/* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright 2007 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ /* Copyright (c) 1983, 1984, 1985, 1986, 1987, 1988, 1989 AT&T */ /* All Rights Reserved */ /* * Portions of this source code were derived from Berkeley 4.3 BSD * under license from the Regents of the University of California. */ #pragma ident "%Z%%M% %I% %E% SMI" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* must be AFTER ! */ #include #include #include #include #include /* _FIOIO */ #include #include #include #include #include #include #include #include #include #include #include static struct instats ins; static int ufs_getpage_ra(struct vnode *, u_offset_t, struct seg *, caddr_t); static int ufs_getpage_miss(struct vnode *, u_offset_t, size_t, struct seg *, caddr_t, struct page **, size_t, enum seg_rw, int); static int ufs_open(struct vnode **, int, struct cred *); static int ufs_close(struct vnode *, int, int, offset_t, struct cred *); static int ufs_read(struct vnode *, struct uio *, int, struct cred *, struct caller_context *); static int ufs_write(struct vnode *, struct uio *, int, struct cred *, struct caller_context *); static int ufs_ioctl(struct vnode *, int, intptr_t, int, struct cred *, int *); static int ufs_getattr(struct vnode *, struct vattr *, int, struct cred *); static int ufs_setattr(struct vnode *, struct vattr *, int, struct cred *, caller_context_t *); static int ufs_access(struct vnode *, int, int, struct cred *); static int ufs_lookup(struct vnode *, char *, struct vnode **, struct pathname *, int, struct vnode *, struct cred *); static int ufs_create(struct vnode *, char *, struct vattr *, enum vcexcl, int, struct vnode **, struct cred *, int); static int ufs_remove(struct vnode *, char *, struct cred *); static int ufs_link(struct vnode *, struct vnode *, char *, struct cred *); static int ufs_rename(struct vnode *, char *, struct vnode *, char *, struct cred *); static int ufs_mkdir(struct vnode *, char *, struct vattr *, struct vnode **, struct cred *); static int ufs_rmdir(struct vnode *, char *, struct vnode *, struct cred *); static int ufs_readdir(struct vnode *, struct uio *, struct cred *, int *); static int ufs_symlink(struct vnode *, char *, struct vattr *, char *, struct cred *); static int ufs_readlink(struct vnode *, struct uio *, struct cred *); static int ufs_fsync(struct vnode *, int, struct cred *); static void ufs_inactive(struct vnode *, struct cred *); static int ufs_fid(struct vnode *, struct fid *); static int ufs_rwlock(struct vnode *, int, caller_context_t *); static void ufs_rwunlock(struct vnode *, int, caller_context_t *); static int ufs_seek(struct vnode *, offset_t, offset_t *); static int ufs_frlock(struct vnode *, int, struct flock64 *, int, offset_t, struct flk_callback *, struct cred *); static int ufs_space(struct vnode *, int, struct flock64 *, int, offset_t, cred_t *, caller_context_t *); static int ufs_getpage(struct vnode *, offset_t, size_t, uint_t *, struct page **, size_t, struct seg *, caddr_t, enum seg_rw, struct cred *); static int ufs_putpage(struct vnode *, offset_t, size_t, int, struct cred *); static int ufs_putpages(struct vnode *, offset_t, size_t, int, struct cred *); static int ufs_map(struct vnode *, offset_t, struct as *, caddr_t *, size_t, uchar_t, uchar_t, uint_t, struct cred *); static int ufs_addmap(struct vnode *, offset_t, struct as *, caddr_t, size_t, uchar_t, uchar_t, uint_t, struct cred *); static int ufs_delmap(struct vnode *, offset_t, struct as *, caddr_t, size_t, uint_t, uint_t, uint_t, struct cred *); static int ufs_poll(vnode_t *, short, int, short *, struct pollhead **); static int ufs_dump(vnode_t *, caddr_t, int, int); static int ufs_l_pathconf(struct vnode *, int, ulong_t *, struct cred *); static int ufs_pageio(struct vnode *, struct page *, u_offset_t, size_t, int, struct cred *); static int ufs_dump(vnode_t *, caddr_t, int, int); static int ufs_dumpctl(vnode_t *, int, int *); static daddr32_t *save_dblks(struct inode *, struct ufsvfs *, daddr32_t *, daddr32_t *, int, int); static int ufs_getsecattr(struct vnode *, vsecattr_t *, int, struct cred *); static int ufs_setsecattr(struct vnode *, vsecattr_t *, int, struct cred *); extern int as_map_locked(struct as *, caddr_t, size_t, int ((*)()), void *); /* * For lockfs: ulockfs begin/end is now inlined in the ufs_xxx functions. * * XXX - ULOCKFS in fs_pathconf and ufs_ioctl is not inlined yet. */ struct vnodeops *ufs_vnodeops; /* NOTE: "not blkd" below means that the operation isn't blocked by lockfs */ const fs_operation_def_t ufs_vnodeops_template[] = { VOPNAME_OPEN, { .vop_open = ufs_open }, /* not blkd */ VOPNAME_CLOSE, { .vop_close = ufs_close }, /* not blkd */ VOPNAME_READ, { .vop_read = ufs_read }, VOPNAME_WRITE, { .vop_write = ufs_write }, VOPNAME_IOCTL, { .vop_ioctl = ufs_ioctl }, VOPNAME_GETATTR, { .vop_getattr = ufs_getattr }, VOPNAME_SETATTR, { .vop_setattr = ufs_setattr }, VOPNAME_ACCESS, { .vop_access = ufs_access }, VOPNAME_LOOKUP, { .vop_lookup = ufs_lookup }, VOPNAME_CREATE, { .vop_create = ufs_create }, VOPNAME_REMOVE, { .vop_remove = ufs_remove }, VOPNAME_LINK, { .vop_link = ufs_link }, VOPNAME_RENAME, { .vop_rename = ufs_rename }, VOPNAME_MKDIR, { .vop_mkdir = ufs_mkdir }, VOPNAME_RMDIR, { .vop_rmdir = ufs_rmdir }, VOPNAME_READDIR, { .vop_readdir = ufs_readdir }, VOPNAME_SYMLINK, { .vop_symlink = ufs_symlink }, VOPNAME_READLINK, { .vop_readlink = ufs_readlink }, VOPNAME_FSYNC, { .vop_fsync = ufs_fsync }, VOPNAME_INACTIVE, { .vop_inactive = ufs_inactive }, /* not blkd */ VOPNAME_FID, { .vop_fid = ufs_fid }, VOPNAME_RWLOCK, { .vop_rwlock = ufs_rwlock }, /* not blkd */ VOPNAME_RWUNLOCK, { .vop_rwunlock = ufs_rwunlock }, /* not blkd */ VOPNAME_SEEK, { .vop_seek = ufs_seek }, VOPNAME_FRLOCK, { .vop_frlock = ufs_frlock }, VOPNAME_SPACE, { .vop_space = ufs_space }, VOPNAME_GETPAGE, { .vop_getpage = ufs_getpage }, VOPNAME_PUTPAGE, { .vop_putpage = ufs_putpage }, VOPNAME_MAP, { .vop_map = ufs_map }, VOPNAME_ADDMAP, { .vop_addmap = ufs_addmap }, /* not blkd */ VOPNAME_DELMAP, { .vop_delmap = ufs_delmap }, /* not blkd */ VOPNAME_POLL, { .vop_poll = ufs_poll }, /* not blkd */ VOPNAME_DUMP, { .vop_dump = ufs_dump }, VOPNAME_PATHCONF, { .vop_pathconf = ufs_l_pathconf }, VOPNAME_PAGEIO, { .vop_pageio = ufs_pageio }, VOPNAME_DUMPCTL, { .vop_dumpctl = ufs_dumpctl }, VOPNAME_GETSECATTR, { .vop_getsecattr = ufs_getsecattr }, VOPNAME_SETSECATTR, { .vop_setsecattr = ufs_setsecattr }, VOPNAME_VNEVENT, { .vop_vnevent = fs_vnevent_support }, NULL, NULL }; #define MAX_BACKFILE_COUNT 9999 /* * Created by ufs_dumpctl() to store a file's disk block info into memory. * Used by ufs_dump() to dump data to disk directly. */ struct dump { struct inode *ip; /* the file we contain */ daddr_t fsbs; /* number of blocks stored */ struct timeval32 time; /* time stamp for the struct */ daddr32_t dblk[1]; /* place holder for block info */ }; static struct dump *dump_info = NULL; /* * Previously there was no special action required for ordinary files. * (Devices are handled through the device file system.) * Now we support Large Files and Large File API requires open to * fail if file is large. * We could take care to prevent data corruption * by doing an atomic check of size and truncate if file is opened with * FTRUNC flag set but traditionally this is being done by the vfs/vnode * layers. So taking care of truncation here is a change in the existing * semantics of VOP_OPEN and therefore we chose not to implement any thing * here. The check for the size of the file > 2GB is being done at the * vfs layer in routine vn_open(). */ /* ARGSUSED */ static int ufs_open(struct vnode **vpp, int flag, struct cred *cr) { return (0); } /*ARGSUSED*/ static int ufs_close(struct vnode *vp, int flag, int count, offset_t offset, struct cred *cr) { cleanlocks(vp, ttoproc(curthread)->p_pid, 0); cleanshares(vp, ttoproc(curthread)->p_pid); /* * Push partially filled cluster at last close. * ``last close'' is approximated because the dnlc * may have a hold on the vnode. * Checking for VBAD here will also act as a forced umount check. */ if (vp->v_count <= 2 && vp->v_type != VBAD) { struct inode *ip = VTOI(vp); if (ip->i_delaylen) { ins.in_poc.value.ul++; (void) ufs_putpages(vp, ip->i_delayoff, ip->i_delaylen, B_ASYNC | B_FREE, cr); ip->i_delaylen = 0; } } return (0); } /*ARGSUSED*/ static int ufs_read(struct vnode *vp, struct uio *uiop, int ioflag, struct cred *cr, struct caller_context *ct) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp; struct ulockfs *ulp = NULL; int error = 0; int intrans = 0; ASSERT(RW_READ_HELD(&ip->i_rwlock)); /* * Mandatory locking needs to be done before ufs_lockfs_begin() * and TRANS_BEGIN_SYNC() calls since mandatory locks can sleep. */ if (MANDLOCK(vp, ip->i_mode)) { /* * ufs_getattr ends up being called by chklock */ error = chklock(vp, FREAD, uiop->uio_loffset, uiop->uio_resid, uiop->uio_fmode, ct); if (error) goto out; } ufsvfsp = ip->i_ufsvfs; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_READ_MASK); if (error) goto out; /* * In the case that a directory is opened for reading as a file * (eg "cat .") with the O_RSYNC, O_SYNC and O_DSYNC flags set. * The locking order had to be changed to avoid a deadlock with * an update taking place on that directory at the same time. */ if ((ip->i_mode & IFMT) == IFDIR) { rw_enter(&ip->i_contents, RW_READER); error = rdip(ip, uiop, ioflag, cr); rw_exit(&ip->i_contents); if (error) { if (ulp) ufs_lockfs_end(ulp); goto out; } if (ulp && (ioflag & FRSYNC) && (ioflag & (FSYNC | FDSYNC)) && TRANS_ISTRANS(ufsvfsp)) { rw_exit(&ip->i_rwlock); TRANS_BEGIN_SYNC(ufsvfsp, TOP_READ_SYNC, TOP_READ_SIZE, error); ASSERT(!error); TRANS_END_SYNC(ufsvfsp, error, TOP_READ_SYNC, TOP_READ_SIZE); rw_enter(&ip->i_rwlock, RW_READER); } } else { /* * Only transact reads to files opened for sync-read and * sync-write on a file system that is not write locked. * * The ``not write locked'' check prevents problems with * enabling/disabling logging on a busy file system. E.g., * logging exists at the beginning of the read but does not * at the end. * */ if (ulp && (ioflag & FRSYNC) && (ioflag & (FSYNC | FDSYNC)) && TRANS_ISTRANS(ufsvfsp)) { TRANS_BEGIN_SYNC(ufsvfsp, TOP_READ_SYNC, TOP_READ_SIZE, error); ASSERT(!error); intrans = 1; } rw_enter(&ip->i_contents, RW_READER); error = rdip(ip, uiop, ioflag, cr); rw_exit(&ip->i_contents); if (intrans) { TRANS_END_SYNC(ufsvfsp, error, TOP_READ_SYNC, TOP_READ_SIZE); } } if (ulp) { ufs_lockfs_end(ulp); } out: return (error); } extern int ufs_HW; /* high water mark */ extern int ufs_LW; /* low water mark */ int ufs_WRITES = 1; /* XXX - enable/disable */ int ufs_throttles = 0; /* throttling count */ int ufs_allow_shared_writes = 1; /* directio shared writes */ static int ufs_check_rewrite(struct inode *ip, struct uio *uiop, int ioflag) { int shared_write; /* * If the FDSYNC flag is set then ignore the global * ufs_allow_shared_writes in this case. */ shared_write = (ioflag & FDSYNC) | ufs_allow_shared_writes; /* * Filter to determine if this request is suitable as a * concurrent rewrite. This write must not allocate blocks * by extending the file or filling in holes. No use trying * through FSYNC descriptors as the inode will be synchronously * updated after the write. The uio structure has not yet been * checked for sanity, so assume nothing. */ return (((ip->i_mode & IFMT) == IFREG) && !(ioflag & FAPPEND) && (uiop->uio_loffset >= (offset_t)0) && (uiop->uio_loffset < ip->i_size) && (uiop->uio_resid > 0) && ((ip->i_size - uiop->uio_loffset) >= uiop->uio_resid) && !(ioflag & FSYNC) && !bmap_has_holes(ip) && shared_write); } /*ARGSUSED*/ static int ufs_write(struct vnode *vp, struct uio *uiop, int ioflag, cred_t *cr, caller_context_t *ct) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp; struct ulockfs *ulp; int retry = 1; int error, resv, resid = 0; int directio_status; int exclusive; int rewriteflg; long start_resid = uiop->uio_resid; ASSERT(RW_LOCK_HELD(&ip->i_rwlock)); retry_mandlock: /* * Mandatory locking needs to be done before ufs_lockfs_begin() * and TRANS_BEGIN_[A]SYNC() calls since mandatory locks can sleep. * Check for forced unmounts normally done in ufs_lockfs_begin(). */ if ((ufsvfsp = ip->i_ufsvfs) == NULL) { error = EIO; goto out; } if (MANDLOCK(vp, ip->i_mode)) { ASSERT(RW_WRITE_HELD(&ip->i_rwlock)); /* * ufs_getattr ends up being called by chklock */ error = chklock(vp, FWRITE, uiop->uio_loffset, uiop->uio_resid, uiop->uio_fmode, ct); if (error) goto out; } /* i_rwlock can change in chklock */ exclusive = rw_write_held(&ip->i_rwlock); rewriteflg = ufs_check_rewrite(ip, uiop, ioflag); /* * Check for fast-path special case of directio re-writes. */ if ((ip->i_flag & IDIRECTIO || ufsvfsp->vfs_forcedirectio) && !exclusive && rewriteflg) { error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_WRITE_MASK); if (error) goto out; rw_enter(&ip->i_contents, RW_READER); error = ufs_directio_write(ip, uiop, ioflag, 1, cr, &directio_status); if (directio_status == DIRECTIO_SUCCESS) { uint_t i_flag_save; if (start_resid != uiop->uio_resid) error = 0; /* * Special treatment of access times for re-writes. * If IMOD is not already set, then convert it * to IMODACC for this operation. This defers * entering a delta into the log until the inode * is flushed. This mimics what is done for read * operations and inode access time. */ mutex_enter(&ip->i_tlock); i_flag_save = ip->i_flag; ip->i_flag |= IUPD | ICHG; ip->i_seq++; ITIMES_NOLOCK(ip); if ((i_flag_save & IMOD) == 0) { ip->i_flag &= ~IMOD; ip->i_flag |= IMODACC; } mutex_exit(&ip->i_tlock); rw_exit(&ip->i_contents); if (ulp) ufs_lockfs_end(ulp); goto out; } rw_exit(&ip->i_contents); if (ulp) ufs_lockfs_end(ulp); } if (!exclusive && !rw_tryupgrade(&ip->i_rwlock)) { rw_exit(&ip->i_rwlock); rw_enter(&ip->i_rwlock, RW_WRITER); /* * Mandatory locking could have been enabled * after dropping the i_rwlock. */ if (MANDLOCK(vp, ip->i_mode)) goto retry_mandlock; } error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_WRITE_MASK); if (error) goto out; /* * Amount of log space needed for this write */ if (!rewriteflg || !(ioflag & FDSYNC)) TRANS_WRITE_RESV(ip, uiop, ulp, &resv, &resid); /* * Throttle writes. */ if (ufs_WRITES && (ip->i_writes > ufs_HW)) { mutex_enter(&ip->i_tlock); while (ip->i_writes > ufs_HW) { ufs_throttles++; cv_wait(&ip->i_wrcv, &ip->i_tlock); } mutex_exit(&ip->i_tlock); } /* * Enter Transaction * * If the write is a rewrite there is no need to open a transaction * if the FDSYNC flag is set and not the FSYNC. In this case just * set the IMODACC flag to modify do the update at a later time * thus avoiding the overhead of the logging transaction that is * not required. */ if (ioflag & (FSYNC|FDSYNC)) { if (ulp) { if (rewriteflg) { uint_t i_flag_save; rw_enter(&ip->i_contents, RW_READER); mutex_enter(&ip->i_tlock); i_flag_save = ip->i_flag; ip->i_flag |= IUPD | ICHG; ip->i_seq++; ITIMES_NOLOCK(ip); if ((i_flag_save & IMOD) == 0) { ip->i_flag &= ~IMOD; ip->i_flag |= IMODACC; } mutex_exit(&ip->i_tlock); rw_exit(&ip->i_contents); } else { int terr = 0; TRANS_BEGIN_SYNC(ufsvfsp, TOP_WRITE_SYNC, resv, terr); ASSERT(!terr); } } } else { if (ulp) TRANS_BEGIN_ASYNC(ufsvfsp, TOP_WRITE, resv); } /* * Write the file */ rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&ip->i_contents, RW_WRITER); if ((ioflag & FAPPEND) != 0 && (ip->i_mode & IFMT) == IFREG) { /* * In append mode start at end of file. */ uiop->uio_loffset = ip->i_size; } /* * Mild optimisation, don't call ufs_trans_write() unless we have to * Also, suppress file system full messages if we will retry. */ if (retry) ip->i_flag |= IQUIET; if (resid) { TRANS_WRITE(ip, uiop, ioflag, error, ulp, cr, resv, resid); } else { error = wrip(ip, uiop, ioflag, cr); } ip->i_flag &= ~IQUIET; rw_exit(&ip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); /* * Leave Transaction */ if (ulp) { if (ioflag & (FSYNC|FDSYNC)) { if (!rewriteflg) { int terr = 0; TRANS_END_SYNC(ufsvfsp, terr, TOP_WRITE_SYNC, resv); if (error == 0) error = terr; } } else { TRANS_END_ASYNC(ufsvfsp, TOP_WRITE, resv); } ufs_lockfs_end(ulp); } out: if ((error == ENOSPC) && retry && TRANS_ISTRANS(ufsvfsp)) { /* * Any blocks tied up in pending deletes? */ ufs_delete_drain_wait(ufsvfsp, 1); retry = 0; goto retry_mandlock; } if (error == ENOSPC && (start_resid != uiop->uio_resid)) error = 0; return (error); } /* * Don't cache write blocks to files with the sticky bit set. * Used to keep swap files from blowing the page cache on a server. */ int stickyhack = 1; /* * Free behind hacks. The pager is busted. * XXX - need to pass the information down to writedone() in a flag like B_SEQ * or B_FREE_IF_TIGHT_ON_MEMORY. */ int freebehind = 1; int smallfile = 0; u_offset_t smallfile64 = 32 * 1024; /* * While we should, in most cases, cache the pages for write, we * may also want to cache the pages for read as long as they are * frequently re-usable. * * If cache_read_ahead = 1, the pages for read will go to the tail * of the cache list when they are released, otherwise go to the head. */ int cache_read_ahead = 0; /* * Freebehind exists so that as we read large files sequentially we * don't consume most of memory with pages from a few files. It takes * longer to re-read from disk multiple small files as it does reading * one large one sequentially. As system memory grows customers need * to retain bigger chunks of files in memory. The advent of the * cachelist opens up of the possibility freeing pages to the head or * tail of the list. * * Not freeing a page is a bet that the page will be read again before * it's segmap slot is needed for something else. If we loose the bet, * it means some other thread is burdened with the page free we did * not do. If we win we save a free and reclaim. * * Freeing it at the tail vs the head of cachelist is a bet that the * page will survive until the next read. It's also saying that this * page is more likely to be re-used than a page freed some time ago * and never reclaimed. * * Freebehind maintains a range of file offset [smallfile1; smallfile2] * * 0 < offset < smallfile1 : pages are not freed. * smallfile1 < offset < smallfile2 : pages freed to tail of cachelist. * smallfile2 < offset : pages freed to head of cachelist. * * The range is computed at most once per second and depends on * freemem and ncpus_online. Both parameters are bounded to be * >= smallfile && >= smallfile64. * * smallfile1 = (free memory / ncpu) / 1000 * smallfile2 = (free memory / ncpu) / 10 * * A few examples values: * * Free Mem (in Bytes) [smallfile1; smallfile2] [smallfile1; smallfile2] * ncpus_online = 4 ncpus_online = 64 * ------------------ ----------------------- ----------------------- * 1G [256K; 25M] [32K; 1.5M] * 10G [2.5M; 250M] [156K; 15M] * 100G [25M; 2.5G] [1.5M; 150M] * */ #define SMALLFILE1_D 1000 #define SMALLFILE2_D 10 static u_offset_t smallfile1 = 32 * 1024; static u_offset_t smallfile2 = 32 * 1024; static clock_t smallfile_update = 0; /* lbolt value of when to recompute */ uint_t smallfile1_d = SMALLFILE1_D; uint_t smallfile2_d = SMALLFILE2_D; /* * wrip does the real work of write requests for ufs. */ int wrip(struct inode *ip, struct uio *uio, int ioflag, struct cred *cr) { rlim64_t limit = uio->uio_llimit; u_offset_t off; u_offset_t old_i_size; struct fs *fs; struct vnode *vp; struct ufsvfs *ufsvfsp; caddr_t base; long start_resid = uio->uio_resid; /* save starting resid */ long premove_resid; /* resid before uiomove() */ uint_t flags; int newpage; int iupdat_flag, directio_status; int n, on, mapon; int error, pagecreate; int do_dqrwlock; /* drop/reacquire vfs_dqrwlock */ int32_t iblocks; int new_iblocks; /* * ip->i_size is incremented before the uiomove * is done on a write. If the move fails (bad user * address) reset ip->i_size. * The better way would be to increment ip->i_size * only if the uiomove succeeds. */ int i_size_changed = 0; o_mode_t type; int i_seq_needed = 0; vp = ITOV(ip); /* * check for forced unmount - should not happen as * the request passed the lockfs checks. */ if ((ufsvfsp = ip->i_ufsvfs) == NULL) return (EIO); fs = ip->i_fs; ASSERT(RW_WRITE_HELD(&ip->i_contents)); /* check for valid filetype */ type = ip->i_mode & IFMT; if ((type != IFREG) && (type != IFDIR) && (type != IFATTRDIR) && (type != IFLNK) && (type != IFSHAD)) { return (EIO); } /* * the actual limit of UFS file size * is UFS_MAXOFFSET_T */ if (limit == RLIM64_INFINITY || limit > MAXOFFSET_T) limit = MAXOFFSET_T; if (uio->uio_loffset >= limit) { proc_t *p = ttoproc(curthread); mutex_enter(&p->p_lock); (void) rctl_action(rctlproc_legacy[RLIMIT_FSIZE], p->p_rctls, p, RCA_UNSAFE_SIGINFO); mutex_exit(&p->p_lock); return (EFBIG); } /* * if largefiles are disallowed, the limit is * the pre-largefiles value of 2GB */ if (ufsvfsp->vfs_lfflags & UFS_LARGEFILES) limit = MIN(UFS_MAXOFFSET_T, limit); else limit = MIN(MAXOFF32_T, limit); if (uio->uio_loffset < (offset_t)0) { return (EINVAL); } if (uio->uio_resid == 0) { return (0); } if (uio->uio_loffset >= limit) return (EFBIG); ip->i_flag |= INOACC; /* don't update ref time in getpage */ if (ioflag & (FSYNC|FDSYNC)) { ip->i_flag |= ISYNC; iupdat_flag = 1; } /* * Try to go direct */ if (ip->i_flag & IDIRECTIO || ufsvfsp->vfs_forcedirectio) { uio->uio_llimit = limit; error = ufs_directio_write(ip, uio, ioflag, 0, cr, &directio_status); /* * If ufs_directio wrote to the file or set the flags, * we need to update i_seq, but it may be deferred. */ if (start_resid != uio->uio_resid || (ip->i_flag & (ICHG|IUPD))) { i_seq_needed = 1; ip->i_flag |= ISEQ; } if (directio_status == DIRECTIO_SUCCESS) goto out; } /* * Behavior with respect to dropping/reacquiring vfs_dqrwlock: * * o shadow inodes: vfs_dqrwlock is not held at all * o quota updates: vfs_dqrwlock is read or write held * o other updates: vfs_dqrwlock is read held * * The first case is the only one where we do not hold * vfs_dqrwlock at all while entering wrip(). * We must make sure not to downgrade/drop vfs_dqrwlock if we * have it as writer, i.e. if we are updating the quota inode. * There is no potential deadlock scenario in this case as * ufs_getpage() takes care of this and avoids reacquiring * vfs_dqrwlock in that case. * * This check is done here since the above conditions do not change * and we possibly loop below, so save a few cycles. */ if ((type == IFSHAD) || (rw_owner(&ufsvfsp->vfs_dqrwlock) == curthread)) { do_dqrwlock = 0; } else { do_dqrwlock = 1; } /* * Large Files: We cast MAXBMASK to offset_t * inorder to mask out the higher bits. Since offset_t * is a signed value, the high order bit set in MAXBMASK * value makes it do the right thing by having all bits 1 * in the higher word. May be removed for _SOLARIS64_. */ fs = ip->i_fs; do { u_offset_t uoff = uio->uio_loffset; off = uoff & (offset_t)MAXBMASK; mapon = (int)(uoff & (offset_t)MAXBOFFSET); on = (int)blkoff(fs, uoff); n = (int)MIN(fs->fs_bsize - on, uio->uio_resid); new_iblocks = 1; if (type == IFREG && uoff + n >= limit) { if (uoff >= limit) { error = EFBIG; goto out; } /* * since uoff + n >= limit, * therefore n >= limit - uoff, and n is an int * so it is safe to cast it to an int */ n = (int)(limit - (rlim64_t)uoff); } if (uoff + n > ip->i_size) { /* * We are extending the length of the file. * bmap is used so that we are sure that * if we need to allocate new blocks, that it * is done here before we up the file size. */ error = bmap_write(ip, uoff, (int)(on + n), mapon == 0, NULL, cr); /* * bmap_write never drops i_contents so if * the flags are set it changed the file. */ if (ip->i_flag & (ICHG|IUPD)) { i_seq_needed = 1; ip->i_flag |= ISEQ; } if (error) break; /* * There is a window of vulnerability here. * The sequence of operations: allocate file * system blocks, uiomove the data into pages, * and then update the size of the file in the * inode, must happen atomically. However, due * to current locking constraints, this can not * be done. */ ASSERT(ip->i_writer == NULL); ip->i_writer = curthread; i_size_changed = 1; /* * If we are writing from the beginning of * the mapping, we can just create the * pages without having to read them. */ pagecreate = (mapon == 0); } else if (n == MAXBSIZE) { /* * Going to do a whole mappings worth, * so we can just create the pages w/o * having to read them in. But before * we do that, we need to make sure any * needed blocks are allocated first. */ iblocks = ip->i_blocks; error = bmap_write(ip, uoff, (int)(on + n), BI_ALLOC_ONLY, NULL, cr); /* * bmap_write never drops i_contents so if * the flags are set it changed the file. */ if (ip->i_flag & (ICHG|IUPD)) { i_seq_needed = 1; ip->i_flag |= ISEQ; } if (error) break; pagecreate = 1; /* * check if the new created page needed the * allocation of new disk blocks. */ if (iblocks == ip->i_blocks) new_iblocks = 0; /* no new blocks allocated */ } else { pagecreate = 0; /* * In sync mode flush the indirect blocks which * may have been allocated and not written on * disk. In above cases bmap_write will allocate * in sync mode. */ if (ioflag & (FSYNC|FDSYNC)) { error = ufs_indirblk_sync(ip, uoff); if (error) break; } } /* * At this point we can enter ufs_getpage() in one * of two ways: * 1) segmap_getmapflt() calls ufs_getpage() when the * forcefault parameter is true (pagecreate == 0) * 2) uiomove() causes a page fault. * * We have to drop the contents lock to prevent the VM * system from trying to reaquire it in ufs_getpage() * should the uiomove cause a pagefault. * * We have to drop the reader vfs_dqrwlock here as well. */ rw_exit(&ip->i_contents); if (do_dqrwlock) { ASSERT(RW_LOCK_HELD(&ufsvfsp->vfs_dqrwlock)); ASSERT(!(RW_WRITE_HELD(&ufsvfsp->vfs_dqrwlock))); rw_exit(&ufsvfsp->vfs_dqrwlock); } newpage = 0; premove_resid = uio->uio_resid; if (vpm_enable) { /* * Copy data. If new pages are created, part of * the page that is not written will be initizliazed * with zeros. */ error = vpm_data_copy(vp, (off + mapon), (uint_t)n, uio, !pagecreate, &newpage, 0, S_WRITE); } else { base = segmap_getmapflt(segkmap, vp, (off + mapon), (uint_t)n, !pagecreate, S_WRITE); /* * segmap_pagecreate() returns 1 if it calls * page_create_va() to allocate any pages. */ if (pagecreate) newpage = segmap_pagecreate(segkmap, base, (size_t)n, 0); error = uiomove(base + mapon, (long)n, UIO_WRITE, uio); } /* * If "newpage" is set, then a new page was created and it * does not contain valid data, so it needs to be initialized * at this point. * Otherwise the page contains old data, which was overwritten * partially or as a whole in uiomove. * If there is only one iovec structure within uio, then * on error uiomove will not be able to update uio->uio_loffset * and we would zero the whole page here! * * If uiomove fails because of an error, the old valid data * is kept instead of filling the rest of the page with zero's. */ if (!vpm_enable && newpage && uio->uio_loffset < roundup(off + mapon + n, PAGESIZE)) { /* * We created pages w/o initializing them completely, * thus we need to zero the part that wasn't set up. * This happens on most EOF write cases and if * we had some sort of error during the uiomove. */ int nzero, nmoved; nmoved = (int)(uio->uio_loffset - (off + mapon)); ASSERT(nmoved >= 0 && nmoved <= n); nzero = roundup(on + n, PAGESIZE) - nmoved; ASSERT(nzero > 0 && mapon + nmoved + nzero <= MAXBSIZE); (void) kzero(base + mapon + nmoved, (uint_t)nzero); } /* * Unlock the pages allocated by page_create_va() * in segmap_pagecreate() */ if (!vpm_enable && newpage) segmap_pageunlock(segkmap, base, (size_t)n, S_WRITE); /* * If the size of the file changed, then update the * size field in the inode now. This can't be done * before the call to segmap_pageunlock or there is * a potential deadlock with callers to ufs_putpage(). * They will be holding i_contents and trying to lock * a page, while this thread is holding a page locked * and trying to acquire i_contents. */ if (i_size_changed) { rw_enter(&ip->i_contents, RW_WRITER); old_i_size = ip->i_size; UFS_SET_ISIZE(uoff + n, ip); TRANS_INODE(ufsvfsp, ip); /* * file has grown larger than 2GB. Set flag * in superblock to indicate this, if it * is not already set. */ if ((ip->i_size > MAXOFF32_T) && !(fs->fs_flags & FSLARGEFILES)) { ASSERT(ufsvfsp->vfs_lfflags & UFS_LARGEFILES); mutex_enter(&ufsvfsp->vfs_lock); fs->fs_flags |= FSLARGEFILES; ufs_sbwrite(ufsvfsp); mutex_exit(&ufsvfsp->vfs_lock); } mutex_enter(&ip->i_tlock); ip->i_writer = NULL; cv_broadcast(&ip->i_wrcv); mutex_exit(&ip->i_tlock); rw_exit(&ip->i_contents); } if (error) { /* * If we failed on a write, we may have already * allocated file blocks as well as pages. It's * hard to undo the block allocation, but we must * be sure to invalidate any pages that may have * been allocated. * * If the page was created without initialization * then we must check if it should be possible * to destroy the new page and to keep the old data * on the disk. * * It is possible to destroy the page without * having to write back its contents only when * - the size of the file keeps unchanged * - bmap_write() did not allocate new disk blocks * it is possible to create big files using "seek" and * write to the end of the file. A "write" to a * position before the end of the file would not * change the size of the file but it would allocate * new disk blocks. * - uiomove intended to overwrite the whole page. * - a new page was created (newpage == 1). */ if (i_size_changed == 0 && new_iblocks == 0 && newpage) { /* unwind what uiomove eventually last did */ uio->uio_resid = premove_resid; /* * destroy the page, do not write ambiguous * data to the disk. */ flags = SM_DESTROY; } else { /* * write the page back to the disk, if dirty, * and remove the page from the cache. */ flags = SM_INVAL; } if (vpm_enable) { /* * Flush pages. */ (void) vpm_sync_pages(vp, off, n, flags); } else { (void) segmap_release(segkmap, base, flags); } } else { flags = 0; /* * Force write back for synchronous write cases. */ if ((ioflag & (FSYNC|FDSYNC)) || type == IFDIR) { /* * If the sticky bit is set but the * execute bit is not set, we do a * synchronous write back and free * the page when done. We set up swap * files to be handled this way to * prevent servers from keeping around * the client's swap pages too long. * XXX - there ought to be a better way. */ if (IS_SWAPVP(vp)) { flags = SM_WRITE | SM_FREE | SM_DONTNEED; iupdat_flag = 0; } else { flags = SM_WRITE; } } else if (n + on == MAXBSIZE || IS_SWAPVP(vp)) { /* * Have written a whole block. * Start an asynchronous write and * mark the buffer to indicate that * it won't be needed again soon. */ flags = SM_WRITE | SM_ASYNC | SM_DONTNEED; } if (vpm_enable) { /* * Flush pages. */ error = vpm_sync_pages(vp, off, n, flags); } else { error = segmap_release(segkmap, base, flags); } /* * If the operation failed and is synchronous, * then we need to unwind what uiomove() last * did so we can potentially return an error to * the caller. If this write operation was * done in two pieces and the first succeeded, * then we won't return an error for the second * piece that failed. However, we only want to * return a resid value that reflects what was * really done. * * Failures for non-synchronous operations can * be ignored since the page subsystem will * retry the operation until it succeeds or the * file system is unmounted. */ if (error) { if ((ioflag & (FSYNC | FDSYNC)) || type == IFDIR) { uio->uio_resid = premove_resid; } else { error = 0; } } } /* * Re-acquire contents lock. * If it was dropped, reacquire reader vfs_dqrwlock as well. */ if (do_dqrwlock) rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&ip->i_contents, RW_WRITER); /* * If the uiomove() failed or if a synchronous * page push failed, fix up i_size. */ if (error) { if (i_size_changed) { /* * The uiomove failed, and we * allocated blocks,so get rid * of them. */ (void) ufs_itrunc(ip, old_i_size, 0, cr); } } else { /* * XXX - Can this be out of the loop? */ ip->i_flag |= IUPD | ICHG; /* * Only do one increase of i_seq for multiple * pieces. Because we drop locks, record * the fact that we changed the timestamp and * are deferring the increase in case another thread * pushes our timestamp update. */ i_seq_needed = 1; ip->i_flag |= ISEQ; if (i_size_changed) ip->i_flag |= IATTCHG; if ((ip->i_mode & (IEXEC | (IEXEC >> 3) | (IEXEC >> 6))) != 0 && (ip->i_mode & (ISUID | ISGID)) != 0 && secpolicy_vnode_setid_retain(cr, (ip->i_mode & ISUID) != 0 && ip->i_uid == 0) != 0) { /* * Clear Set-UID & Set-GID bits on * successful write if not privileged * and at least one of the execute bits * is set. If we always clear Set-GID, * mandatory file and record locking is * unuseable. */ ip->i_mode &= ~(ISUID | ISGID); } } /* * In the case the FDSYNC flag is set and this is a * "rewrite" we won't log a delta. * The FSYNC flag overrides all cases. */ if (!ufs_check_rewrite(ip, uio, ioflag) || !(ioflag & FDSYNC)) { TRANS_INODE(ufsvfsp, ip); } } while (error == 0 && uio->uio_resid > 0 && n != 0); out: /* * Make sure i_seq is increased at least once per write */ if (i_seq_needed) { ip->i_seq++; ip->i_flag &= ~ISEQ; /* no longer deferred */ } /* * Inode is updated according to this table - * * FSYNC FDSYNC(posix.4) * -------------------------- * always@ IATTCHG|IBDWRITE * * @ - If we are doing synchronous write the only time we should * not be sync'ing the ip here is if we have the stickyhack * activated, the file is marked with the sticky bit and * no exec bit, the file length has not been changed and * no new blocks have been allocated during this write. */ if ((ip->i_flag & ISYNC) != 0) { /* * we have eliminated nosync */ if ((ip->i_flag & (IATTCHG|IBDWRITE)) || ((ioflag & FSYNC) && iupdat_flag)) { ufs_iupdat(ip, 1); } } /* * If we've already done a partial-write, terminate * the write but return no error unless the error is ENOSPC * because the caller can detect this and free resources and * try again. */ if ((start_resid != uio->uio_resid) && (error != ENOSPC)) error = 0; ip->i_flag &= ~(INOACC | ISYNC); ITIMES_NOLOCK(ip); return (error); } /* * rdip does the real work of read requests for ufs. */ int rdip(struct inode *ip, struct uio *uio, int ioflag, cred_t *cr) { u_offset_t off; caddr_t base; struct fs *fs; struct ufsvfs *ufsvfsp; struct vnode *vp; long oresid = uio->uio_resid; u_offset_t n, on, mapon; int error = 0; int doupdate = 1; uint_t flags; int dofree, directio_status; krw_t rwtype; o_mode_t type; vp = ITOV(ip); ASSERT(RW_LOCK_HELD(&ip->i_contents)); ufsvfsp = ip->i_ufsvfs; if (ufsvfsp == NULL) return (EIO); fs = ufsvfsp->vfs_fs; /* check for valid filetype */ type = ip->i_mode & IFMT; if ((type != IFREG) && (type != IFDIR) && (type != IFATTRDIR) && (type != IFLNK) && (type != IFSHAD)) { return (EIO); } if (uio->uio_loffset > UFS_MAXOFFSET_T) { error = 0; goto out; } if (uio->uio_loffset < (offset_t)0) { return (EINVAL); } if (uio->uio_resid == 0) { return (0); } if (!ULOCKFS_IS_NOIACC(ITOUL(ip)) && (fs->fs_ronly == 0) && (!ufsvfsp->vfs_noatime)) { mutex_enter(&ip->i_tlock); ip->i_flag |= IACC; mutex_exit(&ip->i_tlock); } /* * Try to go direct */ if (ip->i_flag & IDIRECTIO || ufsvfsp->vfs_forcedirectio) { error = ufs_directio_read(ip, uio, cr, &directio_status); if (directio_status == DIRECTIO_SUCCESS) goto out; } rwtype = (rw_write_held(&ip->i_contents)?RW_WRITER:RW_READER); do { offset_t diff; u_offset_t uoff = uio->uio_loffset; off = uoff & (offset_t)MAXBMASK; mapon = (u_offset_t)(uoff & (offset_t)MAXBOFFSET); on = (u_offset_t)blkoff(fs, uoff); n = MIN((u_offset_t)fs->fs_bsize - on, (u_offset_t)uio->uio_resid); diff = ip->i_size - uoff; if (diff <= (offset_t)0) { error = 0; goto out; } if (diff < (offset_t)n) n = (int)diff; /* * We update smallfile2 and smallfile1 at most every second. */ if (lbolt >= smallfile_update) { uint64_t percpufreeb; if (smallfile1_d == 0) smallfile1_d = SMALLFILE1_D; if (smallfile2_d == 0) smallfile2_d = SMALLFILE2_D; percpufreeb = ptob((uint64_t)freemem) / ncpus_online; smallfile1 = percpufreeb / smallfile1_d; smallfile2 = percpufreeb / smallfile2_d; smallfile1 = MAX(smallfile1, smallfile); smallfile1 = MAX(smallfile1, smallfile64); smallfile2 = MAX(smallfile1, smallfile2); smallfile_update = lbolt + hz; } dofree = freebehind && ip->i_nextr == (off & PAGEMASK) && off > smallfile1; /* * At this point we can enter ufs_getpage() in one of two * ways: * 1) segmap_getmapflt() calls ufs_getpage() when the * forcefault parameter is true (value of 1 is passed) * 2) uiomove() causes a page fault. * * We cannot hold onto an i_contents reader lock without * risking deadlock in ufs_getpage() so drop a reader lock. * The ufs_getpage() dolock logic already allows for a * thread holding i_contents as writer to work properly * so we keep a writer lock. */ if (rwtype == RW_READER) rw_exit(&ip->i_contents); if (vpm_enable) { /* * Copy data. */ error = vpm_data_copy(vp, (off + mapon), (uint_t)n, uio, 1, NULL, 0, S_READ); } else { base = segmap_getmapflt(segkmap, vp, (off + mapon), (uint_t)n, 1, S_READ); error = uiomove(base + mapon, (long)n, UIO_READ, uio); } flags = 0; if (!error) { /* * If reading sequential we won't need this * buffer again soon. For offsets in range * [smallfile1, smallfile2] release the pages * at the tail of the cache list, larger * offsets are released at the head. */ if (dofree) { flags = SM_FREE | SM_ASYNC; if ((cache_read_ahead == 0) && (off > smallfile2)) flags |= SM_DONTNEED; } /* * In POSIX SYNC (FSYNC and FDSYNC) read mode, * we want to make sure that the page which has * been read, is written on disk if it is dirty. * And corresponding indirect blocks should also * be flushed out. */ if ((ioflag & FRSYNC) && (ioflag & (FSYNC|FDSYNC))) { flags &= ~SM_ASYNC; flags |= SM_WRITE; } if (vpm_enable) { error = vpm_sync_pages(vp, off, n, flags); } else { error = segmap_release(segkmap, base, flags); } } else { if (vpm_enable) { (void) vpm_sync_pages(vp, off, n, flags); } else { (void) segmap_release(segkmap, base, flags); } } if (rwtype == RW_READER) rw_enter(&ip->i_contents, rwtype); } while (error == 0 && uio->uio_resid > 0 && n != 0); out: /* * Inode is updated according to this table if FRSYNC is set. * * FSYNC FDSYNC(posix.4) * -------------------------- * always IATTCHG|IBDWRITE */ /* * The inode is not updated if we're logging and the inode is a * directory with FRSYNC, FSYNC and FDSYNC flags set. */ if (ioflag & FRSYNC) { if (TRANS_ISTRANS(ufsvfsp) && ((ip->i_mode & IFMT) == IFDIR)) { doupdate = 0; } if (doupdate) { if ((ioflag & FSYNC) || ((ioflag & FDSYNC) && (ip->i_flag & (IATTCHG|IBDWRITE)))) { ufs_iupdat(ip, 1); } } } /* * If we've already done a partial read, terminate * the read but return no error. */ if (oresid != uio->uio_resid) error = 0; ITIMES(ip); return (error); } /* ARGSUSED */ static int ufs_ioctl( struct vnode *vp, int cmd, intptr_t arg, int flag, struct cred *cr, int *rvalp) { struct lockfs lockfs, lockfs_out; struct ufsvfs *ufsvfsp = VTOI(vp)->i_ufsvfs; char *comment, *original_comment; struct fs *fs; struct ulockfs *ulp; offset_t off; extern int maxphys; int error; int issync; int trans_size; /* * forcibly unmounted */ if (ufsvfsp == NULL) { return (EIO); } fs = ufsvfsp->vfs_fs; if (cmd == Q_QUOTACTL) { error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_QUOTA_MASK); if (error) return (error); if (ulp) { TRANS_BEGIN_ASYNC(ufsvfsp, TOP_QUOTA, TOP_SETQUOTA_SIZE(fs)); } error = quotactl(vp, arg, flag, cr); if (ulp) { TRANS_END_ASYNC(ufsvfsp, TOP_QUOTA, TOP_SETQUOTA_SIZE(fs)); ufs_lockfs_end(ulp); } return (error); } switch (cmd) { case _FIOLFS: /* * file system locking */ if (secpolicy_fs_config(cr, ufsvfsp->vfs_vfs) != 0) return (EPERM); if ((flag & DATAMODEL_MASK) == DATAMODEL_NATIVE) { if (copyin((caddr_t)arg, &lockfs, sizeof (struct lockfs))) return (EFAULT); } #ifdef _SYSCALL32_IMPL else { struct lockfs32 lockfs32; /* Translate ILP32 lockfs to LP64 lockfs */ if (copyin((caddr_t)arg, &lockfs32, sizeof (struct lockfs32))) return (EFAULT); lockfs.lf_lock = (ulong_t)lockfs32.lf_lock; lockfs.lf_flags = (ulong_t)lockfs32.lf_flags; lockfs.lf_key = (ulong_t)lockfs32.lf_key; lockfs.lf_comlen = (ulong_t)lockfs32.lf_comlen; lockfs.lf_comment = (caddr_t)(uintptr_t)lockfs32.lf_comment; } #endif /* _SYSCALL32_IMPL */ if (lockfs.lf_comlen) { if (lockfs.lf_comlen > LOCKFS_MAXCOMMENTLEN) return (ENAMETOOLONG); comment = kmem_alloc(lockfs.lf_comlen, KM_SLEEP); if (copyin(lockfs.lf_comment, comment, lockfs.lf_comlen)) { kmem_free(comment, lockfs.lf_comlen); return (EFAULT); } original_comment = lockfs.lf_comment; lockfs.lf_comment = comment; } if ((error = ufs_fiolfs(vp, &lockfs, 0)) == 0) { lockfs.lf_comment = original_comment; if ((flag & DATAMODEL_MASK) == DATAMODEL_NATIVE) { (void) copyout(&lockfs, (caddr_t)arg, sizeof (struct lockfs)); } #ifdef _SYSCALL32_IMPL else { struct lockfs32 lockfs32; /* Translate LP64 to ILP32 lockfs */ lockfs32.lf_lock = (uint32_t)lockfs.lf_lock; lockfs32.lf_flags = (uint32_t)lockfs.lf_flags; lockfs32.lf_key = (uint32_t)lockfs.lf_key; lockfs32.lf_comlen = (uint32_t)lockfs.lf_comlen; lockfs32.lf_comment = (uint32_t)(uintptr_t) lockfs.lf_comment; (void) copyout(&lockfs32, (caddr_t)arg, sizeof (struct lockfs32)); } #endif /* _SYSCALL32_IMPL */ } else { if (lockfs.lf_comlen) kmem_free(comment, lockfs.lf_comlen); } return (error); case _FIOLFSS: /* * get file system locking status */ if ((flag & DATAMODEL_MASK) == DATAMODEL_NATIVE) { if (copyin((caddr_t)arg, &lockfs, sizeof (struct lockfs))) return (EFAULT); } #ifdef _SYSCALL32_IMPL else { struct lockfs32 lockfs32; /* Translate ILP32 lockfs to LP64 lockfs */ if (copyin((caddr_t)arg, &lockfs32, sizeof (struct lockfs32))) return (EFAULT); lockfs.lf_lock = (ulong_t)lockfs32.lf_lock; lockfs.lf_flags = (ulong_t)lockfs32.lf_flags; lockfs.lf_key = (ulong_t)lockfs32.lf_key; lockfs.lf_comlen = (ulong_t)lockfs32.lf_comlen; lockfs.lf_comment = (caddr_t)(uintptr_t)lockfs32.lf_comment; } #endif /* _SYSCALL32_IMPL */ if (error = ufs_fiolfss(vp, &lockfs_out)) return (error); lockfs.lf_lock = lockfs_out.lf_lock; lockfs.lf_key = lockfs_out.lf_key; lockfs.lf_flags = lockfs_out.lf_flags; lockfs.lf_comlen = MIN(lockfs.lf_comlen, lockfs_out.lf_comlen); if ((flag & DATAMODEL_MASK) == DATAMODEL_NATIVE) { if (copyout(&lockfs, (caddr_t)arg, sizeof (struct lockfs))) return (EFAULT); } #ifdef _SYSCALL32_IMPL else { /* Translate LP64 to ILP32 lockfs */ struct lockfs32 lockfs32; lockfs32.lf_lock = (uint32_t)lockfs.lf_lock; lockfs32.lf_flags = (uint32_t)lockfs.lf_flags; lockfs32.lf_key = (uint32_t)lockfs.lf_key; lockfs32.lf_comlen = (uint32_t)lockfs.lf_comlen; lockfs32.lf_comment = (uint32_t)(uintptr_t)lockfs.lf_comment; if (copyout(&lockfs32, (caddr_t)arg, sizeof (struct lockfs32))) return (EFAULT); } #endif /* _SYSCALL32_IMPL */ if (lockfs.lf_comlen && lockfs.lf_comment && lockfs_out.lf_comment) if (copyout(lockfs_out.lf_comment, lockfs.lf_comment, lockfs.lf_comlen)) return (EFAULT); return (0); case _FIOSATIME: /* * set access time */ /* * if mounted w/o atime, return quietly. * I briefly thought about returning ENOSYS, but * figured that most apps would consider this fatal * but the idea is to make this as seamless as poss. */ if (ufsvfsp->vfs_noatime) return (0); error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_SETATTR_MASK); if (error) return (error); if (ulp) { trans_size = (int)TOP_SETATTR_SIZE(VTOI(vp)); TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_SETATTR, trans_size); } error = ufs_fiosatime(vp, (struct timeval *)arg, flag, cr); if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_SETATTR, trans_size); ufs_lockfs_end(ulp); } return (error); case _FIOSDIO: /* * set delayed-io */ return (ufs_fiosdio(vp, (uint_t *)arg, flag, cr)); case _FIOGDIO: /* * get delayed-io */ return (ufs_fiogdio(vp, (uint_t *)arg, flag, cr)); case _FIOIO: /* * inode open */ error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_VGET_MASK); if (error) return (error); error = ufs_fioio(vp, (struct fioio *)arg, flag, cr); if (ulp) { ufs_lockfs_end(ulp); } return (error); case _FIOFFS: /* * file system flush (push w/invalidate) */ if ((caddr_t)arg != NULL) return (EINVAL); return (ufs_fioffs(vp, NULL, cr)); case _FIOISBUSY: /* * Contract-private interface for Legato * Purge this vnode from the DNLC and decide * if this vnode is busy (*arg == 1) or not * (*arg == 0) */ if (secpolicy_fs_config(cr, ufsvfsp->vfs_vfs) != 0) return (EPERM); error = ufs_fioisbusy(vp, (int *)arg, cr); return (error); case _FIODIRECTIO: return (ufs_fiodirectio(vp, (int)arg, cr)); case _FIOTUNE: /* * Tune the file system (aka setting fs attributes) */ error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_SETATTR_MASK); if (error) return (error); error = ufs_fiotune(vp, (struct fiotune *)arg, cr); if (ulp) ufs_lockfs_end(ulp); return (error); case _FIOLOGENABLE: if (secpolicy_fs_config(cr, ufsvfsp->vfs_vfs) != 0) return (EPERM); return (ufs_fiologenable(vp, (void *)arg, cr, flag)); case _FIOLOGDISABLE: if (secpolicy_fs_config(cr, ufsvfsp->vfs_vfs) != 0) return (EPERM); return (ufs_fiologdisable(vp, (void *)arg, cr, flag)); case _FIOISLOG: return (ufs_fioislog(vp, (void *)arg, cr, flag)); case _FIOSNAPSHOTCREATE_MULTI: { struct fiosnapcreate_multi fc, *fcp; size_t fcm_size; if (copyin((void *)arg, &fc, sizeof (fc))) return (EFAULT); if (fc.backfilecount > MAX_BACKFILE_COUNT) return (EINVAL); fcm_size = sizeof (struct fiosnapcreate_multi) + (fc.backfilecount - 1) * sizeof (int); fcp = (struct fiosnapcreate_multi *) kmem_alloc(fcm_size, KM_SLEEP); if (copyin((void *)arg, fcp, fcm_size)) { kmem_free(fcp, fcm_size); return (EFAULT); } error = ufs_snap_create(vp, fcp, cr); if (!error && copyout(fcp, (void *)arg, fcm_size)) error = EFAULT; kmem_free(fcp, fcm_size); return (error); } case _FIOSNAPSHOTDELETE: { struct fiosnapdelete fc; if (copyin((void *)arg, &fc, sizeof (fc))) return (EFAULT); error = ufs_snap_delete(vp, &fc, cr); if (!error && copyout(&fc, (void *)arg, sizeof (fc))) error = EFAULT; return (error); } case _FIOGETSUPERBLOCK: if (copyout(fs, (void *)arg, SBSIZE)) return (EFAULT); return (0); case _FIOGETMAXPHYS: if (copyout(&maxphys, (void *)arg, sizeof (maxphys))) return (EFAULT); return (0); /* * The following 3 ioctls are for TSufs support * although could potentially be used elsewhere */ case _FIO_SET_LUFS_DEBUG: if (secpolicy_fs_config(cr, ufsvfsp->vfs_vfs) != 0) return (EPERM); lufs_debug = (uint32_t)arg; return (0); case _FIO_SET_LUFS_ERROR: if (secpolicy_fs_config(cr, ufsvfsp->vfs_vfs) != 0) return (EPERM); TRANS_SETERROR(ufsvfsp); return (0); case _FIO_GET_TOP_STATS: { fio_lufs_stats_t *ls; ml_unit_t *ul = ufsvfsp->vfs_log; ls = kmem_zalloc(sizeof (*ls), KM_SLEEP); ls->ls_debug = ul->un_debug; /* return debug value */ /* Copy stucture if statistics are being kept */ if (ul->un_logmap->mtm_tops) { ls->ls_topstats = *(ul->un_logmap->mtm_tops); } error = 0; if (copyout(ls, (void *)arg, sizeof (*ls))) error = EFAULT; kmem_free(ls, sizeof (*ls)); return (error); } case _FIO_SEEK_DATA: case _FIO_SEEK_HOLE: if (ddi_copyin((void *)arg, &off, sizeof (off), flag)) return (EFAULT); /* offset paramater is in/out */ error = ufs_fio_holey(vp, cmd, &off); if (error) return (error); if (ddi_copyout(&off, (void *)arg, sizeof (off), flag)) return (EFAULT); return (0); default: return (ENOTTY); } } /* ARGSUSED */ static int ufs_getattr(struct vnode *vp, struct vattr *vap, int flags, struct cred *cr) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp; int err; if (vap->va_mask == AT_SIZE) { /* * for performance, if only the size is requested don't bother * with anything else. */ UFS_GET_ISIZE(&vap->va_size, ip); return (0); } /* * inlined lockfs checks */ ufsvfsp = ip->i_ufsvfs; if ((ufsvfsp == NULL) || ULOCKFS_IS_HLOCK(&ufsvfsp->vfs_ulockfs)) { err = EIO; goto out; } rw_enter(&ip->i_contents, RW_READER); /* * Return all the attributes. This should be refined so * that it only returns what's asked for. */ /* * Copy from inode table. */ vap->va_type = vp->v_type; vap->va_mode = ip->i_mode & MODEMASK; /* * If there is an ACL and there is a mask entry, then do the * extra work that completes the equivalent of an acltomode(3) * call. According to POSIX P1003.1e, the acl mask should be * returned in the group permissions field. * * - start with the original permission and mode bits (from above) * - clear the group owner bits * - add in the mask bits. */ if (ip->i_ufs_acl && ip->i_ufs_acl->aclass.acl_ismask) { vap->va_mode &= ~((VREAD | VWRITE | VEXEC) >> 3); vap->va_mode |= (ip->i_ufs_acl->aclass.acl_maskbits & PERMMASK) << 3; } vap->va_uid = ip->i_uid; vap->va_gid = ip->i_gid; vap->va_fsid = ip->i_dev; vap->va_nodeid = (ino64_t)ip->i_number; vap->va_nlink = ip->i_nlink; vap->va_size = ip->i_size; if (vp->v_type == VCHR || vp->v_type == VBLK) vap->va_rdev = ip->i_rdev; else vap->va_rdev = 0; /* not a b/c spec. */ mutex_enter(&ip->i_tlock); ITIMES_NOLOCK(ip); /* mark correct time in inode */ vap->va_seq = ip->i_seq; vap->va_atime.tv_sec = (time_t)ip->i_atime.tv_sec; vap->va_atime.tv_nsec = ip->i_atime.tv_usec*1000; vap->va_mtime.tv_sec = (time_t)ip->i_mtime.tv_sec; vap->va_mtime.tv_nsec = ip->i_mtime.tv_usec*1000; vap->va_ctime.tv_sec = (time_t)ip->i_ctime.tv_sec; vap->va_ctime.tv_nsec = ip->i_ctime.tv_usec*1000; mutex_exit(&ip->i_tlock); switch (ip->i_mode & IFMT) { case IFBLK: vap->va_blksize = MAXBSIZE; /* was BLKDEV_IOSIZE */ break; case IFCHR: vap->va_blksize = MAXBSIZE; break; default: vap->va_blksize = ip->i_fs->fs_bsize; break; } vap->va_nblocks = (fsblkcnt64_t)ip->i_blocks; rw_exit(&ip->i_contents); err = 0; out: return (err); } /*ARGSUSED4*/ static int ufs_setattr( struct vnode *vp, struct vattr *vap, int flags, struct cred *cr, caller_context_t *ct) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp = ip->i_ufsvfs; struct fs *fs; struct ulockfs *ulp; char *errmsg1; char *errmsg2; long blocks; long int mask = vap->va_mask; size_t len1, len2; int issync; int trans_size; int dotrans; int dorwlock; int error; int owner_change; int dodqlock; timestruc_t now; vattr_t oldva; int retry = 1; int indeadlock; /* * Cannot set these attributes. */ if (mask & AT_NOSET) { error = EINVAL; goto out; } /* * check for forced unmount */ if (ufsvfsp == NULL) return (EIO); fs = ufsvfsp->vfs_fs; if (fs->fs_ronly != 0) return (EROFS); again: errmsg1 = NULL; errmsg2 = NULL; dotrans = 0; dorwlock = 0; dodqlock = 0; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_SETATTR_MASK); if (error) goto out; /* * Acquire i_rwlock before TRANS_BEGIN_CSYNC() if this is a file. * This follows the protocol for read()/write(). */ if (vp->v_type != VDIR) { /* * ufs_tryirwlock uses rw_tryenter and checks for SLOCK to * avoid i_rwlock, ufs_lockfs_begin deadlock. If deadlock * possible, retries the operation. */ ufs_tryirwlock(&ip->i_rwlock, RW_WRITER, retry_file); if (indeadlock) { if (ulp) ufs_lockfs_end(ulp); goto again; } dorwlock = 1; } /* * Truncate file. Must have write permission and not be a directory. */ if (mask & AT_SIZE) { rw_enter(&ip->i_contents, RW_WRITER); if (vp->v_type == VDIR) { error = EISDIR; goto update_inode; } if (error = ufs_iaccess(ip, IWRITE, cr)) goto update_inode; rw_exit(&ip->i_contents); error = TRANS_ITRUNC(ip, vap->va_size, 0, cr); if (error) { rw_enter(&ip->i_contents, RW_WRITER); goto update_inode; } } if (ulp) { trans_size = (int)TOP_SETATTR_SIZE(ip); TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_SETATTR, trans_size); ++dotrans; } /* * Acquire i_rwlock after TRANS_BEGIN_CSYNC() if this is a directory. * This follows the protocol established by * ufs_link/create/remove/rename/mkdir/rmdir/symlink. */ if (vp->v_type == VDIR) { ufs_tryirwlock_trans(&ip->i_rwlock, RW_WRITER, TOP_SETATTR, retry_dir); if (indeadlock) goto again; dorwlock = 1; } /* * Grab quota lock if we are changing the file's owner. */ if (mask & AT_UID) { rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); dodqlock = 1; } rw_enter(&ip->i_contents, RW_WRITER); oldva.va_mode = ip->i_mode; oldva.va_uid = ip->i_uid; oldva.va_gid = ip->i_gid; vap->va_mask &= ~AT_SIZE; /* * ufs_iaccess is "close enough"; that's because it doesn't * map the defines. */ error = secpolicy_vnode_setattr(cr, vp, vap, &oldva, flags, ufs_iaccess, ip); if (error) goto update_inode; mask = vap->va_mask; /* * Change file access modes. */ if (mask & AT_MODE) { ip->i_mode = (ip->i_mode & IFMT) | (vap->va_mode & ~IFMT); TRANS_INODE(ufsvfsp, ip); ip->i_flag |= ICHG; if (stickyhack) { mutex_enter(&vp->v_lock); if ((ip->i_mode & (ISVTX | IEXEC | IFDIR)) == ISVTX) vp->v_flag |= VSWAPLIKE; else vp->v_flag &= ~VSWAPLIKE; mutex_exit(&vp->v_lock); } } if (mask & (AT_UID|AT_GID)) { if (mask & AT_UID) { /* * Don't change ownership of the quota inode. */ if (ufsvfsp->vfs_qinod == ip) { ASSERT(ufsvfsp->vfs_qflags & MQ_ENABLED); error = EINVAL; goto update_inode; } /* * No real ownership change. */ if (ip->i_uid == vap->va_uid) { blocks = 0; owner_change = 0; } /* * Remove the blocks and the file, from the old user's * quota. */ else { blocks = ip->i_blocks; owner_change = 1; (void) chkdq(ip, -blocks, /* force */ 1, cr, (char **)NULL, (size_t *)NULL); (void) chkiq(ufsvfsp, /* change */ -1, ip, (uid_t)ip->i_uid, /* force */ 1, cr, (char **)NULL, (size_t *)NULL); dqrele(ip->i_dquot); } ip->i_uid = vap->va_uid; /* * There is a real ownership change. */ if (owner_change) { /* * Add the blocks and the file to the new * user's quota. */ ip->i_dquot = getinoquota(ip); (void) chkdq(ip, blocks, /* force */ 1, cr, &errmsg1, &len1); (void) chkiq(ufsvfsp, /* change */ 1, (struct inode *)NULL, (uid_t)ip->i_uid, /* force */ 1, cr, &errmsg2, &len2); } } if (mask & AT_GID) { ip->i_gid = vap->va_gid; } TRANS_INODE(ufsvfsp, ip); ip->i_flag |= ICHG; } /* * Change file access or modified times. */ if (mask & (AT_ATIME|AT_MTIME)) { /* Check that the time value is within ufs range */ if (((mask & AT_ATIME) && TIMESPEC_OVERFLOW(&vap->va_atime)) || ((mask & AT_MTIME) && TIMESPEC_OVERFLOW(&vap->va_mtime))) { error = EOVERFLOW; goto update_inode; } /* * if the "noaccess" mount option is set and only atime * update is requested, do nothing. No error is returned. */ if ((ufsvfsp->vfs_noatime) && ((mask & (AT_ATIME|AT_MTIME)) == AT_ATIME)) goto skip_atime; if (mask & AT_ATIME) { ip->i_atime.tv_sec = vap->va_atime.tv_sec; ip->i_atime.tv_usec = vap->va_atime.tv_nsec / 1000; ip->i_flag &= ~IACC; } if (mask & AT_MTIME) { ip->i_mtime.tv_sec = vap->va_mtime.tv_sec; ip->i_mtime.tv_usec = vap->va_mtime.tv_nsec / 1000; gethrestime(&now); if (now.tv_sec > TIME32_MAX) { /* * In 2038, ctime sticks forever.. */ ip->i_ctime.tv_sec = TIME32_MAX; ip->i_ctime.tv_usec = 0; } else { ip->i_ctime.tv_sec = now.tv_sec; ip->i_ctime.tv_usec = now.tv_nsec / 1000; } ip->i_flag &= ~(IUPD|ICHG); ip->i_flag |= IMODTIME; } TRANS_INODE(ufsvfsp, ip); ip->i_flag |= IMOD; } skip_atime: /* * The presence of a shadow inode may indicate an ACL, but does * not imply an ACL. Future FSD types should be handled here too * and check for the presence of the attribute-specific data * before referencing it. */ if (ip->i_shadow) { /* * XXX if ufs_iupdat is changed to sandbagged write fix * ufs_acl_setattr to push