6253 F_GETLK doesn't always return lock owner

[illumos-gate.git] / usr / src / uts / common / sys / fs / ufs_inode.h

/*
 * 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 (c) 1983, 2010, Oracle and/or its affiliates. All rights reserved.
 */

/*      Copyright (c) 1984, 1986, 1987, 1988, 1989 AT&T */
/*        All Rights Reserved   */

/*
 * University Copyright- Copyright (c) 1982, 1986, 1988
 * The Regents of the University of California
 * All Rights Reserved
 *
 * University Acknowledgment- Portions of this document are derived from
 * software developed by the University of California, Berkeley, and its
 * contributors.
 */

#ifndef _SYS_FS_UFS_INODE_H
#define _SYS_FS_UFS_INODE_H

#include <sys/isa_defs.h>
#include <sys/fbuf.h>
#include <sys/fdbuffer.h>
#include <sys/fcntl.h>
#include <sys/uio.h>
#include <sys/t_lock.h>
#include <sys/thread.h>
#include <sys/cred.h>
#include <sys/time.h>
#include <sys/types32.h>
#include <sys/fs/ufs_fs.h>
#include <sys/fs/ufs_lockfs.h>
#include <sys/fs/ufs_trans.h>
#include <sys/kstat.h>
#include <sys/fs/ufs_acl.h>
#include <sys/fs/ufs_panic.h>
#include <sys/dnlc.h>

#ifdef _KERNEL
#include <sys/vfs_opreg.h>
#endif

#ifdef  __cplusplus
extern "C" {
#endif

/*
 * The I node is the focus of all local file activity in UNIX.
 * There is a unique inode allocated for each active file,
 * each current directory, each mounted-on file, each mapping,
 * and the root.  An inode is `named' by its dev/inumber pair.
 * Data in icommon is read in from permanent inode on volume.
 *
 * Each inode has 5 locks associated with it:
 *      i_rwlock:       Serializes ufs_write and ufs_setattr request
 *                      and allows ufs_read requests to proceed in parallel.
 *                      Serializes reads/updates to directories.
 *      vfs_dqrwlock:   Manages quota sub-system quiescence.  See below.
 *      i_contents:     Protects almost all of the fields in the inode
 *                      except for those listed below. When held
 *                      in writer mode also protects those fields
 *                      listed under i_tlock.
 *      i_tlock:        When i_tlock is held with the i_contents reader
 *                      lock the i_atime, i_mtime, i_ctime,
 *                      i_delayoff, i_delaylen, i_nextrio, i_writes, i_flag
 *                      i_seq, i_writer & i_mapcnt fields are protected.
 *                      For more i_flag locking info see below.
 *      ih_lock:        Protects inode hash chain buckets
 *      ifree_lock:     Protects inode freelist
 *
 * Lock ordering:
 *      i_rwlock > i_contents > i_tlock
 *      i_rwlock > vfs_dqrwlock > i_contents(writer) > i_tlock
 *      i_contents > i_tlock
 *      vfs_dqrwlock > i_contents(writer) > i_tlock
 *      ih_lock > i_contents > i_tlock
 *
 * Making major changes to quota sub-system state, while the file
 * system is mounted required the addition of another lock.  The
 * primary lock in the quota sub-system is vfs_dqrwlock in the ufsvfs
 * structure.  This lock is used to manage quota sub-system quiescence
 * for a particular file system. Major changes to quota sub-system
 * state (disabling quotas, enabling quotas, and setting new quota
 * limits) all require the file system to be quiescent and grabbing
 * vfs_dqrwlock as writer accomplishes this.  On the other hand,
 * grabbing vfs_dqrwlock as reader makes the quota sub-system
 * non-quiescent and lets the quota sub-system know that now is not a
 * good time to change major quota sub-system state.  Typically
 * vfs_dqrwlock is grabbed for reading before i_contents is grabbed for
 * writing.  However, there are cases where vfs_dqrwlock is grabbed for
 * reading without a corresponding i_contents write grab because there
 * is no relevant inode.  There are also cases where i_contents is
 * grabbed for writing when a vfs_dqrwlock read grab is not needed
 * because the inode changes do not affect quotas.
 *
 * Unfortunately, performance considerations have required that we be more
 * intelligent about using i_tlock when updating i_flag.  Ideally, we would
 * have simply separated out several of the bits in i_flag into their own
 * ints to avoid problems.  But, instead, we have implemented the following
 * rules:
 *
 *      o You can update any i_flag field while holding the writer-contents,
 *        or by holding the reader-contents AND holding i_tlock.
 *        You can only call ITIMES_NOLOCK while holding the writer-contents,
 *        or by holding the reader-contents AND holding i_tlock.
 *
 *      o For a directory, holding the reader-rw_lock is sufficient for setting
 *        IACC.
 *
 *      o Races with IREF are avoided by holding the reader contents lock
 *        and by holding i_tlock in ufs_rmidle, ufs_putapage, and ufs_getpage.
 *        And by holding the writer-contents in ufs_iinactive.
 *
 *      o The callers are no longer required to handle the calls to ITIMES
 *        and ITIMES_NOLOCK.  The functions that set the i_flag bits are
 *        responsible for managing those calls.  The exceptions are the
 *        bmap routines.
 *
 * SVR4 Extended Fundamental Type (EFT) support:
 *      The inode structure has been enhanced to support
 *      32-bit user-id, 32-bit group-id, and 32-bit device number.
 *      Standard SVR4 ufs also supports 32-bit mode field.  For the reason
 *      of backward compatibility with the previous ufs disk format,
 *      32-bit mode field is not supported.
 *
 *      The current inode structure is 100% backward compatible with
 *      the previous inode structure if no user-id or group-id exceeds
 *      USHRT_MAX, and no major or minor number of a device number
 *      stored in an inode exceeds 255.
 *
 * Rules for managing i_seq:
 *      o i_seq is locked under the same rules as i_flag
 *      o The i_ctime or i_mtime MUST never change without increasing
 *        the value of i_seq.
 *      o You may increase the value of i_seq without the timestamps
 *        changing, this may decrease the callers performance but will
 *        be functionally correct.
 *      o The common case is when IUPD or ICHG is set, increase i_seq
 *        and immediately call ITIMES* or ufs_iupdat to create a new timestamp.
 *      o A less common case is the setting of IUPD or ICHG and while still
 *        holding the correct lock defer the timestamp and i_seq update
 *        until later, but it must still be done before the lock is released.
 *        bmap_write is an example of this, where the caller does the update.
 *      o If multiple changes are being made with the timestamps being
 *        updated only at the end, a single increase of i_seq is allowed.
 *      o If changes are made with IUPD or ICHG being set, but
 *        the controlling lock is being dropped before the timestamp is
 *        updated, there is a risk that another thread will also change
 *        the file, update i_flag, and push just one timestamp update.
 *        There is also the risk that another thread calls ITIMES or
 *        ufs_iupdat without setting IUPD|ICHG and thus not changing i_seq,
 *        this will cause ufs_imark to change the timestamps without changing
 *        i_seq. If the controlling lock is dropped, ISEQ must be set to
 *        force i_seq to be increased on next ufs_imark, but i_seq MUST still
 *        be increased by the original setting thread before its deferred
 *        call to ITIMES to insure it is increased the correct number of times.
 */

