs, ifp, LFS_ORPHAN_NEXTFREE(fs));
LFS_BWRITE_LOG(bp);
}
/*
* Free orphans discovered during mount. This is a separate stage
* because it requires fs->lfs_suflags to be set up, which is not done
* by the time we run lfs_order_freelist. It's possible that we could
* run lfs_order_freelist later (i.e., set up fs->lfs_suflags sooner)
* but that requires more thought than I can put into this at the
* moment.
*/
void
lfs_free_orphans(struct lfs *fs, ino_t *orphan, size_t norphan)
{
size_t i;
for (i = 0; i < norphan; i++) {
ino_t ino = orphan[i];
unsigned segno;
struct vnode *vp;
struct inode *ip;
struct buf *bp;
IFILE *ifp;
SEGUSE *sup;
int error;
/* Get the segment the inode is in on disk. */
LFS_IENTRY(ifp, fs, ino, bp);
segno = lfs_dtosn(fs, lfs_if_getdaddr(fs, ifp));
brelse(bp, 0);
/*
* Try to get the vnode. If we can't, tough -- hope
* you have backups!
*/
error = VFS_VGET(fs->lfs_ivnode->v_mount, ino, &vp);
if (error) {
printf("orphan %jd vget error %d\n", (intmax_t)ino,
error);
continue;
}
/*
* Sanity-check the inode.
*
* XXX What to do if it is still referenced?
*/
ip = VTOI(vp);
if (ip->i_nlink != 0)
printf("orphan %jd nlink %d\n", (intmax_t)ino,
ip->i_nlink);
/*
* Truncate the inode, to free any blocks allocated for
* it, and release it, to free the inode number.
*
* XXX Isn't it redundant to truncate? Won't vput do
* that for us?
*/
error = lfs_truncate(vp, 0, 0, NOCRED);
if (error)
printf("orphan %jd truncate error %d", (intmax_t)ino,
error);
vput(vp);
/* Update the number of bytes in the segment summary. */
LFS_SEGENTRY(sup, fs, segno, bp);
KASSERT(sup->su_nbytes >= DINOSIZE(fs));
sup->su_nbytes -= DINOSIZE(fs);
LFS_WRITESEGENTRY(sup, fs, segno, bp);
/* Drop the on-disk address. */
LFS_IENTRY(ifp, fs, ino, bp);
lfs_if_setdaddr(fs, ifp, LFS_UNUSED_DADDR);
LFS_BWRITE_LOG(bp);
}
if (orphan)
kmem_free(orphan, sizeof(orphan[0]) * norphan);
}