ip to keep acls consistent * * Suppress out of inodes messages if we will retry. */ if (retry) ip->i_flag |= IQUIET; error = ufs_acl_setattr(ip, vap, cr); ip->i_flag &= ~IQUIET; } update_inode: /* * Setattr always increases the sequence number */ ip->i_seq++; /* * if nfsd and not logging; push synchronously */ if ((curthread->t_flag & T_DONTPEND) && !TRANS_ISTRANS(ufsvfsp)) { ufs_iupdat(ip, 1); } else { ITIMES_NOLOCK(ip); } rw_exit(&ip->i_contents); if (dodqlock) { rw_exit(&ufsvfsp->vfs_dqrwlock); } if (dorwlock) rw_exit(&ip->i_rwlock); if (ulp) { if (dotrans) { int terr = 0; TRANS_END_CSYNC(ufsvfsp, terr, issync, TOP_SETATTR, trans_size); if (error == 0) error = terr; } ufs_lockfs_end(ulp); } out: /* * If out of inodes or blocks, see if we can free something * up from the delete queue. */ if ((error == ENOSPC) && retry && TRANS_ISTRANS(ufsvfsp)) { ufs_delete_drain_wait(ufsvfsp, 1); retry = 0; if (errmsg1 != NULL) kmem_free(errmsg1, len1); if (errmsg2 != NULL) kmem_free(errmsg2, len2); goto again; } if (errmsg1 != NULL) { uprintf(errmsg1); kmem_free(errmsg1, len1); } if (errmsg2 != NULL) { uprintf(errmsg2); kmem_free(errmsg2, len2); } return (error); } /*ARGSUSED*/ static int ufs_access(struct vnode *vp, int mode, int flags, struct cred *cr) { struct inode *ip = VTOI(vp); int error; if (ip->i_ufsvfs == NULL) return (EIO); rw_enter(&ip->i_contents, RW_READER); /* * The ufs_iaccess function wants to be called with * mode bits expressed as "ufs specific" bits. * I.e., VWRITE|VREAD|VEXEC do not make sense to * ufs_iaccess() but IWRITE|IREAD|IEXEC do. * But since they're the same we just pass the vnode mode * bit but just verify that assumption at compile time. */ #if IWRITE != VWRITE || IREAD != VREAD || IEXEC != VEXEC #error "ufs_access needs to map Vmodes to Imodes" #endif error = ufs_iaccess(ip, mode, cr); rw_exit(&ip->i_contents); return (error); } /* ARGSUSED */ static int ufs_readlink(struct vnode *vp, struct uio *uiop, struct cred *cr) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp; struct ulockfs *ulp; int error; int fastsymlink; if (vp->v_type != VLNK) { error = EINVAL; goto nolockout; } /* * If the symbolic link is empty there is nothing to read. * Fast-track these empty symbolic links */ if (ip->i_size == 0) { error = 0; goto nolockout; } ufsvfsp = ip->i_ufsvfs; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_READLINK_MASK); if (error) goto nolockout; /* * The ip->i_rwlock protects the data blocks used for FASTSYMLINK */ again: fastsymlink = 0; if (ip->i_flag & IFASTSYMLNK) { rw_enter(&ip->i_rwlock, RW_READER); rw_enter(&ip->i_contents, RW_READER); if (ip->i_flag & IFASTSYMLNK) { if (!ULOCKFS_IS_NOIACC(ITOUL(ip)) && (ip->i_fs->fs_ronly == 0) && (!ufsvfsp->vfs_noatime)) { mutex_enter(&ip->i_tlock); ip->i_flag |= IACC; mutex_exit(&ip->i_tlock); } error = uiomove((caddr_t)&ip->i_db[1], MIN(ip->i_size, uiop->uio_resid), UIO_READ, uiop); ITIMES(ip); ++fastsymlink; } rw_exit(&ip->i_contents); rw_exit(&ip->i_rwlock); } if (!fastsymlink) { ssize_t size; /* number of bytes read */ caddr_t basep; /* pointer to input data */ ino_t ino; long igen; struct uio tuio; /* temp uio struct */ struct uio *tuiop; iovec_t tiov; /* temp iovec struct */ char kbuf[FSL_SIZE]; /* buffer to hold fast symlink */ int tflag = 0; /* flag to indicate temp vars used */ ino = ip->i_number; igen = ip->i_gen; size = uiop->uio_resid; basep = uiop->uio_iov->iov_base; tuiop = uiop; rw_enter(&ip->i_rwlock, RW_WRITER); rw_enter(&ip->i_contents, RW_WRITER); if (ip->i_flag & IFASTSYMLNK) { rw_exit(&ip->i_contents); rw_exit(&ip->i_rwlock); goto again; } /* can this be a fast symlink and is it a user buffer? */ if (ip->i_size <= FSL_SIZE && (uiop->uio_segflg == UIO_USERSPACE || uiop->uio_segflg == UIO_USERISPACE)) { bzero(&tuio, sizeof (struct uio)); /* * setup a kernel buffer to read link into. this * is to fix a race condition where the user buffer * got corrupted before copying it into the inode. */ size = ip->i_size; tiov.iov_len = size; tiov.iov_base = kbuf; tuio.uio_iov = &tiov; tuio.uio_iovcnt = 1; tuio.uio_offset = uiop->uio_offset; tuio.uio_segflg = UIO_SYSSPACE; tuio.uio_fmode = uiop->uio_fmode; tuio.uio_extflg = uiop->uio_extflg; tuio.uio_limit = uiop->uio_limit; tuio.uio_resid = size; basep = tuio.uio_iov->iov_base; tuiop = &tuio; tflag = 1; } error = rdip(ip, tuiop, 0, cr); if (!(error == 0 && ip->i_number == ino && ip->i_gen == igen)) { rw_exit(&ip->i_contents); rw_exit(&ip->i_rwlock); goto out; } if (tflag == 0) size -= uiop->uio_resid; if ((tflag == 0 && ip->i_size <= FSL_SIZE && ip->i_size == size) || (tflag == 1 && tuio.uio_resid == 0)) { error = kcopy(basep, &ip->i_db[1], ip->i_size); if (error == 0) { ip->i_flag |= IFASTSYMLNK; /* * free page */ (void) VOP_PUTPAGE(ITOV(ip), (offset_t)0, PAGESIZE, (B_DONTNEED | B_FREE | B_FORCE | B_ASYNC), cr); } else { int i; /* error, clear garbage left behind */ for (i = 1; i < NDADDR; i++) ip->i_db[i] = 0; for (i = 0; i < NIADDR; i++) ip->i_ib[i] = 0; } } if (tflag == 1) { /* now, copy it into the user buffer */ error = uiomove((caddr_t)kbuf, MIN(size, uiop->uio_resid), UIO_READ, uiop); } rw_exit(&ip->i_contents); rw_exit(&ip->i_rwlock); } out: if (ulp) { ufs_lockfs_end(ulp); } nolockout: return (error); } /* ARGSUSED */ static int ufs_fsync(struct vnode *vp, int syncflag, struct cred *cr) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp = ip->i_ufsvfs; struct ulockfs *ulp; int error; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_FSYNC_MASK); if (error) return (error); if (TRANS_ISTRANS(ufsvfsp)) { /* * First push out any data pages */ if (vn_has_cached_data(vp) && !(syncflag & FNODSYNC) && (vp->v_type != VCHR) && !(IS_SWAPVP(vp))) { error = VOP_PUTPAGE(vp, (offset_t)0, (size_t)0, 0, CRED()); if (error) goto out; } /* * Delta any delayed inode times updates * and push inode to log. * All other inode deltas will have already been delta'd * and will be pushed during the commit. */ if (!(syncflag & FDSYNC) && ((ip->i_flag & (IMOD|IMODACC)) == IMODACC)) { if (ulp) { TRANS_BEGIN_ASYNC(ufsvfsp, TOP_FSYNC, TOP_SYNCIP_SIZE); } rw_enter(&ip->i_contents, RW_READER); mutex_enter(&ip->i_tlock); ip->i_flag &= ~IMODTIME; mutex_exit(&ip->i_tlock); ufs_iupdat(ip, I_SYNC); rw_exit(&ip->i_contents); if (ulp) { TRANS_END_ASYNC(ufsvfsp, TOP_FSYNC, TOP_SYNCIP_SIZE); } } /* * Commit the Moby transaction * * Deltas have already been made so we just need to * commit them with a synchronous transaction. * TRANS_BEGIN_SYNC() will return an error * if there are no deltas to commit, for an * empty transaction. */ if (ulp) { TRANS_BEGIN_SYNC(ufsvfsp, TOP_FSYNC, TOP_COMMIT_SIZE, error); if (error) { error = 0; /* commit wasn't needed */ goto out; } TRANS_END_SYNC(ufsvfsp, error, TOP_FSYNC, TOP_COMMIT_SIZE); } } else { /* not logging */ if (!(IS_SWAPVP(vp))) if (syncflag & FNODSYNC) { /* Just update the inode only */ TRANS_IUPDAT(ip, 1); error = 0; } else if (syncflag & FDSYNC) /* Do data-synchronous writes */ error = TRANS_SYNCIP(ip, 0, I_DSYNC, TOP_FSYNC); else /* Do synchronous writes */ error = TRANS_SYNCIP(ip, 0, I_SYNC, TOP_FSYNC); rw_enter(&ip->i_contents, RW_WRITER); if (!error) error = ufs_sync_indir(ip); rw_exit(&ip->i_contents); } out: if (ulp) { ufs_lockfs_end(ulp); } return (error); } /*ARGSUSED*/ static void ufs_inactive(struct vnode *vp, struct cred *cr) { ufs_iinactive(VTOI(vp)); } /* * Unix file system operations having to do with directory manipulation. */ int ufs_lookup_idle_count = 2; /* Number of inodes to idle each time */ /* ARGSUSED */ static int ufs_lookup(struct vnode *dvp, char *nm, struct vnode **vpp, struct pathname *pnp, int flags, struct vnode *rdir, struct cred *cr) { struct inode *ip; struct inode *sip; struct inode *xip; struct ufsvfs *ufsvfsp; struct ulockfs *ulp; struct vnode *vp; int error; /* * Check flags for type of lookup (regular file or attribute file) */ ip = VTOI(dvp); if (flags & LOOKUP_XATTR) { /* * We don't allow recursive attributes... * Maybe someday we will. */ if ((ip->i_cflags & IXATTR)) { return (EINVAL); } if ((vp = dnlc_lookup(dvp, XATTR_DIR_NAME)) == NULL) { error = ufs_xattr_getattrdir(dvp, &sip, flags, cr); if (error) { *vpp = NULL; goto out; } vp = ITOV(sip); dnlc_update(dvp, XATTR_DIR_NAME, vp); } /* * Check accessibility of directory. */ if (vp == DNLC_NO_VNODE) { VN_RELE(vp); error = ENOENT; goto out; } if ((error = ufs_iaccess(VTOI(vp), IEXEC, cr)) != 0) { VN_RELE(vp); goto out; } *vpp = vp; return (0); } /* * Check for a null component, which we should treat as * looking at dvp from within it's parent, so we don't * need a call to ufs_iaccess(), as it has already been * done. */ if (nm[0] == 0) { VN_HOLD(dvp); error = 0; *vpp = dvp; goto out; } /* * Check for "." ie itself. this is a quick check and * avoids adding "." into the dnlc (which have been seen * to occupy >10% of the cache). */ if ((nm[0] == '.') && (nm[1] == 0)) { /* * Don't return without checking accessibility * of the directory. We only need the lock if * we are going to return it. */ if ((error = ufs_iaccess(ip, IEXEC, cr)) == 0) { VN_HOLD(dvp); *vpp = dvp; } goto out; } /* * Fast path: Check the directory name lookup cache. */ if (vp = dnlc_lookup(dvp, nm)) { /* * Check accessibility of directory. */ if ((error = ufs_iaccess(ip, IEXEC, cr)) != 0) { VN_RELE(vp); goto out; } if (vp == DNLC_NO_VNODE) { VN_RELE(vp); error = ENOENT; goto out; } xip = VTOI(vp); ulp = NULL; goto fastpath; } /* * Keep the idle queue from getting too long by * idling two inodes before attempting to allocate another. * This operation must be performed before entering * lockfs or a transaction. */ if (ufs_idle_q.uq_ne > ufs_idle_q.uq_hiwat) if ((curthread->t_flag & T_DONTBLOCK) == 0) { ins.in_lidles.value.ul += ufs_lookup_idle_count; ufs_idle_some(ufs_lookup_idle_count); } retry_lookup: ufsvfsp = ip->i_ufsvfs; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_LOOKUP_MASK); if (error) goto out; error = ufs_dirlook(ip, nm, &xip, cr, 1); fastpath: if (error == 0) { ip = xip; *vpp = ITOV(ip); /* * If vnode is a device return special vnode instead. */ if (IS_DEVVP(*vpp)) { struct vnode *newvp; newvp = specvp(*vpp, (*vpp)->v_rdev, (*vpp)->v_type, cr); VN_RELE(*vpp); if (newvp == NULL) error = ENOSYS; else *vpp = newvp; } } if (ulp) { ufs_lockfs_end(ulp); } if (error == EAGAIN) goto retry_lookup; out: return (error); } static int ufs_create(struct vnode *dvp, char *name, struct vattr *vap, enum vcexcl excl, int mode, struct vnode **vpp, struct cred *cr, int flag) { struct inode *ip; struct inode *xip; struct inode *dip; struct vnode *xvp; struct ufsvfs *ufsvfsp; struct ulockfs *ulp; int error; int issync; int truncflag; int trans_size; int noentry; int defer_dip_seq_update = 0; /* need to defer update of dip->i_seq */ int retry = 1; int indeadlock; again: ip = VTOI(dvp); ufsvfsp = ip->i_ufsvfs; truncflag = 0; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_CREATE_MASK); if (error) goto out; if (ulp) { trans_size = (int)TOP_CREATE_SIZE(ip); TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_CREATE, trans_size); } if ((vap->va_mode & VSVTX) && secpolicy_vnode_stky_modify(cr) != 0) vap->va_mode &= ~VSVTX; if (*name == '\0') { /* * Null component name refers to the directory itself. */ VN_HOLD(dvp); /* * Even though this is an error case, we need to grab the * quota lock since the error handling code below is common. */ rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&ip->i_contents, RW_WRITER); error = EEXIST; } else { xip = NULL; noentry = 0; /* * ufs_tryirwlock_trans uses rw_tryenter and checks for SLOCK * to avoid i_rwlock, ufs_lockfs_begin deadlock. If deadlock * possible, retries the operation. */ ufs_tryirwlock_trans(&ip->i_rwlock, RW_WRITER, TOP_CREATE, retry_dir); if (indeadlock) goto again; xvp = dnlc_lookup(dvp, name); if (xvp == DNLC_NO_VNODE) { noentry = 1; VN_RELE(xvp); xvp = NULL; } if (xvp) { rw_exit(&ip->i_rwlock); if (error = ufs_iaccess(ip, IEXEC, cr)) { VN_RELE(xvp); } else { error = EEXIST; xip = VTOI(xvp); } } else { /* * Suppress file system full message if we will retry */ error = ufs_direnter_cm(ip, name, DE_CREATE, vap, &xip, cr, (noentry | (retry ? IQUIET : 0))); if (error == EAGAIN) { if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_CREATE, trans_size); ufs_lockfs_end(ulp); } goto again; } rw_exit(&ip->i_rwlock); } ip = xip; if (ip != NULL) { rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&ip->i_contents, RW_WRITER); } } /* * If the file already exists and this is a non-exclusive create, * check permissions and allow access for non-directories. * Read-only create of an existing directory is also allowed. * We fail an exclusive create of anything which already exists. */ if (error == EEXIST) { dip = VTOI(dvp); if (excl == NONEXCL) { if ((((ip->i_mode & IFMT) == IFDIR) || ((ip->i_mode & IFMT) == IFATTRDIR)) && (mode & IWRITE)) error = EISDIR; else if (mode) error = ufs_iaccess(ip, mode, cr); else error = 0; } if (error) { rw_exit(&ip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); VN_RELE(ITOV(ip)); goto unlock; } /* * If the error EEXIST was set, then i_seq can not * have been updated. The sequence number interface * is defined such that a non-error VOP_CREATE must * increase the dir va_seq it by at least one. If we * have cleared the error, increase i_seq. Note that * we are increasing the dir i_seq and in rare cases * ip may actually be from the dvp, so we already have * the locks and it will not be subject to truncation. * In case we have to update i_seq of the parent * directory dip, we have to defer it till we have * released our locks on ip due to lock ordering requirements. */ if (ip != dip) defer_dip_seq_update = 1; else ip->i_seq++; if (((ip->i_mode & IFMT) == IFREG) && (vap->va_mask & AT_SIZE) && vap->va_size == 0) { /* * Truncate regular files, if requested by caller. * Grab i_rwlock to make sure no one else is * currently writing to the file (we promised * bmap we would do this). * Must get the locks in the correct order. */ if (ip->i_size == 0) { ip->i_flag |= ICHG | IUPD; ip->i_seq++; TRANS_INODE(ufsvfsp, ip); } else { /* * Large Files: Why this check here? * Though we do it in vn_create() we really * want to guarantee that we do not destroy * Large file data by atomically checking * the size while holding the contents * lock. */ if (flag && !(flag & FOFFMAX) && ((ip->i_mode & IFMT) == IFREG) && (ip->i_size > (offset_t)MAXOFF32_T)) { rw_exit(&ip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); error = EOVERFLOW; goto unlock; } if (TRANS_ISTRANS(ufsvfsp)) truncflag++; else { rw_exit(&ip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); ufs_tryirwlock_trans(&ip->i_rwlock, RW_WRITER, TOP_CREATE, retry_file); if (indeadlock) { VN_RELE(ITOV(ip)); goto again; } rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&ip->i_contents, RW_WRITER); (void) ufs_itrunc(ip, (u_offset_t)0, 0, cr); rw_exit(&ip->i_rwlock); } } } } if (error) { if (ip != NULL) { rw_exit(&ufsvfsp->vfs_dqrwlock); rw_exit(&ip->i_contents); } goto unlock; } *vpp = ITOV(ip); ITIMES(ip); rw_exit(&ip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); /* * If vnode is a device return special vnode instead. */ if (!error && IS_DEVVP(*vpp)) { struct vnode *newvp; newvp = specvp(*vpp, (*vpp)->v_rdev, (*vpp)->v_type, cr); VN_RELE(*vpp); if (newvp == NULL) { error = ENOSYS; goto unlock; } truncflag = 0; *vpp = newvp; } unlock: /* * Do the deferred update of the parent directory's sequence * number now. */ if (defer_dip_seq_update == 1) { rw_enter(&dip->i_contents, RW_READER); mutex_enter(&dip->i_tlock); dip->i_seq++; mutex_exit(&dip->i_tlock); rw_exit(&dip->i_contents); } if (ulp) { int terr = 0; TRANS_END_CSYNC(ufsvfsp, terr, issync, TOP_CREATE, trans_size); /* * If we haven't had a more interesting failure * already, then anything that might've happened * here should be reported. */ if (error == 0) error = terr; } if (!error && truncflag) { ufs_tryirwlock(&ip->i_rwlock, RW_WRITER, retry_trunc); if (indeadlock) { if (ulp) ufs_lockfs_end(ulp); VN_RELE(ITOV(ip)); goto again; } (void) TRANS_ITRUNC(ip, (u_offset_t)0, 0, cr); rw_exit(&ip->i_rwlock); } if (ulp) ufs_lockfs_end(ulp); /* * If no inodes available, try to free one up out of the * pending delete queue. */ if ((error == ENOSPC) && retry && TRANS_ISTRANS(ufsvfsp)) { ufs_delete_drain_wait(ufsvfsp, 1); retry = 0; goto again; } out: return (error); } extern int ufs_idle_max; /*ARGSUSED*/ static int ufs_remove(struct vnode *vp, char *nm, struct cred *cr) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp = ip->i_ufsvfs; struct ulockfs *ulp; vnode_t *rmvp = NULL; /* Vnode corresponding to name being removed */ int indeadlock; int error; int issync; int trans_size; /* * don't let the delete queue get too long */ if (ufsvfsp == NULL) { error = EIO; goto out; } if (ufsvfsp->vfs_delete.uq_ne > ufs_idle_max) ufs_delete_drain(vp->v_vfsp, 1, 1); retry_remove: error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_REMOVE_MASK); if (error) goto out; if (ulp) TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_REMOVE, trans_size = (int)TOP_REMOVE_SIZE(VTOI(vp))); /* * ufs_tryirwlock_trans uses rw_tryenter and checks for SLOCK * to avoid i_rwlock, ufs_lockfs_begin deadlock. If deadlock * possible, retries the operation. */ ufs_tryirwlock_trans(&ip->i_rwlock, RW_WRITER, TOP_REMOVE, retry); if (indeadlock) goto retry_remove; error = ufs_dirremove(ip, nm, (struct inode *)0, (struct vnode *)0, DR_REMOVE, cr, &rmvp); rw_exit(&ip->i_rwlock); if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_REMOVE, trans_size); ufs_lockfs_end(ulp); } /* * This must be called after the remove transaction is closed. */ if (rmvp != NULL) { /* Only send the event if there were no errors */ if (error == 0) vnevent_remove(rmvp); VN_RELE(rmvp); } out: return (error); } /* * Link a file or a directory. Only privileged processes are allowed to * make links to directories. */ static int ufs_link(struct vnode *tdvp, struct vnode *svp, char *tnm, struct cred *cr) { struct inode *sip; struct inode *tdp = VTOI(tdvp); struct ufsvfs *ufsvfsp = tdp->i_ufsvfs; struct ulockfs *ulp; struct vnode *realvp; int error; int issync; int trans_size; int isdev; int indeadlock; retry_link: error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_LINK_MASK); if (error) goto out; if (ulp) TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_LINK, trans_size = (int)TOP_LINK_SIZE(VTOI(tdvp))); if (VOP_REALVP(svp, &realvp) == 0) svp = realvp; /* * Make sure link for extended attributes is valid * We only support hard linking of attr in ATTRDIR to ATTRDIR * * Make certain we don't attempt to look at a device node as * a ufs inode. */ isdev = IS_DEVVP(svp); if (((isdev == 0) && ((VTOI(svp)->i_cflags & IXATTR) == 0) && ((tdp->i_mode & IFMT) == IFATTRDIR)) || ((isdev == 0) && (VTOI(svp)->i_cflags & IXATTR) && ((tdp->i_mode & IFMT) == IFDIR))) { error = EINVAL; goto unlock; } sip = VTOI(svp); if ((svp->v_type == VDIR && secpolicy_fs_linkdir(cr, ufsvfsp->vfs_vfs) != 0) || (sip->i_uid != crgetuid(cr) && secpolicy_basic_link(cr) != 0)) { error = EPERM; goto unlock; } /* * ufs_tryirwlock_trans uses rw_tryenter and checks for SLOCK * to avoid i_rwlock, ufs_lockfs_begin deadlock. If deadlock * possible, retries the operation. */ ufs_tryirwlock_trans(&tdp->i_rwlock, RW_WRITER, TOP_LINK, retry); if (indeadlock) goto retry_link; error = ufs_direnter_lr(tdp, tnm, DE_LINK, (struct inode *)0, sip, cr, NULL); rw_exit(&tdp->i_rwlock); unlock: if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_LINK, trans_size); ufs_lockfs_end(ulp); } out: return (error); } uint64_t ufs_rename_retry_cnt; uint64_t ufs_rename_upgrade_retry_cnt; uint64_t ufs_rename_dircheck_retry_cnt; clock_t ufs_rename_backoff_delay = 1; /* * Rename a file or directory. * We are given the vnode and entry string of the source and the * vnode and entry string of the place we want to move the source * to (the target). The essential operation is: * unlink(target); * link(source, target); * unlink(source); * but "atomically". Can't do full commit without saving state in * the inode on disk, which isn't feasible at this time. Best we * can do is always guarantee that the TARGET exists. */ /*ARGSUSED*/ static int ufs_rename( struct vnode *sdvp, /* old (source) parent vnode */ char *snm, /* old (source) entry name */ struct vnode *tdvp, /* new (target) parent vnode */ char *tnm, /* new (target) entry name */ struct cred *cr) { struct inode *sip = NULL; /* source inode */ struct inode *ip = NULL; /* check inode */ struct inode *sdp; /* old (source) parent inode */ struct inode *tdp; /* new (target) parent inode */ struct vnode *tvp = NULL; /* target vnode, if it exists */ struct vnode *realvp; struct ufsvfs *ufsvfsp; struct ulockfs *ulp; struct ufs_slot slot; timestruc_t now; int error; int issync; int trans_size; krwlock_t *first_lock; krwlock_t *second_lock; krwlock_t *reverse_lock; sdp = VTOI(sdvp); slot.fbp = NULL; ufsvfsp = sdp->i_ufsvfs; retry_rename: error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_RENAME_MASK); if (error) goto out; if (ulp) TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_RENAME, trans_size = (int)TOP_RENAME_SIZE(sdp)); if (VOP_REALVP(tdvp, &realvp) == 0) tdvp = realvp; tdp = VTOI(tdvp); /* * We only allow renaming of attributes from ATTRDIR to ATTRDIR. */ if ((tdp->i_mode & IFMT) != (sdp->i_mode & IFMT)) { error = EINVAL; goto unlock; } /* * Look up inode of file we're supposed to rename. */ gethrestime(&now); if (error = ufs_dirlook(sdp, snm, &sip, cr, 0)) { if (error == EAGAIN) { if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_RENAME, trans_size); ufs_lockfs_end(ulp); } goto retry_rename; } goto unlock; } /* * Lock both the source and target directories (they may be * the same) to provide the atomicity semantics that was * previously provided by the per file system vfs_rename_lock * * with vfs_rename_lock removed to allow simultaneous renames * within a file system, ufs_dircheckpath can deadlock while * traversing back to ensure that source is not a parent directory * of target parent directory. This is because we get into * ufs_dircheckpath with the sdp and tdp locks held as RW_WRITER. * If the tdp and sdp of the simultaneous renames happen to be * in the path of each other, it can lead to a deadlock. This * can be avoided by getting the locks as RW_READER here and then * upgrading to RW_WRITER after completing the ufs_dircheckpath. * * We hold the target directory's i_rwlock after calling * ufs_lockfs_begin but in many other operations (like ufs_readdir) * VOP_RWLOCK is explicitly called by the filesystem independent code * before calling the file system operation. In these cases the order * is reversed (i.e i_rwlock is taken first and then ufs_lockfs_begin * is called). This is fine as long as ufs_lockfs_begin acts as a VOP * counter but with ufs_quiesce setting the SLOCK bit this becomes a * synchronizing object which might lead to a deadlock. So we use * rw_tryenter instead of rw_enter. If we fail to get this lock and * find that SLOCK bit is set, we call ufs_lockfs_end and restart the * operation. */ retry: first_lock = &tdp->i_rwlock; second_lock = &sdp->i_rwlock; retry_firstlock: if (!rw_tryenter(first_lock, RW_READER)) { /* * We didn't get the lock. Check if the SLOCK is set in the * ufsvfs. If yes, we might be in a deadlock. Safer to give up * and wait for SLOCK to be cleared. */ if (ulp && ULOCKFS_IS_SLOCK(ulp)) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_RENAME, trans_size); ufs_lockfs_end(ulp); goto retry_rename; } else { /* * SLOCK isn't set so this is a genuine synchronization * case. Let's try again after giving them a breather. */ delay(RETRY_LOCK_DELAY); goto retry_firstlock; } } /* * Need to check if the tdp and sdp are same !!! */ if ((tdp != sdp) && (!rw_tryenter(second_lock, RW_READER))) { /* * We didn't get the lock. Check if the SLOCK is set in the * ufsvfs. If yes, we might be in a deadlock. Safer to give up * and wait for SLOCK to be cleared. */ rw_exit(first_lock); if (ulp && ULOCKFS_IS_SLOCK(ulp)) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_RENAME, trans_size); ufs_lockfs_end(ulp); goto retry_rename; } else { /* * So we couldn't get the second level peer lock *and* * the SLOCK bit isn't set. Too bad we can be * contentding with someone wanting these locks otherway * round. Reverse the locks in case there is a heavy * contention for the second level lock. */ reverse_lock = first_lock; first_lock = second_lock; second_lock = reverse_lock; ufs_rename_retry_cnt++; goto retry_firstlock; } } if (sip == tdp) { error = EINVAL; goto errout; } /* * Make sure we can delete the source entry. This requires * write permission on the containing directory. * Check for sticky directories. */ rw_enter(&sdp->i_contents, RW_READER); rw_enter(&sip->i_contents, RW_READER); if ((error = ufs_iaccess(sdp, IWRITE, cr)) != 0 || (error = ufs_sticky_remove_access(sdp, sip, cr)) != 0) { rw_exit(&sip->i_contents); rw_exit(&sdp->i_contents); goto errout; } /* * If this is a rename of a directory and the parent is * different (".." must be changed), then the source * directory must not be in the directory hierarchy * above the target, as this would orphan everything * below the source directory. Also the user must have * write permission in the source so as to be able to * change "..". */ if ((((sip->i_mode & IFMT) == IFDIR) || ((sip->i_mode & IFMT) == IFATTRDIR)) && sdp != tdp) { ino_t inum; if ((error = ufs_iaccess(sip, IWRITE, cr))) { rw_exit(&sip->i_contents); rw_exit(&sdp->i_contents); goto errout; } inum = sip->i_number; rw_exit(&sip->i_contents); rw_exit(&sdp->i_contents); if ((error = ufs_dircheckpath(inum, tdp, sdp, cr))) { /* * If we got EAGAIN ufs_dircheckpath detected a * potential deadlock and backed out. We need * to retry the operation since sdp and tdp have * to be released to avoid the deadlock. */ if (error == EAGAIN) { rw_exit(&tdp->i_rwlock); if (tdp != sdp) rw_exit(&sdp->i_rwlock); delay(ufs_rename_backoff_delay); ufs_rename_dircheck_retry_cnt++; goto retry; } goto errout; } } else { rw_exit(&sip->i_contents); rw_exit(&sdp->i_contents); } /* * Check for renaming '.' or '..' or alias of '.' */ if (strcmp(snm, ".") == 0 || strcmp(snm, "..") == 0 || sdp == sip) { error = EINVAL; goto errout; } /* * Simultaneous renames can deadlock in ufs_dircheckpath since it * tries to traverse back the file tree with both tdp and sdp held * as RW_WRITER. To avoid that we have to hold the tdp and sdp locks * as RW_READERS till ufs_dircheckpath is done. * Now that ufs_dircheckpath is done with, we can upgrade the locks * to RW_WRITER. */ if (!rw_tryupgrade(&tdp->i_rwlock)) { /* * The upgrade failed. We got to give away the lock * as to avoid deadlocking with someone else who is * waiting for writer lock. With the lock gone, we * cannot be sure the checks done above will hold * good when we eventually get them back as writer. * So if we can't upgrade we drop the locks and retry * everything again. */ rw_exit(&tdp->i_rwlock); if (tdp != sdp) rw_exit(&sdp->i_rwlock); delay(ufs_rename_backoff_delay); ufs_rename_upgrade_retry_cnt++; goto retry; } if (tdp != sdp) { if (!rw_tryupgrade(&sdp->i_rwlock)) { /* * The upgrade failed. We got to give away the lock * as to avoid deadlocking with someone else who is * waiting for writer lock. With the lock gone, we * cannot be sure the checks done above will hold * good when we eventually get them back as writer. * So if we can't upgrade we drop the locks and retry * everything again. */ rw_exit(&tdp->i_rwlock); rw_exit(&sdp->i_rwlock); delay(ufs_rename_backoff_delay); ufs_rename_upgrade_retry_cnt++; goto retry; } } /* * Now that all the locks are held check to make sure another thread * didn't slip in and take out the sip. */ slot.status = NONE; if ((sip->i_ctime.tv_usec * 1000) > now.tv_nsec || sip->i_ctime.tv_sec > now.tv_sec) { rw_enter(&sdp->i_ufsvfs->vfs_dqrwlock, RW_READER); rw_enter(&sdp->i_contents, RW_WRITER); error = ufs_dircheckforname(sdp, snm, strlen(snm), &slot, &ip, cr, 0); rw_exit(&sdp->i_contents); rw_exit(&sdp->i_ufsvfs->vfs_dqrwlock); if (error) { goto errout; } if (ip == NULL) { error = ENOENT; goto errout; } else { /* * If the inode was found need to drop the v_count * so as not to keep the filesystem from being * unmounted at a later time. */ VN_RELE(ITOV(ip)); } /* * Release the slot.fbp that has the page mapped and * locked SE_SHARED, and could be used in in * ufs_direnter_lr() which needs to get the SE_EXCL lock * on said page. */ if (slot.fbp) { fbrelse(slot.fbp, S_OTHER); slot.fbp = NULL; } } /* * Link source to the target. If a target exists, return its * vnode pointer in tvp. We'll release it after sending the * vnevent. */ if (error = ufs_direnter_lr(tdp, tnm, DE_RENAME, sdp, sip, cr, &tvp)) { /* * ESAME isn't really an error; it indicates that the * operation should not be done because the source and target * are the same file, but that no error should be reported. */ if (error == ESAME) error = 0; goto errout; } /* * Unlink the source. * Remove the source entry. ufs_dirremove() checks that the entry * still reflects sip, and returns an error if it doesn't. * If the entry has changed just forget about it. Release * the source inode. */ if ((error = ufs_dirremove(sdp, snm, sip, (struct vnode *)0, DR_RENAME, cr, NULL)) == ENOENT) error = 0; errout: if (slot.fbp) fbrelse(slot.fbp, S_OTHER); rw_exit(&tdp->i_rwlock); if (sdp != tdp) { rw_exit(&sdp->i_rwlock); } unlock: if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_RENAME, trans_size); ufs_lockfs_end(ulp); } /* * If no errors, send the appropriate events on the source * and destination (a.k.a, target) vnodes, if they exist. * This has to be done after the rename transaction has closed. */ if (error == 0) { if (tvp != NULL) vnevent_rename_dest(tvp); /* * Note that if ufs_direnter_lr() returned ESAME then * this event will still be sent. This isn't expected * to be a problem for anticipated usage by consumers. */ if (sip != NULL) vnevent_rename_src(ITOV(sip)); } if (tvp != NULL) VN_RELE(tvp); if (sip != NULL) VN_RELE(ITOV(sip)); out: return (error); } /*ARGSUSED*/ static int ufs_mkdir(struct vnode *dvp, char *dirname, struct vattr *vap, struct vnode **vpp, struct cred *cr) { struct inode *ip; struct inode *xip; struct ufsvfs *ufsvfsp; struct ulockfs *ulp; int error; int issync; int trans_size; int indeadlock; int retry = 1; ASSERT((vap->va_mask & (AT_TYPE|AT_MODE)) == (AT_TYPE|AT_MODE)); /* * Can't make directory in attr hidden dir */ if ((VTOI(dvp)->i_mode & IFMT) == IFATTRDIR) return (EINVAL); again: ip = VTOI(dvp); ufsvfsp = ip->i_ufsvfs; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_MKDIR_MASK); if (error) goto out; if (ulp) TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_MKDIR, trans_size = (int)TOP_MKDIR_SIZE(ip)); /* * ufs_tryirwlock_trans uses rw_tryenter and checks for SLOCK * to avoid i_rwlock, ufs_lockfs_begin deadlock. If deadlock * possible, retries the operation. */ ufs_tryirwlock_trans(&ip->i_rwlock, RW_WRITER, TOP_MKDIR, retry); if (indeadlock) goto again; error = ufs_direnter_cm(ip, dirname, DE_MKDIR, vap, &xip, cr, (retry ? IQUIET : 0)); if (error == EAGAIN) { if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_MKDIR, trans_size); ufs_lockfs_end(ulp); } goto again; } rw_exit(&ip->i_rwlock); if (error == 0) { ip = xip; *vpp = ITOV(ip); } else if (error == EEXIST) VN_RELE(ITOV(xip)); if (ulp) { int terr = 0; TRANS_END_CSYNC(ufsvfsp, terr, issync, TOP_MKDIR, trans_size); ufs_lockfs_end(ulp); if (error == 0) error = terr; } out: if ((error == ENOSPC) && retry && TRANS_ISTRANS(ufsvfsp)) { ufs_delete_drain_wait(ufsvfsp, 1); retry = 0; goto again; } return (error); } /*ARGSUSED*/ static int ufs_rmdir(struct vnode *vp, char *nm, struct vnode *cdir, struct cred *cr) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp = ip->i_ufsvfs; struct ulockfs *ulp; vnode_t *rmvp = NULL; /* Vnode of removed directory */ int error; int issync; int trans_size; int indeadlock; /* * don't let the delete queue get too long */ if (ufsvfsp == NULL) { error = EIO; goto out; } if (ufsvfsp->vfs_delete.uq_ne > ufs_idle_max) ufs_delete_drain(vp->v_vfsp, 1, 1); retry_rmdir: error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_RMDIR_MASK); if (error) goto out; if (ulp) TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_RMDIR, trans_size = TOP_RMDIR_SIZE); /* * ufs_tryirwlock_trans uses rw_tryenter and checks for SLOCK * to avoid i_rwlock, ufs_lockfs_begin deadlock. If deadlock * possible, retries the operation. */ ufs_tryirwlock_trans(&ip->i_rwlock, RW_WRITER, TOP_RMDIR, retry); if (indeadlock) goto retry_rmdir; error = ufs_dirremove(ip, nm, (struct inode *)0, cdir, DR_RMDIR, cr, &rmvp); rw_exit(&ip->i_rwlock); if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_RMDIR, trans_size); ufs_lockfs_end(ulp); } /* * This must be done AFTER the rmdir transaction has closed. */ if (rmvp != NULL) { /* Only send the event if there were no errors */ if (error == 0) vnevent_rmdir(rmvp); VN_RELE(rmvp); } out: return (error); } /* ARGSUSED */ static int ufs_readdir( struct vnode *vp, struct uio *uiop, struct cred *cr, int *eofp) { struct iovec *iovp; struct inode *ip; struct direct *idp; struct dirent64 *odp; struct fbuf *fbp; struct ufsvfs *ufsvfsp; struct ulockfs *ulp; caddr_t outbuf; size_t bufsize; uint_t offset; uint_t bytes_wanted, total_bytes_wanted; int incount = 0; int outcount = 0; int error; ip = VTOI(vp); ASSERT(RW_READ_HELD(&ip->i_rwlock)); if (uiop->uio_loffset >= MAXOFF32_T) { if (eofp) *eofp = 1; return (0); } /* * Check if we have been called with a valid iov_len * and bail out if not, otherwise we may potentially loop * forever further down. */ if (uiop->uio_iov->iov_len <= 0) { error = EINVAL; goto out; } /* * Large Files: When we come here we are guaranteed that * uio_offset can be used safely. The high word is zero. */ ufsvfsp = ip->i_ufsvfs; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_READDIR_MASK); if (error) goto out; iovp = uiop->uio_iov; total_bytes_wanted = iovp->iov_len; /* Large Files: directory files should not be "large" */ ASSERT(ip->i_size <= MAXOFF32_T); /* Force offset to be valid (to guard against bogus lseek() values) */ offset = (uint_t)uiop->uio_offset & ~(DIRBLKSIZ - 1); /* Quit if at end of file or link count of zero (posix) */ if (offset >= (uint_t)ip->i_size || ip->i_nlink <= 0) { if (eofp) *eofp = 1; error = 0; goto unlock; } /* * Get space to change directory entries into fs independent format. * Do fast alloc for the most commonly used-request size (filesystem * block size). */ if (uiop->uio_segflg != UIO_SYSSPACE || uiop->uio_iovcnt != 1) { bufsize = total_bytes_wanted; outbuf = kmem_alloc(bufsize, KM_SLEEP); odp = (struct dirent64 *)outbuf; } else { bufsize = total_bytes_wanted; odp = (struct dirent64 *)iovp->iov_base; } nextblk: bytes_wanted = total_bytes_wanted; /* Truncate request to file size */ if (offset + bytes_wanted > (int)ip->i_size) bytes_wanted = (int)(ip->i_size - offset); /* Comply with MAXBSIZE boundary restrictions of fbread() */ if ((offset & MAXBOFFSET) + bytes_wanted > MAXBSIZE) bytes_wanted = MAXBSIZE - (offset & MAXBOFFSET); /* * Read in the next chunk. * We are still holding the i_rwlock. */ error = fbread(vp, (offset_t)offset, bytes_wanted, S_OTHER, &fbp); if (error) goto update_inode; if (!ULOCKFS_IS_NOIACC(ITOUL(ip)) && (ip->i_fs->fs_ronly == 0) && (!ufsvfsp->vfs_noatime)) { ip->i_flag |= IACC; } incount = 0; idp = (struct direct *)fbp->fb_addr; if (idp->d_ino == 0 && idp->d_reclen == 0 && idp->d_namlen == 0) { cmn_err(CE_WARN, "ufs_readdir: bad dir, inumber = %llu, " "fs = %s\n", (u_longlong_t)ip->i_number, ufsvfsp->vfs_fs->fs_fsmnt); fbrelse(fbp, S_OTHER); error = ENXIO; goto update_inode; } /* Transform to file-system independent format */ while (incount < bytes_wanted) { /* * If the current directory entry is mangled, then skip * to the next block. It would be nice to set the FSBAD * flag in the super-block so that a fsck is forced on * next reboot, but locking is a problem. */ if (idp->d_reclen & 0x3) { offset = (offset + DIRBLKSIZ) & ~(DIRBLKSIZ-1); break; } /* Skip to requested offset and skip empty entries */ if (idp->d_ino != 0 && offset >= (uint_t)uiop->uio_offset) { ushort_t this_reclen = DIRENT64_RECLEN(idp->d_namlen); /* Buffer too small for any entries */ if (!outcount && this_reclen > bufsize) { fbrelse(fbp, S_OTHER); error = EINVAL; goto update_inode; } /* If would overrun the buffer, quit */ if (outcount + this_reclen > bufsize) { break; } /* Take this entry */ odp->d_ino = (ino64_t)idp->d_ino; odp->d_reclen = (ushort_t)this_reclen; odp->d_off = (offset_t)(offset + idp->d_reclen); /* use strncpy(9f) to zero out uninitialized bytes */ ASSERT(strlen(idp->d_name) + 1 <= DIRENT64_NAMELEN(this_reclen)); (void) strncpy(odp->d_name, idp->d_name, DIRENT64_NAMELEN(this_reclen)); outcount += odp->d_reclen; odp = (struct dirent64 *) ((intptr_t)odp + odp->d_reclen); ASSERT(outcount <= bufsize); } if (idp->d_reclen) { incount += idp->d_reclen; offset += idp->d_reclen; idp = (struct direct *)((intptr_t)idp + idp->d_reclen); } else { offset = (offset + DIRBLKSIZ) & ~(DIRBLKSIZ-1); break; } } /* Release the chunk */ fbrelse(fbp, S_OTHER); /* Read whole block, but got no entries, read another if not eof */ /* * Large Files: casting i_size to int here is not a problem * because directory sizes are always less than MAXOFF32_T. * See assertion above. */ if (offset < (int)ip->i_size && !outcount) goto nextblk; /* Copy out the entry data */ if (uiop->uio_segflg == UIO_SYSSPACE && uiop->uio_iovcnt == 1) { iovp->iov_base += outcount; iovp->iov_len -= outcount; uiop->uio_resid -= outcount; uiop->uio_offset = offset; } else if ((error = uiomove(outbuf, (long)outcount, UIO_READ, uiop)) == 0) uiop->uio_offset = offset; update_inode: ITIMES(ip); if (uiop->uio_segflg != UIO_SYSSPACE || uiop->uio_iovcnt != 1) kmem_free(outbuf, bufsize); if (eofp && error == 0) *eofp = (uiop->uio_offset >= (int)ip->i_size); unlock: if (ulp) { ufs_lockfs_end(ulp); } out: return (error); } /*ARGSUSED*/ static int ufs_symlink( struct vnode *dvp, /* ptr to parent dir vnode */ char *linkname, /* name of symbolic link */ struct vattr *vap, /* attributes */ char *target, /* target path */ struct cred *cr) /* user credentials */ { struct inode *ip, *dip = VTOI(dvp); struct ufsvfs *ufsvfsp = dip->i_ufsvfs; struct ulockfs *ulp; int error; int issync; int trans_size; int residual; int ioflag; int retry = 1; /* * No symlinks in attrdirs at this time */ if ((VTOI(dvp)->i_mode & IFMT) == IFATTRDIR) return (EINVAL); again: ip = (struct inode *)NULL; vap->va_type = VLNK; vap->va_rdev = 0; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_SYMLINK_MASK); if (error) goto out; if (ulp) TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_SYMLINK, trans_size = (int)TOP_SYMLINK_SIZE(dip)); /* * We must create the inode before the directory entry, to avoid * racing with readlink(). ufs_dirmakeinode requires that we * hold the quota lock as reader, and directory locks as writer. */ rw_enter(&dip->i_rwlock, RW_WRITER); rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&dip->i_contents, RW_WRITER); /* * Suppress any out of inodes messages if we will retry on * ENOSP */ if (retry) dip->i_flag |= IQUIET; error = ufs_dirmakeinode(dip, &ip, vap, DE_SYMLINK, cr); dip->i_flag &= ~IQUIET; rw_exit(&dip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); rw_exit(&dip->i_rwlock); if (error) goto unlock; /* * OK. The inode has been created. Write out the data of the * symbolic link. Since symbolic links are metadata, and should * remain consistent across a system crash, we need to force the * data out synchronously. * * (This is a change from the semantics in earlier releases, which * only created symbolic links synchronously if the semi-documented * 'syncdir' option was set, or if we were being invoked by the NFS * server, which requires symbolic links to be created synchronously.) * * We need to pass in a pointer for the residual length; otherwise * ufs_rdwri() will always return EIO if it can't write the data, * even if the error was really ENOSPC or EDQUOT. */ ioflag = FWRITE | FDSYNC; residual = 0; rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&ip->i_contents, RW_WRITER); /* * Suppress file system full messages if we will retry */ if (retry) ip->i_flag |= IQUIET; error = ufs_rdwri(UIO_WRITE, ioflag, ip, target, strlen(target), (offset_t)0, UIO_SYSSPACE, &residual, cr); ip->i_flag &= ~IQUIET; if (error) { rw_exit(&ip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); goto remove; } /* * If the link's data is small enough, we can cache it in the inode. * This is a "fast symbolic link". We don't use the first direct * block because that's actually used to point at the symbolic link's * contents on disk; but we know that none of the other direct or * indirect blocks can be used because symbolic links are restricted * to be smaller than a file system block. */ ASSERT(MAXPATHLEN <= VBSIZE(ITOV(ip))); if (ip->i_size > 0 && ip->i_size <= FSL_SIZE) { if (kcopy(target, &ip->i_db[1], ip->i_size) == 0) { ip->i_flag |= IFASTSYMLNK; } else { int i; /* error, clear garbage left behind */ for (i = 1; i < NDADDR; i++) ip->i_db[i] = 0; for (i = 0; i < NIADDR; i++) ip->i_ib[i] = 0; } } rw_exit(&ip->i_contents); rw_exit(&ufsvfsp->vfs_dqrwlock); /* * OK. We've successfully created the symbolic link. All that * remains is to insert it into the appropriate directory. */ rw_enter(&dip->i_rwlock, RW_WRITER); error = ufs_direnter_lr(dip, linkname, DE_SYMLINK, NULL, ip, cr, NULL); rw_exit(&dip->i_rwlock); /* * Fall through into remove-on-error code. We're either done, or we * need to remove the inode (if we couldn't insert it). */ remove: if (error && (ip != NULL)) { rw_enter(&ip->i_contents, RW_WRITER); ip->i_nlink--; ip->i_flag |= ICHG; ip->i_seq++; ufs_setreclaim(ip); rw_exit(&ip->i_contents); } unlock: if (ip != NULL) VN_RELE(ITOV(ip)); if (ulp) { int terr = 0; TRANS_END_CSYNC(ufsvfsp, terr, issync, TOP_SYMLINK, trans_size); ufs_lockfs_end(ulp); if (error == 0) error = terr; } /* * We may have failed due to lack of an inode or of a block to * store the target in. Try flushing the delete queue to free * logically-available things up and try again. */ if ((error == ENOSPC) && retry && TRANS_ISTRANS(ufsvfsp)) { ufs_delete_drain_wait(ufsvfsp, 1); retry = 0; goto again; } out: return (error); } /* * Ufs specific routine used to do ufs io. */ int ufs_rdwri(enum uio_rw rw, int ioflag, struct inode *ip, caddr_t base, ssize_t len, offset_t offset, enum uio_seg seg, int *aresid, struct cred *cr) { struct uio auio; struct iovec aiov; int error; ASSERT(RW_LOCK_HELD(&ip->i_contents)); bzero((caddr_t)&auio, sizeof (uio_t)); bzero((caddr_t)&aiov, sizeof (iovec_t)); aiov.iov_base = base; aiov.iov_len = len; auio.uio_iov = &aiov; auio.uio_iovcnt = 1; auio.uio_loffset = offset; auio.uio_segflg = (short)seg; auio.uio_resid = len; if (rw == UIO_WRITE) { auio.uio_fmode = FWRITE; auio.uio_extflg = UIO_COPY_DEFAULT; auio.uio_llimit = curproc->p_fsz_ctl; error = wrip(ip, &auio, ioflag, cr); } else { auio.uio_fmode = FREAD; auio.uio_extflg = UIO_COPY_CACHED; auio.uio_llimit = MAXOFFSET_T; error = rdip(ip, &auio, ioflag, cr); } if (aresid) { *aresid = auio.uio_resid; } else if (auio.uio_resid) { error = EIO; } return (error); } static int ufs_fid(vp, fidp) struct vnode *vp; struct fid *fidp; { struct ufid *ufid; struct inode *ip = VTOI(vp); if (ip->i_ufsvfs == NULL) return (EIO); if (fidp->fid_len < (sizeof (struct ufid) - sizeof (ushort_t))) { fidp->fid_len = sizeof (struct ufid) - sizeof (ushort_t); return (ENOSPC); } ufid = (struct ufid *)fidp; bzero((char *)ufid, sizeof (struct ufid)); ufid->ufid_len = sizeof (struct ufid) - sizeof (ushort_t); ufid->ufid_ino = ip->i_number; ufid->ufid_gen = ip->i_gen; return (0); } /* ARGSUSED2 */ static int ufs_rwlock(struct vnode *vp, int write_lock, caller_context_t *ctp) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp; int forcedirectio; /* * Read case is easy. */ if (!write_lock) { rw_enter(&ip->i_rwlock, RW_READER); return (V_WRITELOCK_FALSE); } /* * Caller has requested a writer lock, but that inhibits any * concurrency in the VOPs that follow. Acquire the lock shared * and defer exclusive access until it is known to be needed in * other VOP handlers. Some cases can be determined here. */ /* * If directio is not set, there is no chance of concurrency, * so just acquire the lock exclusive. Beware of a forced * unmount before looking at the mount option. */ ufsvfsp = ip->i_ufsvfs; forcedirectio = ufsvfsp ? ufsvfsp->vfs_forcedirectio : 0; if (!(ip->i_flag & IDIRECTIO || forcedirectio) || !ufs_allow_shared_writes) { rw_enter(&ip->i_rwlock, RW_WRITER); return (V_WRITELOCK_TRUE); } /* * Mandatory locking forces acquiring i_rwlock exclusive. */ if (MANDLOCK(vp, ip->i_mode)) { rw_enter(&ip->i_rwlock, RW_WRITER); return (V_WRITELOCK_TRUE); } /* * Acquire the lock shared in case a concurrent write follows. * Mandatory locking could have become enabled before the lock * was acquired. Re-check and upgrade if needed. */ rw_enter(&ip->i_rwlock, RW_READER); if (MANDLOCK(vp, ip->i_mode)) { rw_exit(&ip->i_rwlock); rw_enter(&ip->i_rwlock, RW_WRITER); return (V_WRITELOCK_TRUE); } return (V_WRITELOCK_FALSE); } /*ARGSUSED*/ static void ufs_rwunlock(struct vnode *vp, int write_lock, caller_context_t *ctp) { struct inode *ip = VTOI(vp); rw_exit(&ip->i_rwlock); } /* ARGSUSED */ static int ufs_seek(struct vnode *vp, offset_t ooff, offset_t *noffp) { return ((*noffp < 0 || *noffp > MAXOFFSET_T) ? EINVAL : 0); } /* ARGSUSED */ static int ufs_frlock(struct vnode *vp, int cmd, struct flock64 *bfp, int flag, offset_t offset, struct flk_callback *flk_cbp, struct cred *cr) { struct inode *ip = VTOI(vp); if (ip->i_ufsvfs == NULL) return (EIO); /* * If file is being mapped, disallow frlock. * XXX I am not holding tlock while checking i_mapcnt because the * current locking strategy drops all locks before calling fs_frlock. * So, mapcnt could change before we enter fs_frlock making is * meaningless to have held tlock in the first place. */ if (ip->i_mapcnt > 0 && MANDLOCK(vp, ip->i_mode)) return (EAGAIN); return (fs_frlock(vp, cmd, bfp, flag, offset, flk_cbp, cr)); } /* ARGSUSED */ static int ufs_space(struct vnode *vp, int cmd, struct flock64 *bfp, int flag, offset_t offset, cred_t *cr, caller_context_t *ct) { struct ufsvfs *ufsvfsp = VTOI(vp)->i_ufsvfs; struct ulockfs *ulp; int error; if ((error = convoff(vp, bfp, 0, offset)) == 0) { if (cmd == F_FREESP) { error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_SPACE_MASK); if (error) return (error); error = ufs_freesp(vp, bfp, flag, cr); } else if (cmd == F_ALLOCSP) { error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_FALLOCATE_MASK); if (error) return (error); error = ufs_allocsp(vp, bfp, cr); } else return (EINVAL); /* Command not handled here */ if (ulp) ufs_lockfs_end(ulp); } return (error); } /* * Used to determine if read ahead should be done. Also used to * to determine when write back occurs. */ #define CLUSTSZ(ip) ((ip)->i_ufsvfs->vfs_ioclustsz) /* * A faster version of ufs_getpage. * * We optimize by inlining the pvn_getpages iterator, eliminating * calls to bmap_read if file doesn't have UFS holes, and avoiding * the overhead of page_exists(). * * When files has UFS_HOLES and ufs_getpage is called with S_READ, * we set *protp to PROT_READ to avoid calling bmap_read. This approach * victimizes performance when a file with UFS holes is faulted * first in the S_READ mode, and then in the S_WRITE mode. We will get * two MMU faults in this case. * * XXX - the inode fields which control the sequential mode are not * protected by any mutex. The read ahead will act wild if * multiple processes will access the file concurrently and * some of them in sequential mode. One particulary bad case * is if another thread will change the value of i_nextrio between * the time this thread tests the i_nextrio value and then reads it * again to use it as the offset for the read ahead. */ static int ufs_getpage(struct vnode *vp, offset_t off, size_t len, uint_t *protp, page_t *plarr[], size_t plsz, struct seg *seg, caddr_t addr, enum seg_rw rw, struct cred *cr) { u_offset_t uoff = (u_offset_t)off; /* type conversion */ u_offset_t pgoff; u_offset_t eoff; struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp = ip->i_ufsvfs; struct fs *fs; struct ulockfs *ulp; page_t **pl; caddr_t pgaddr; krw_t rwtype; int err; int has_holes; int beyond_eof; int seqmode; int pgsize = PAGESIZE; int dolock; int do_qlock; int trans_size; ASSERT((uoff & PAGEOFFSET) == 0); if (protp) *protp = PROT_ALL; /* * Obey the lockfs protocol */ err = ufs_lockfs_begin_getpage(ufsvfsp, &ulp, seg, rw == S_READ || rw == S_EXEC, protp); if (err) goto out; fs = ufsvfsp->vfs_fs; if (ulp && (rw == S_CREATE || rw == S_WRITE) && !(vp->v_flag & VISSWAP)) { /* * Try to start a transaction, will return if blocking is * expected to occur and the address space is not the * kernel address space. */ trans_size = TOP_GETPAGE_SIZE(ip); if (seg->s_as != &kas) { TRANS_TRY_BEGIN_ASYNC(ufsvfsp, TOP_GETPAGE, trans_size, err) if (err == EWOULDBLOCK) { /* * Use EDEADLK here because the VM code * can normally never see this error. */ err = EDEADLK; ufs_lockfs_end(ulp); goto out; } } else { TRANS_BEGIN_ASYNC(ufsvfsp, TOP_GETPAGE, trans_size); } } if (vp->v_flag & VNOMAP) { err = ENOSYS; goto unlock; } seqmode = ip->i_nextr == uoff && rw != S_CREATE; rwtype = RW_READER; /* start as a reader */ dolock = (rw_owner(&ip->i_contents) != curthread); /* * If this thread owns the lock, i.e., this thread grabbed it * as writer somewhere above, then we don't need to grab the * lock as reader in this routine. */ do_qlock = (rw_owner(&ufsvfsp->vfs_dqrwlock) != curthread); retrylock: if (dolock) { /* * Grab the quota lock if we need to call * bmap_write() below (with i_contents as writer). */ if (do_qlock && rwtype == RW_WRITER) rw_enter(&ufsvfsp->vfs_dqrwlock, RW_READER); rw_enter(&ip->i_contents, rwtype); } /* * We may be getting called as a side effect of a bmap using * fbread() when the blocks might be being allocated and the * size has not yet been up'ed. In this case we want to be * able to return zero pages if we get back UFS_HOLE from * calling bmap for a non write case here. We also might have * to read some frags from the disk into a page if we are * extending the number of frags for a given lbn in bmap(). * Large Files: The read of i_size here is atomic because * i_contents is held here. If dolock is zero, the lock * is held in bmap routines. */ beyond_eof = uoff + len > ip->i_size + PAGEOFFSET; if (beyond_eof && seg != segkmap) { if (dolock) { rw_exit(&ip->i_contents); if (do_qlock && rwtype == RW_WRITER) rw_exit(&ufsvfsp->vfs_dqrwlock); } err = EFAULT; goto unlock; } /* * Must hold i_contents lock throughout the call to pvn_getpages * since locked pages are returned from each call to ufs_getapage. * Must *not* return locked pages and then try for contents lock * due to lock ordering requirements (inode > page) */ has_holes = bmap_has_holes(ip); if ((rw == S_WRITE || rw == S_CREATE) && has_holes && !beyond_eof) { int blk_size; u_offset_t offset; /* * We must acquire the RW_WRITER lock in order to * call bmap_write(). */ if (dolock && rwtype == RW_READER) { rwtype = RW_WRITER; /* * Grab the quota lock before * upgrading i_contents, but if we can't grab it * don't wait here due to lock order: * vfs_dqrwlock > i_contents. */ if (do_qlock && rw_tryenter(&ufsvfsp->vfs_dqrwlock, RW_READER) == 0) { rw_exit(&ip->i_contents); goto retrylock; } if (!rw_tryupgrade(&ip->i_contents)) { rw_exit(&ip->i_contents); if (do_qlock) rw_exit(&ufsvfsp->vfs_dqrwlock); goto retrylock; } } /* * May be allocating disk blocks for holes here as * a result of mmap faults. write(2) does the bmap_write * in rdip/wrip, not here. We are not dealing with frags * in this case. */ /* * Large Files: We cast fs_bmask field to offset_t * just as we do for MAXBMASK because uoff is a 64-bit * data type. fs_bmask will still be a 32-bit type * as we cannot change any ondisk data structures. */ offset = uoff & (offset_t)fs->fs_bmask; while (offset < uoff + len) { blk_size = (int)blksize(fs, ip, lblkno(fs, offset)); err = bmap_write(ip, offset, blk_size, BI_NORMAL, NULL, cr); if (ip->i_flag & (ICHG|IUPD)) ip->i_seq++; if (err) goto update_inode; offset += blk_size; /* XXX - make this contig */ } } /* * Can be a reader from now on. */ if (dolock && rwtype == RW_WRITER) { rw_downgrade(&ip->i_contents); /* * We can release vfs_dqrwlock early so do it, but make * sure we don't try to release it again at the bottom. */ if (do_qlock) { rw_exit(&ufsvfsp->vfs_dqrwlock); do_qlock = 0; } } /* * We remove PROT_WRITE in cases when the file has UFS holes * because we don't want to call bmap_read() to check each * page if it is backed with a disk block. */ if (protp && has_holes && rw != S_WRITE && rw != S_CREATE) *protp &= ~PROT_WRITE; err = 0; /* * The loop looks up pages in the range [off, off + len). * For each page, we first check if we should initiate an asynchronous * read ahead before we call page_lookup (we may sleep in page_lookup * for a previously initiated disk read). */ eoff = (uoff + len); for (pgoff = uoff, pgaddr = addr, pl = plarr; pgoff < eoff; /* empty */) { page_t *pp; u_offset_t nextrio; se_t se; int retval; se = ((rw == S_CREATE || rw == S_OTHER) ? SE_EXCL : SE_SHARED); /* Handle async getpage (faultahead) */ if (plarr == NULL) { ip->i_nextrio = pgoff; (void) ufs_getpage_ra(vp, pgoff, seg, pgaddr); pgoff += pgsize; pgaddr += pgsize; continue; } /* * Check if we should initiate read ahead of next cluster. * We call page_exists only when we need to confirm that * we have the current page before we initiate the read ahead. */ nextrio = ip->i_nextrio; if (seqmode && pgoff + CLUSTSZ(ip) >= nextrio && pgoff <= nextrio && nextrio < ip->i_size && page_exists(vp, pgoff)) { retval = ufs_getpage_ra(vp, pgoff, seg, pgaddr); /* * We always read ahead the next cluster of data * starting from i_nextrio. If the page (vp,nextrio) * is actually in core at this point, the routine * ufs_getpage_ra() will stop pre-fetching data * until we read that page in a synchronized manner * through ufs_getpage_miss(). So, we should increase * i_nextrio if the page (vp, nextrio) exists. */ if ((retval == 0) && page_exists(vp, nextrio)) { ip->i_nextrio = nextrio + pgsize; } } if ((pp = page_lookup(vp, pgoff, se)) != NULL) { /* * We found the page in the page cache. */ *pl++ = pp; pgoff += pgsize; pgaddr += pgsize; len -= pgsize; plsz -= pgsize; } else { /* * We have to create the page, or read it from disk. */ if (err = ufs_getpage_miss(vp, pgoff, len, seg, pgaddr, pl, plsz, rw, seqmode)) goto error; while (*pl != NULL) { pl++; pgoff += pgsize; pgaddr += pgsize; len -= pgsize; plsz -= pgsize; } } } /* * Return pages up to plsz if they are in the page cache. * We cannot return pages if there is a chance that they are * backed with a UFS hole and rw is S_WRITE or S_CREATE. */ if (plarr && !(has_holes && (rw == S_WRITE || rw == S_CREATE))) { ASSERT((protp == NULL) || !(has_holes && (*protp & PROT_WRITE))); eoff = pgoff + plsz; while (pgoff < eoff) { page_t *pp; if ((pp = page_lookup_nowait(vp, pgoff, SE_SHARED)) == NULL) break; *pl++ = pp; pgoff += pgsize; plsz -= pgsize; } } if (plarr) *pl = NULL; /* Terminate page list */ ip->i_nextr = pgoff; error: if (err && plarr) { /* * Release any pages we have locked. */ while (pl > &plarr[0]) page_unlock(*--pl); plarr[0] = NULL; } update_inode: /* * If the inode is not already marked for IACC (in rdip() for read) * and the inode is not marked for no access time update (in wrip() * for write) then update the inode access time and mod time now. */ if ((ip->i_flag & (IACC | INOACC)) == 0) { if ((rw != S_OTHER) && (ip->i_mode & IFMT) != IFDIR) { if (!ULOCKFS_IS_NOIACC(ITOUL(ip)) && (fs->fs_ronly == 0) && (!ufsvfsp->vfs_noatime)) { mutex_enter(&ip->i_tlock); ip->i_flag |= IACC; ITIMES_NOLOCK(ip); mutex_exit(&ip->i_tlock); } } } if (dolock) { rw_exit(&ip->i_contents); if (do_qlock && rwtype == RW_WRITER) rw_exit(&ufsvfsp->vfs_dqrwlock); } unlock: if (ulp) { if ((rw == S_CREATE || rw == S_WRITE) && !(vp->v_flag & VISSWAP)) { TRANS_END_ASYNC(ufsvfsp, TOP_GETPAGE, trans_size); } ufs_lockfs_end(ulp); } out: return (err); } /* * ufs_getpage_miss is called when ufs_getpage missed the page in the page * cache. The page is either read from the disk, or it's created. * A page is created (without disk read) if rw == S_CREATE, or if * the page is not backed with a real disk block (UFS hole). */ /* ARGSUSED */ static int ufs_getpage_miss(struct vnode *vp, u_offset_t off, size_t len, struct seg *seg, caddr_t addr, page_t *pl[], size_t plsz, enum seg_rw rw, int seq) { struct inode *ip = VTOI(vp); page_t *pp; daddr_t bn; size_t io_len; int crpage = 0; int err; int contig; int bsize = ip->i_fs->fs_bsize; /* * Figure out whether the page can be created, or must be * must be read from the disk. */ if (rw == S_CREATE) crpage = 1; else { contig = 0; if (err = bmap_read(ip, off, &bn, &contig)) return (err); crpage = (bn == UFS_HOLE); /* * If its also a fallocated block that hasn't been written to * yet, we will treat it just like a UFS_HOLE and create * a zero page for it */ if (ISFALLOCBLK(ip, bn)) crpage = 1; } if (crpage) { if ((pp = page_create_va(vp, off, PAGESIZE, PG_WAIT, seg, addr)) == NULL) { return (ufs_fault(vp, "ufs_getpage_miss: page_create == NULL")); } if (rw != S_CREATE) pagezero(pp, 0, PAGESIZE); io_len = PAGESIZE; } else { u_offset_t io_off; uint_t xlen; struct buf *bp; ufsvfs_t *ufsvfsp = ip->i_ufsvfs; /* * If access is not in sequential order, we read from disk * in bsize units. * * We limit the size of the transfer to bsize if we are reading * from the beginning of the file. Note in this situation we * will hedge our bets and initiate an async read ahead of * the second block. */ if (!seq || off == 0) contig = MIN(contig, bsize); pp = pvn_read_kluster(vp, off, seg, addr, &io_off, &io_len, off, contig, 0); /* * Some other thread has entered the page. * ufs_getpage will retry page_lookup. */ if (pp == NULL) { pl[0] = NULL; return (0); } /* * Zero part of the page which we are not * going to read from the disk. */ xlen = io_len & PAGEOFFSET; if (xlen != 0) pagezero(pp->p_prev, xlen, PAGESIZE - xlen); bp = pageio_setup(pp, io_len, ip->i_devvp, B_READ); bp->b_edev = ip->i_dev; bp->b_dev = cmpdev(ip->i_dev); bp->b_blkno = bn; bp->b_un.b_addr = (caddr_t)0; bp->b_file = ip->i_vnode; bp->b_offset = off; if (ufsvfsp->vfs_log) { lufs_read_strategy(ufsvfsp->vfs_log, bp); } else if (ufsvfsp->vfs_snapshot) { fssnap_strategy(&ufsvfsp->vfs_snapshot, bp); } else { ufsvfsp->vfs_iotstamp = lbolt; ub.ub_getpages.value.ul++; (void) bdev_strategy(bp); lwp_stat_update(LWP_STAT_INBLK, 1); } ip->i_nextrio = off + ((io_len + PAGESIZE - 1) & PAGEMASK); /* * If the file access is sequential, initiate read ahead * of the next cluster. */ if (seq && ip->i_nextrio < ip->i_size) (void) ufs_getpage_ra(vp, off, seg, addr); err = biowait(bp); pageio_done(bp); if (err) { pvn_read_done(pp, B_ERROR); return (err); } } pvn_plist_init(pp, pl, plsz, off, io_len, rw); return (0); } /* * Read ahead a cluster from the disk. Returns the length in bytes. */ static int ufs_getpage_ra(struct vnode *vp, u_offset_t off, struct seg *seg, caddr_t addr) { struct inode *ip = VTOI(vp); page_t *pp; u_offset_t io_off = ip->i_nextrio; ufsvfs_t *ufsvfsp; caddr_t addr2 = addr + (io_off - off); struct buf *bp; daddr_t bn; size_t io_len; int err; int contig; int xlen; int bsize = ip->i_fs->fs_bsize; /* * If the directio advisory is in effect on this file, * then do not do buffered read ahead. Read ahead makes * it more difficult on threads using directio as they * will be forced to flush the pages from this vnode. */ if ((ufsvfsp = ip->i_ufsvfs) == NULL) return (0); if (ip->i_flag & IDIRECTIO || ufsvfsp->vfs_forcedirectio) return (0); /* * Is this test needed? */ if (addr2 >= seg->s_base + seg->s_size) return (0); contig = 0; err = bmap_read(ip, io_off, &bn, &contig); /* * If its a UFS_HOLE or a fallocated block, do not perform * any read ahead's since there probably is nothing to read ahead */ if (err || bn == UFS_HOLE || ISFALLOCBLK(ip, bn)) return (0); /* * Limit the transfer size to bsize if this is the 2nd block. */ if (io_off == (u_offset_t)bsize) contig = MIN(contig, bsize); if ((pp = pvn_read_kluster(vp, io_off, seg, addr2, &io_off, &io_len, io_off, contig, 1)) == NULL) return (0); /* * Zero part of page which we are not going to read from disk */ if ((xlen = (io_len & PAGEOFFSET)) > 0) pagezero(pp->p_prev, xlen, PAGESIZE - xlen); ip->i_nextrio = (io_off + io_len + PAGESIZE - 1) & PAGEMASK; bp = pageio_setup(pp, io_len, ip->i_devvp, B_READ | B_ASYNC); bp->b_edev = ip->i_dev; bp->b_dev = cmpdev(ip->i_dev); bp->b_blkno = bn; bp->b_un.b_addr = (caddr_t)0; bp->b_file = ip->i_vnode; bp->b_offset = off; if (ufsvfsp->vfs_log) { lufs_read_strategy(ufsvfsp->vfs_log, bp); } else if (ufsvfsp->vfs_snapshot) { fssnap_strategy(&ufsvfsp->vfs_snapshot, bp); } else { ufsvfsp->vfs_iotstamp = lbolt; ub.ub_getras.value.ul++; (void) bdev_strategy(bp); lwp_stat_update(LWP_STAT_INBLK, 1); } return (io_len); } int ufs_delay = 1; /* * Flags are composed of {B_INVAL, B_FREE, B_DONTNEED, B_FORCE, B_ASYNC} * * LMXXX - the inode really ought to contain a pointer to one of these * async args. Stuff gunk in there and just hand the whole mess off. * This would replace i_delaylen, i_delayoff. */ /*ARGSUSED*/ static int ufs_putpage(struct vnode *vp, offset_t off, size_t len, int flags, struct cred *cr) { struct inode *ip = VTOI(vp); int err = 0; if (vp->v_count == 0) { return (ufs_fault(vp, "ufs_putpage: bad v_count == 0")); } /* * XXX - Why should this check be made here? */ if (vp->v_flag & VNOMAP) { err = ENOSYS; goto errout; } if (ip->i_ufsvfs == NULL) { err = EIO; goto errout; } if (flags & B_ASYNC) { if (ufs_delay && len && (flags & ~(B_ASYNC|B_DONTNEED|B_FREE)) == 0) { mutex_enter(&ip->i_tlock); /* * If nobody stalled, start a new cluster. */ if (ip->i_delaylen == 0) { ip->i_delayoff = off; ip->i_delaylen = len; mutex_exit(&ip->i_tlock); goto errout; } /* * If we have a full cluster or they are not contig, * then push last cluster and start over. */ if (ip->i_delaylen >= CLUSTSZ(ip) || ip->i_delayoff + ip->i_delaylen != off) { u_offset_t doff; size_t dlen; doff = ip->i_delayoff; dlen = ip->i_delaylen; ip->i_delayoff = off; ip->i_delaylen = len; mutex_exit(&ip->i_tlock); err = ufs_putpages(vp, doff, dlen, flags, cr); /* LMXXX - flags are new val, not old */ goto errout; } /* * There is something there, it's not full, and * it is contig. */ ip->i_delaylen += len; mutex_exit(&ip->i_tlock); goto errout; } /* * Must have weird flags or we are not clustering. */ } err = ufs_putpages(vp, off, len, flags, cr); errout: return (err); } /* * If len == 0, do from off to EOF. * * The normal cases should be len == 0 & off == 0 (entire vp list), * len == MAXBSIZE (from segmap_release actions), and len == PAGESIZE * (from pageout). */ /*ARGSUSED*/ static int ufs_putpages( struct vnode *vp, offset_t off, size_t len, int flags, struct cred *cr) { u_offset_t io_off; u_offset_t eoff; struct inode *ip = VTOI(vp); page_t *pp; size_t io_len; int err = 0; int dolock; if (vp->v_count == 0) return (ufs_fault(vp, "ufs_putpages: v_count == 0")); /* * Acquire the readers/write inode lock before locking * any pages in this inode. * The inode lock is held during i/o. */ if (len == 0) { mutex_enter(&ip->i_tlock); ip->i_delayoff = ip->i_delaylen = 0; mutex_exit(&ip->i_tlock); } dolock = (rw_owner(&ip->i_contents) != curthread); if (dolock) { /* * Must synchronize this thread and any possible thread * operating in the window of vulnerability in wrip(). * It is dangerous to allow both a thread doing a putpage * and a thread writing, so serialize them. The exception * is when the thread in wrip() does something which causes * a putpage operation. Then, the thread must be allowed * to continue. It may encounter a bmap_read problem in * ufs_putapage, but that is handled in ufs_putapage. * Allow async writers to proceed, we don't want to block * the pageout daemon. */ if (ip->i_writer == curthread) rw_enter(&ip->i_contents, RW_READER); else { for (;;) { rw_enter(&ip->i_contents, RW_READER); mutex_enter(&ip->i_tlock); /* * If there is no thread in the critical * section of wrip(), then proceed. * Otherwise, wait until there isn't one. */ if (ip->i_writer == NULL) { mutex_exit(&ip->i_tlock); break; } rw_exit(&ip->i_contents); /* * Bounce async writers when we have a writer * working on this file so we don't deadlock * the pageout daemon. */ if (flags & B_ASYNC) { mutex_exit(&ip->i_tlock); return (0); } cv_wait(&ip->i_wrcv, &ip->i_tlock); mutex_exit(&ip->i_tlock); } } } if (!vn_has_cached_data(vp)) { if (dolock) rw_exit(&ip->i_contents); return (0); } if (len == 0) { /* * Search the entire vp list for pages >= off. */ err = pvn_vplist_dirty(vp, (u_offset_t)off, ufs_putapage, flags, cr); } else { /* * Loop over all offsets in the range looking for * pages to deal with. */ if ((eoff = blkroundup(ip->i_fs, ip->i_size)) != 0) eoff = MIN(off + len, eoff); else eoff = off + len; for (io_off = off; io_off < eoff; io_off += io_len) { /* * If we are not invalidating, synchronously * freeing or writing pages, use the routine * page_lookup_nowait() to prevent reclaiming * them from the free list. */ if ((flags & B_INVAL) || ((flags & B_ASYNC) == 0)) { pp = page_lookup(vp, io_off, (flags & (B_INVAL | B_FREE)) ? SE_EXCL : SE_SHARED); } else { pp = page_lookup_nowait(vp, io_off, (flags & B_FREE) ? SE_EXCL : SE_SHARED); } if (pp == NULL || pvn_getdirty(pp, flags) == 0) io_len = PAGESIZE; else { u_offset_t *io_offp = &io_off; err = ufs_putapage(vp, pp, io_offp, &io_len, flags, cr); if (err != 0) break; /* * "io_off" and "io_len" are returned as * the range of pages we actually wrote. * This allows us to skip ahead more quickly * since several pages may've been dealt * with by this iteration of the loop. */ } } } if (err == 0 && off == 0 && (len == 0 || len >= ip->i_size)) { /* * We have just sync'ed back all the pages on * the inode, turn off the IMODTIME flag. */ mutex_enter(&ip->i_tlock); ip->i_flag &= ~IMODTIME; mutex_exit(&ip->i_tlock); } if (dolock) rw_exit(&ip->i_contents); return (err); } static void ufs_iodone(buf_t *bp) { struct inode *ip; ASSERT((bp->b_pages->p_vnode != NULL) && !(bp->b_flags & B_READ)); bp->b_iodone = NULL; ip = VTOI(bp->b_pages->p_vnode); mutex_enter(&ip->i_tlock); if (ip->i_writes >= ufs_LW) { if ((ip->i_writes -= bp->b_bcount) <= ufs_LW) if (ufs_WRITES) cv_broadcast(&ip->i_wrcv); /* wake all up */ } else { ip->i_writes -= bp->b_bcount; } mutex_exit(&ip->i_tlock); iodone(bp); } /* * Write out a single page, possibly klustering adjacent * dirty pages. The inode lock must be held. * * LMXXX - bsize < pagesize not done. */ /*ARGSUSED*/ int ufs_putapage( struct vnode *vp, page_t *pp, u_offset_t *offp, size_t *lenp, /* return values */ int flags, struct cred *cr) { u_offset_t io_off; u_offset_t off; struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp = ip->i_ufsvfs; struct fs *fs; struct buf *bp; size_t io_len; daddr_t bn; int err; int contig; ASSERT(RW_LOCK_HELD(&ip->i_contents)); if (ufsvfsp == NULL) { err = EIO; goto out_trace; } fs = ip->i_fs; ASSERT(fs->fs_ronly == 0); /* * If the modified time on the inode has not already been * set elsewhere (e.g. for write/setattr) we set the time now. * This gives us approximate modified times for mmap'ed files * which are modified via stores in the user address space. */ if ((ip->i_flag & IMODTIME) == 0) { mutex_enter(&ip->i_tlock); ip->i_flag |= IUPD; ip->i_seq++; ITIMES_NOLOCK(ip); mutex_exit(&ip->i_tlock); } /* * Align the request to a block boundry (for old file systems), * and go ask bmap() how contiguous things are for this file. */ off = pp->p_offset & (offset_t)fs->fs_bmask; /* block align it */ contig = 0; err = bmap_read(ip, off, &bn, &contig); if (err) goto out; if (bn == UFS_HOLE) { /* putpage never allocates */ /* * logging device is in error mode; simply return EIO */ if (TRANS_ISERROR(ufsvfsp)) { err = EIO; goto out; } /* * Oops, the thread in the window in wrip() did some * sort of operation which caused a putpage in the bad * range. In this case, just return an error which will * cause the software modified bit on the page to set * and the page will get written out again later. */ if (ip->i_writer == curthread) { err = EIO; goto out; } /* * If the pager is trying to push a page in the bad range * just tell him to try again later when things are better. */ if (flags & B_ASYNC) { err = EAGAIN; goto out; } err = ufs_fault(ITOV(ip), "ufs_putapage: bn == UFS_HOLE"); goto out; } /* * If it is an fallocate'd block, reverse the negativity since * we are now writing to it */ if (ISFALLOCBLK(ip, bn)) { err = bmap_set_bn(vp, off, dbtofsb(fs, -bn)); if (err) goto out; bn = -bn; } /* * Take the length (of contiguous bytes) passed back from bmap() * and _try_ and get a set of pages covering that extent. */ pp = pvn_write_kluster(vp, pp, &io_off, &io_len, off, contig, flags); /* * May have run out of memory and not clustered backwards. * off p_offset * [ pp - 1 ][ pp ] * [ block ] * We told bmap off, so we have to adjust the bn accordingly. */ if (io_off > off) { bn += btod(io_off - off); contig -= (io_off - off); } /* * bmap was carefull to tell us the right size so use that. * There might be unallocated frags at the end. * LMXXX - bzero the end of the page? We must be writing after EOF. */ if (io_len > contig) { ASSERT(io_len - contig < fs->fs_bsize); io_len -= (io_len - contig); } /* * Handle the case where we are writing the last page after EOF. * * XXX - just a patch for i-mt3. */ if (io_len == 0) { ASSERT(pp->p_offset >= (u_offset_t)(roundup(ip->i_size, PAGESIZE))); io_len = PAGESIZE; } bp = pageio_setup(pp, io_len, ip->i_devvp, B_WRITE | flags); ULOCKFS_SET_MOD(ITOUL(ip)); bp->b_edev = ip->i_dev; bp->b_dev = cmpdev(ip->i_dev); bp->b_blkno = bn; bp->b_un.b_addr = (caddr_t)0; bp->b_file = ip->i_vnode; if (TRANS_ISTRANS(ufsvfsp)) { if ((ip->i_mode & IFMT) == IFSHAD) { TRANS_BUF(ufsvfsp, 0, io_len, bp, DT_SHAD); } else if (ufsvfsp->vfs_qinod == ip) { TRANS_DELTA(ufsvfsp, ldbtob(bn), bp->b_bcount, DT_QR, 0, 0); } } /* write throttle */ ASSERT(bp->b_iodone == NULL); bp->b_iodone = (int (*)())ufs_iodone; mutex_enter(&ip->i_tlock); ip->i_writes += bp->b_bcount; mutex_exit(&ip->i_tlock); if (bp->b_flags & B_ASYNC) { if (ufsvfsp->vfs_log) { lufs_write_strategy(ufsvfsp->vfs_log, bp); } else if (ufsvfsp->vfs_snapshot) { fssnap_strategy(&ufsvfsp->vfs_snapshot, bp); } else { ufsvfsp->vfs_iotstamp = lbolt; ub.ub_putasyncs.value.ul++; (void) bdev_strategy(bp); lwp_stat_update(LWP_STAT_OUBLK, 1); } } else { if (ufsvfsp->vfs_log) { lufs_write_strategy(ufsvfsp->vfs_log, bp); } else if (ufsvfsp->vfs_snapshot) { fssnap_strategy(&ufsvfsp->vfs_snapshot, bp); } else { ufsvfsp->vfs_iotstamp = lbolt; ub.ub_putsyncs.value.ul++; (void) bdev_strategy(bp); lwp_stat_update(LWP_STAT_OUBLK, 1); } err = biowait(bp); pageio_done(bp); pvn_write_done(pp, ((err) ? B_ERROR : 0) | B_WRITE | flags); } pp = NULL; out: if (err != 0 && pp != NULL) pvn_write_done(pp, B_ERROR | B_WRITE | flags); if (offp) *offp = io_off; if (lenp) *lenp = io_len; out_trace: return (err); } /* ARGSUSED */ static int ufs_map(struct vnode *vp, offset_t off, struct as *as, caddr_t *addrp, size_t len, uchar_t prot, uchar_t maxprot, uint_t flags, struct cred *cr) { struct segvn_crargs vn_a; struct ufsvfs *ufsvfsp = VTOI(vp)->i_ufsvfs; struct ulockfs *ulp; int error; retry_map: error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_MAP_MASK); if (error) goto out; if (vp->v_flag & VNOMAP) { error = ENOSYS; goto unlock; } if (off < (offset_t)0 || (offset_t)(off + len) < (offset_t)0) { error = ENXIO; goto unlock; } if (vp->v_type != VREG) { error = ENODEV; goto unlock; } /* * If file is being locked, disallow mapping. */ if (vn_has_mandatory_locks(vp, VTOI(vp)->i_mode)) { error = EAGAIN; goto unlock; } as_rangelock(as); if ((flags & MAP_FIXED) == 0) { map_addr(addrp, len, off, 1, flags); if (*addrp == NULL) { as_rangeunlock(as); error = ENOMEM; goto unlock; } } else { /* * User specified address - blow away any previous mappings */ (void) as_unmap(as, *addrp, len); } vn_a.vp = vp; vn_a.offset = (u_offset_t)off; vn_a.type = flags & MAP_TYPE; vn_a.prot = prot; vn_a.maxprot = maxprot; vn_a.cred = cr; vn_a.amp = NULL; vn_a.flags = flags & ~MAP_TYPE; vn_a.szc = 0; vn_a.lgrp_mem_policy_flags = 0; retry_lock: if (!AS_LOCK_TRYENTER(ias, &as->a_lock, RW_WRITER)) { /* * We didn't get the lock. Check if the SLOCK is set in the * ufsvfs. If yes, we might be in a deadlock. Safer to give up * and wait for SLOCK to be cleared. */ if (ulp && ULOCKFS_IS_SLOCK(ulp)) { as_rangeunlock(as); ufs_lockfs_end(ulp); goto retry_map; } else { /* * SLOCK isn't set so this is a genuine synchronization * case. Let's try again after giving them a breather. */ delay(RETRY_LOCK_DELAY); goto retry_lock; } } error = as_map_locked(as, *addrp, len, segvn_create, &vn_a); as_rangeunlock(as); unlock: if (ulp) { ufs_lockfs_end(ulp); } out: return (error); } /* ARGSUSED */ static int ufs_addmap(struct vnode *vp, offset_t off, struct as *as, caddr_t addr, size_t len, uchar_t prot, uchar_t maxprot, uint_t flags, struct cred *cr) { struct inode *ip = VTOI(vp); if (vp->v_flag & VNOMAP) { return (ENOSYS); } mutex_enter(&ip->i_tlock); ip->i_mapcnt += btopr(len); mutex_exit(&ip->i_tlock); return (0); } /*ARGSUSED*/ static int ufs_delmap(struct vnode *vp, offset_t off, struct as *as, caddr_t addr, size_t len, uint_t prot, uint_t maxprot, uint_t flags, struct cred *cr) { struct inode *ip = VTOI(vp); if (vp->v_flag & VNOMAP) { return (ENOSYS); } mutex_enter(&ip->i_tlock); ip->i_mapcnt -= btopr(len); /* Count released mappings */ ASSERT(ip->i_mapcnt >= 0); mutex_exit(&ip->i_tlock); return (0); } /* * Return the answer requested to poll() for non-device files */ struct pollhead ufs_pollhd; /* ARGSUSED */ int ufs_poll(vnode_t *vp, short ev, int any, short *revp, struct pollhead **phpp) { struct ufsvfs *ufsvfsp; *revp = 0; ufsvfsp = VTOI(vp)->i_ufsvfs; if (!ufsvfsp) { *revp = POLLHUP; goto out; } if (ULOCKFS_IS_HLOCK(&ufsvfsp->vfs_ulockfs) || ULOCKFS_IS_ELOCK(&ufsvfsp->vfs_ulockfs)) { *revp |= POLLERR; } else { if ((ev & POLLOUT) && !ufsvfsp->vfs_fs->fs_ronly && !ULOCKFS_IS_WLOCK(&ufsvfsp->vfs_ulockfs)) *revp |= POLLOUT; if ((ev & POLLWRBAND) && !ufsvfsp->vfs_fs->fs_ronly && !ULOCKFS_IS_WLOCK(&ufsvfsp->vfs_ulockfs)) *revp |= POLLWRBAND; if (ev & POLLIN) *revp |= POLLIN; if (ev & POLLRDNORM) *revp |= POLLRDNORM; if (ev & POLLRDBAND) *revp |= POLLRDBAND; } if ((ev & POLLPRI) && (*revp & (POLLERR|POLLHUP))) *revp |= POLLPRI; out: *phpp = !any && !*revp ? &ufs_pollhd : (struct pollhead *)NULL; return (0); } /* ARGSUSED */ static int ufs_l_pathconf(struct vnode *vp, int cmd, ulong_t *valp, struct cred *cr) { struct ufsvfs *ufsvfsp = VTOI(vp)->i_ufsvfs; struct ulockfs *ulp = NULL; struct inode *sip = NULL; int error; struct inode *ip = VTOI(vp); int issync; error = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_PATHCONF_MASK); if (error) return (error); switch (cmd) { /* * Have to handle _PC_NAME_MAX here, because the normal way * [fs_pathconf() -> VOP_STATVFS() -> ufs_statvfs()] * results in a lock ordering reversal between * ufs_lockfs_{begin,end}() and * ufs_thread_{suspend,continue}(). * * Keep in sync with ufs_statvfs(). */ case _PC_NAME_MAX: *valp = MAXNAMLEN; break; case _PC_FILESIZEBITS: if (ufsvfsp->vfs_lfflags & UFS_LARGEFILES) *valp = UFS_FILESIZE_BITS; else *valp = 32; break; case _PC_XATTR_EXISTS: if (vp->v_vfsp->vfs_flag & VFS_XATTR) { error = ufs_xattr_getattrdir(vp, &sip, LOOKUP_XATTR, cr); if (error == 0 && sip != NULL) { /* Start transaction */ if (ulp) { TRANS_BEGIN_CSYNC(ufsvfsp, issync, TOP_RMDIR, TOP_RMDIR_SIZE); } /* * Is directory empty */ rw_enter(&sip->i_rwlock, RW_WRITER); rw_enter(&sip->i_contents, RW_WRITER); if (ufs_xattrdirempty(sip, sip->i_number, CRED())) { rw_enter(&ip->i_contents, RW_WRITER); ufs_unhook_shadow(ip, sip); rw_exit(&ip->i_contents); *valp = 0; } else *valp = 1; rw_exit(&sip->i_contents); rw_exit(&sip->i_rwlock); if (ulp) { TRANS_END_CSYNC(ufsvfsp, error, issync, TOP_RMDIR, TOP_RMDIR_SIZE); } VN_RELE(ITOV(sip)); } else if (error == ENOENT) { *valp = 0; error = 0; } } else { error = fs_pathconf(vp, cmd, valp, cr); } break; case _PC_ACL_ENABLED: *valp = _ACL_ACLENT_ENABLED; break; case _PC_MIN_HOLE_SIZE: *valp = (ulong_t)ip->i_fs->fs_bsize; break; default: error = fs_pathconf(vp, cmd, valp, cr); } if (ulp != NULL) { ufs_lockfs_end(ulp); } return (error); } int ufs_pageio_writes, ufs_pageio_reads; /*ARGSUSED*/ static int ufs_pageio(struct vnode *vp, page_t *pp, u_offset_t io_off, size_t io_len, int flags, struct cred *cr) { struct inode *ip = VTOI(vp); struct ufsvfs *ufsvfsp; page_t *npp = NULL, *opp = NULL, *cpp = pp; struct buf *bp; daddr_t bn; size_t done_len = 0, cur_len = 0; int err = 0; int contig = 0; int dolock; int vmpss = 0; struct ulockfs *ulp; if ((flags & B_READ) && pp != NULL && pp->p_vnode == vp && vp->v_mpssdata != NULL) { vmpss = 1; } dolock = (rw_owner(&ip->i_contents) != curthread); /* * We need a better check. Ideally, we would use another * vnodeops so that hlocked and forcibly unmounted file * systems would return EIO where appropriate and w/o the * need for these checks. */ if ((ufsvfsp = ip->i_ufsvfs) == NULL) return (EIO); /* * For vmpss (pp can be NULL) case respect the quiesce protocol. * ul_lock must be taken before locking pages so we can't use it here * if pp is non NULL because segvn already locked pages * SE_EXCL. Instead we rely on the fact that a forced umount or * applying a filesystem lock via ufs_fiolfs() will block in the * implicit call to ufs_flush() until we unlock the pages after the * return to segvn. Other ufs_quiesce() callers keep ufs_quiesce_pend * above 0 until they are done. We have to be careful not to increment * ul_vnops_cnt here after forceful unmount hlocks the file system. * * If pp is NULL use ul_lock to make sure we don't increment * ul_vnops_cnt after forceful unmount hlocks the file system. */ if (vmpss || pp == NULL) { ulp = &ufsvfsp->vfs_ulockfs; if (pp == NULL) mutex_enter(&ulp->ul_lock); if (ulp->ul_fs_lock & ULOCKFS_GETREAD_MASK) { if (pp == NULL) { mutex_exit(&ulp->ul_lock); } return (vmpss ? EIO : EINVAL); } atomic_add_long(&ulp->ul_vnops_cnt, 1); if (pp == NULL) mutex_exit(&ulp->ul_lock); if (ufs_quiesce_pend) { if (!atomic_add_long_nv(&ulp->ul_vnops_cnt, -1)) cv_broadcast(&ulp->ul_cv); return (vmpss ? EIO : EINVAL); } } if (dolock) { /* * segvn may call VOP_PAGEIO() instead of VOP_GETPAGE() to * handle a fault against a segment that maps vnode pages with * large mappings. Segvn creates pages and holds them locked * SE_EXCL during VOP_PAGEIO() call. In this case we have to * use rw_tryenter() to avoid a potential deadlock since in * lock order i_contents needs to be taken first. * Segvn will retry via VOP_GETPAGE() if VOP_PAGEIO() fails. */ if (!vmpss) { rw_enter(&ip->i_contents, RW_READER); } else if (!rw_tryenter(&ip->i_contents, RW_READER)) { if (!atomic_add_long_nv(&ulp->ul_vnops_cnt, -1)) cv_broadcast(&ulp->ul_cv); return (EDEADLK); } } /* * Return an error to segvn because the pagefault request is beyond * PAGESIZE rounded EOF. */ if (vmpss && btopr(io_off + io_len) > btopr(ip->i_size)) { if (dolock) rw_exit(&ip->i_contents); if (!atomic_add_long_nv(&ulp->ul_vnops_cnt, -1)) cv_broadcast(&ulp->ul_cv); return (EFAULT); } if (pp == NULL) { if (bmap_has_holes(ip)) { err = ENOSYS; } else { err = EINVAL; } if (dolock) rw_exit(&ip->i_contents); if (!atomic_add_long_nv(&ulp->ul_vnops_cnt, -1)) cv_broadcast(&ulp->ul_cv); return (err); } /* * Break the io request into chunks, one for each contiguous * stretch of disk blocks in the target file. */ while (done_len < io_len) { ASSERT(cpp); contig = 0; if (err = bmap_read(ip, (u_offset_t)(io_off + done_len), &bn, &contig)) break; if (bn == UFS_HOLE) { /* No holey swapfiles */ if (vmpss) { err = EFAULT; break; } err = ufs_fault(ITOV(ip), "ufs_pageio: bn == UFS_HOLE"); break; } cur_len = MIN(io_len - done_len, contig); /* * Zero out a page beyond EOF, when the last block of * a file is a UFS fragment so that ufs_pageio() can be used * instead of ufs_getpage() to handle faults against * segvn segments that use large pages. */ page_list_break(&cpp, &npp, btopr(cur_len)); if ((flags & B_READ) && (cur_len & PAGEOFFSET)) { size_t xlen = cur_len & PAGEOFFSET; pagezero(cpp->p_prev, xlen, PAGESIZE - xlen); } bp = pageio_setup(cpp, cur_len, ip->i_devvp, flags); ASSERT(bp != NULL); bp->b_edev = ip->i_dev; bp->b_dev = cmpdev(ip->i_dev); bp->b_blkno = bn; bp->b_un.b_addr = (caddr_t)0; bp->b_file = ip->i_vnode; ufsvfsp->vfs_iotstamp = lbolt; ub.ub_pageios.value.ul++; if (ufsvfsp->vfs_snapshot) fssnap_strategy(&(ufsvfsp->vfs_snapshot), bp); else (void) bdev_strategy(bp); if (flags & B_READ) ufs_pageio_reads++; else ufs_pageio_writes++; if (flags & B_READ) lwp_stat_update(LWP_STAT_INBLK, 1); else lwp_stat_update(LWP_STAT_OUBLK, 1); /* * If the request is not B_ASYNC, wait for i/o to complete * and re-assemble the page list to return to the caller. * If it is B_ASYNC we leave the page list in pieces and * cleanup() will dispose of them. */ if ((flags & B_ASYNC) == 0) { err = biowait(bp); pageio_done(bp); if (err) break; page_list_concat(&opp, &cpp); } cpp = npp; npp = NULL; if (flags & B_READ) cur_len = P2ROUNDUP_TYPED(cur_len, PAGESIZE, size_t); done_len += cur_len; } ASSERT(err || (cpp == NULL && npp == NULL && done_len == io_len)); if (err) { if (flags & B_ASYNC) { /* Cleanup unprocessed parts of list */ page_list_concat(&cpp, &npp); if (flags & B_READ) pvn_read_done(cpp, B_ERROR); else pvn_write_done(cpp, B_ERROR); } else { /* Re-assemble list and let caller clean up */ page_list_concat(&opp, &cpp); page_list_concat(&opp, &npp); } } if (vmpss && !(ip->i_flag & IACC) && !ULOCKFS_IS_NOIACC(ulp) && ufsvfsp->vfs_fs->fs_ronly == 0 && !ufsvfsp->vfs_noatime) { mutex_enter(&ip->i_tlock); ip->i_flag |= IACC; ITIMES_NOLOCK(ip); mutex_exit(&ip->i_tlock); } if (dolock) rw_exit(&ip->i_contents); if (vmpss && !atomic_add_long_nv(&ulp->ul_vnops_cnt, -1)) cv_broadcast(&ulp->ul_cv); return (err); } /* * Called when the kernel is in a frozen state to dump data * directly to the device. It uses a private dump data structure, * set up by dump_ctl, to locate the correct disk block to which to dump. */ static int ufs_dump(vnode_t *vp, caddr_t addr, int ldbn, int dblks) { u_offset_t file_size; struct inode *ip = VTOI(vp); struct fs *fs = ip->i_fs; daddr_t dbn, lfsbn; int disk_blks = fs->fs_bsize >> DEV_BSHIFT; int error = 0; int ndbs, nfsbs; /* * forced unmount case */ if (ip->i_ufsvfs == NULL) return (EIO); /* * Validate the inode that it has not been modified since * the dump structure is allocated. */ mutex_enter(&ip->i_tlock); if ((dump_info == NULL) || (dump_info->ip != ip) || (dump_info->time.tv_sec != ip->i_mtime.tv_sec) || (dump_info->time.tv_usec != ip->i_mtime.tv_usec)) { mutex_exit(&ip->i_tlock); return (-1); } mutex_exit(&ip->i_tlock); /* * See that the file has room for this write */ UFS_GET_ISIZE(&file_size, ip); if (ldbtob((offset_t)(ldbn + dblks)) > file_size) return (ENOSPC); /* * Find the physical disk block numbers from the dump * private data structure directly and write out the data * in contiguous block lumps */ while (dblks > 0 && !error) { lfsbn = (daddr_t)lblkno(fs, ldbtob((offset_t)ldbn)); dbn = fsbtodb(fs, dump_info->dblk[lfsbn]) + ldbn % disk_blks; nfsbs = 1; ndbs = disk_blks - ldbn % disk_blks; while (ndbs < dblks && fsbtodb(fs, dump_info->dblk[lfsbn + nfsbs]) == dbn + ndbs) { nfsbs++; ndbs += disk_blks; } if (ndbs > dblks) ndbs = dblks; error = bdev_dump(ip->i_dev, addr, dbn, ndbs); addr += ldbtob((offset_t)ndbs); dblks -= ndbs; ldbn += ndbs; } return (error); } /* * Prepare the file system before and after the dump operation. * * action = DUMP_ALLOC: * Preparation before dump, allocate dump private data structure * to hold all the direct and indirect block info for dump. * * action = DUMP_FREE: * Clean up after dump, deallocate the dump private data structure. * * action = DUMP_SCAN: * Scan dump_info for *blkp DEV_BSIZE blocks of contig fs space; * if found, the starting file-relative DEV_BSIZE lbn is written * to *bklp; that lbn is intended for use with VOP_DUMP() */ static int ufs_dumpctl(vnode_t *vp, int action, int *blkp) { struct inode *ip = VTOI(vp); ufsvfs_t *ufsvfsp = ip->i_ufsvfs; struct fs *fs; daddr32_t *dblk, *storeblk; daddr32_t *nextblk, *endblk; struct buf *bp; int i, entry, entries; int n, ncontig; /* * check for forced unmount */ if (ufsvfsp == NULL) return (EIO); if (action == DUMP_ALLOC) { /* * alloc and record dump_info */ if (dump_info != NULL) return (EINVAL); ASSERT(vp->v_type == VREG); fs = ufsvfsp->vfs_fs; rw_enter(&ip->i_contents, RW_READER); if (bmap_has_holes(ip)) { rw_exit(&ip->i_contents); return (EFAULT); } /* * calculate and allocate space needed according to i_size */ entries = (int)lblkno(fs, blkroundup(fs, ip->i_size)); dump_info = kmem_alloc(sizeof (struct dump) + (entries - 1) * sizeof (daddr32_t), KM_NOSLEEP); if (dump_info == NULL) { rw_exit(&ip->i_contents); return (ENOMEM); } /* Start saving the info */ dump_info->fsbs = entries; dump_info->ip = ip; storeblk = &dump_info->dblk[0]; /* Direct Blocks */ for (entry = 0; entry < NDADDR && entry < entries; entry++) *storeblk++ = ip->i_db[entry]; /* Indirect Blocks */ for (i = 0; i < NIADDR; i++) { int error = 0; bp = UFS_BREAD(ufsvfsp, ip->i_dev, fsbtodb(fs, ip->i_ib[i]), fs->fs_bsize); if (bp->b_flags & B_ERROR) error = EIO; else { dblk = bp->b_un.b_daddr; if ((storeblk = save_dblks(ip, ufsvfsp, storeblk, dblk, i, entries)) == NULL) error = EIO; } brelse(bp); if (error != 0) { kmem_free(dump_info, sizeof (struct dump) + (entries - 1) * sizeof (daddr32_t)); rw_exit(&ip->i_contents); dump_info = NULL; return (error); } } /* and time stamp the information */ mutex_enter(&ip->i_tlock); dump_info->time = ip->i_mtime; mutex_exit(&ip->i_tlock); rw_exit(&ip->i_contents); } else if (action == DUMP_FREE) { /* * free dump_info */ if (dump_info == NULL) return (EINVAL); entries = dump_info->fsbs - 1; kmem_free(dump_info, sizeof (struct dump) + entries * sizeof (daddr32_t)); dump_info = NULL; } else if (action == DUMP_SCAN) { /* * scan dump_info */ if (dump_info == NULL) return (EINVAL); dblk = dump_info->dblk; nextblk = dblk + 1; endblk = dblk + dump_info->fsbs - 1; fs = ufsvfsp->vfs_fs; ncontig = *blkp >> (fs->fs_bshift - DEV_BSHIFT); /* * scan dblk[] entries; contig fs space is found when: * ((current blkno + frags per block) == next blkno) */ n = 0; while (n < ncontig && dblk < endblk) { if ((*dblk + fs->fs_frag) == *nextblk) n++; else n = 0; dblk++; nextblk++; } /* * index is where size bytes of contig space begins; * conversion from index to the file's DEV_BSIZE lbn * is equivalent to: (index * fs_bsize) / DEV_BSIZE */ if (n == ncontig) { i = (dblk - dump_info->dblk) - ncontig; *blkp = i << (fs->fs_bshift - DEV_BSHIFT); } else return (EFAULT); } return (0); } /* * Recursive helper function for ufs_dumpctl(). It follows the indirect file * system blocks until it reaches the the disk block addresses, which are * then stored into the given buffer, storeblk. */ static daddr32_t * save_dblks(struct inode *ip, struct ufsvfs *ufsvfsp, daddr32_t *storeblk, daddr32_t *dblk, int level, int entries) { struct fs *fs = ufsvfsp->vfs_fs; struct buf *bp; int i; if (level == 0) { for (i = 0; i < NINDIR(fs); i++) { if (storeblk - dump_info->dblk >= entries) break; *storeblk++ = dblk[i]; } return (storeblk); } for (i = 0; i < NINDIR(fs); i++) { if (storeblk - dump_info->dblk >= entries) break; bp = UFS_BREAD(ufsvfsp, ip->i_dev, fsbtodb(fs, dblk[i]), fs->fs_bsize); if (bp->b_flags & B_ERROR) { brelse(bp); return (NULL); } storeblk = save_dblks(ip, ufsvfsp, storeblk, bp->b_un.b_daddr, level - 1, entries); brelse(bp); if (storeblk == NULL) return (NULL); } return (storeblk); } /* ARGSUSED */ static int ufs_getsecattr(struct vnode *vp, vsecattr_t *vsap, int flag, struct cred *cr) { struct inode *ip = VTOI(vp); struct ulockfs *ulp; struct ufsvfs *ufsvfsp = ip->i_ufsvfs; ulong_t vsa_mask = vsap->vsa_mask; int err = EINVAL; vsa_mask &= (VSA_ACL | VSA_ACLCNT | VSA_DFACL | VSA_DFACLCNT); /* * Only grab locks if needed - they're not needed to check vsa_mask * or if the mask contains no acl flags. */ if (vsa_mask != 0) { if (err = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_GETATTR_MASK)) return (err); rw_enter(&ip->i_contents, RW_READER); err = ufs_acl_get(ip, vsap, flag, cr); rw_exit(&ip->i_contents); if (ulp) ufs_lockfs_end(ulp); } return (err); } /* ARGSUSED */ static int ufs_setsecattr(struct vnode *vp, vsecattr_t *vsap, int flag, struct cred *cr) { struct inode *ip = VTOI(vp); struct ulockfs *ulp = NULL; struct ufsvfs *ufsvfsp = VTOI(vp)->i_ufsvfs; ulong_t vsa_mask = vsap->vsa_mask; int err; int haverwlock = 1; int trans_size; int donetrans = 0; int retry = 1; ASSERT(RW_LOCK_HELD(&ip->i_rwlock)); /* Abort now if the request is either empty or invalid. */ vsa_mask &= (VSA_ACL | VSA_ACLCNT | VSA_DFACL | VSA_DFACLCNT); if ((vsa_mask == 0) || ((vsap->vsa_aclentp == NULL) && (vsap->vsa_dfaclentp == NULL))) { err = EINVAL; goto out; } /* * Following convention, if this is a directory then we acquire the * inode's i_rwlock after starting a UFS logging transaction; * otherwise, we acquire it beforehand. Since we were called (and * must therefore return) with the lock held, we will have to drop it, * and later reacquire it, if operating on a directory. */ if (vp->v_type == VDIR) { rw_exit(&ip->i_rwlock); haverwlock = 0; } else { /* Upgrade the lock if required. */ if (!rw_write_held(&ip->i_rwlock)) { rw_exit(&ip->i_rwlock); rw_enter(&ip->i_rwlock, RW_WRITER); } } again: ASSERT(!(vp->v_type == VDIR && haverwlock)); if (err = ufs_lockfs_begin(ufsvfsp, &ulp, ULOCKFS_SETATTR_MASK)) { ulp = NULL; retry = 0; goto out; } /* * Check that the file system supports this operation. Note that * ufs_lockfs_begin() will have checked that the file system had * not been forcibly unmounted. */ if (ufsvfsp->vfs_fs->fs_ronly) { err = EROFS; goto out; } if (ufsvfsp->vfs_nosetsec) { err = ENOSYS; goto out; } if (ulp) { TRANS_BEGIN_ASYNC(ufsvfsp, TOP_SETSECATTR, trans_size = TOP_SETSECATTR_SIZE(VTOI(vp))); donetrans = 1; } if (vp->v_type == VDIR) { rw_enter(&ip->i_rwlock, RW_WRITER); haverwlock = 1; } ASSERT(haverwlock); /* Do the actual work. */ rw_enter(&ip->i_contents, RW_WRITER); /* * Suppress out of inodes messages if we will retry. */ if (retry) ip->i_flag |= IQUIET; err = ufs_acl_set(ip, vsap, flag, cr); ip->i_flag &= ~IQUIET; rw_exit(&ip->i_contents); out: if (ulp) { if (donetrans) { /* * top_end_async() can eventually call * top_end_sync(), which can block. We must * therefore observe the lock-ordering protocol * here as well. */ if (vp->v_type == VDIR) { rw_exit(&ip->i_rwlock); haverwlock = 0; } TRANS_END_ASYNC(ufsvfsp, TOP_SETSECATTR, trans_size); } ufs_lockfs_end(ulp); } /* * If no inodes available, try scaring a logically- * free one out of the delete queue to someplace * that we can find it. */ if ((err == ENOSPC) && retry && TRANS_ISTRANS(ufsvfsp)) { ufs_delete_drain_wait(ufsvfsp, 1); retry = 0; if (vp->v_type == VDIR && haverwlock) { rw_exit(&ip->i_rwlock); haverwlock = 0; } goto again; } /* * If we need to reacquire the lock then it is safe to do so * as a reader. This is because ufs_rwunlock(), which will be * called by our caller after we return, does not differentiate * between shared and exclusive locks. */ if (!haverwlock) { ASSERT(vp->v_type == VDIR); rw_enter(&ip->i_rwlock, RW_READER); } return (err); }