#define UID_LONG  (o_uid_t)65535
                                /* flag value to indicate uid is 32-bit long */
#define GID_LONG  (o_uid_t)65535
                                /* flag value to indicate gid is 32-bit long */

#define NDADDR  12              /* direct addresses in inode */
#define NIADDR  3               /* indirect addresses in inode */
#define FSL_SIZE (NDADDR + NIADDR - 1) * sizeof (daddr32_t)
                                /* max fast symbolic name length is 56 */

#define i_fs    i_ufsvfs->vfs_bufp->b_un.b_fs
#define i_vfs   i_vnode->v_vfsp

struct  icommon {
        o_mode_t ic_smode;      /*  0: mode and type of file */
        short   ic_nlink;       /*  2: number of links to file */
        o_uid_t ic_suid;        /*  4: owner's user id */
        o_gid_t ic_sgid;        /*  6: owner's group id */
        u_offset_t ic_lsize;    /*  8: number of bytes in file */
#ifdef _KERNEL
        struct timeval32 ic_atime;      /* 16: time last accessed */
        struct timeval32 ic_mtime;      /* 24: time last modified */
        struct timeval32 ic_ctime;      /* 32: last time inode changed */
#else
        time32_t ic_atime;      /* 16: time last accessed */
        int32_t ic_atspare;
        time32_t ic_mtime;      /* 24: time last modified */
        int32_t ic_mtspare;
        time32_t ic_ctime;      /* 32: last time inode changed */
        int32_t ic_ctspare;
#endif
        daddr32_t       ic_db[NDADDR];  /* 40: disk block addresses */
        daddr32_t       ic_ib[NIADDR];  /* 88: indirect blocks */
        int32_t ic_flags;       /* 100: cflags */
        int32_t ic_blocks;      /* 104: 512 byte blocks actually held */
        int32_t ic_gen;         /* 108: generation number */
        int32_t ic_shadow;      /* 112: shadow inode */
        uid_t   ic_uid;         /* 116: long EFT version of uid */
        gid_t   ic_gid;         /* 120: long EFT version of gid */
        uint32_t ic_oeftflag;   /* 124: extended attr directory ino, 0 = none */
};

/*
 * Large directories can be cached. Directory caching can take the following
 * states:
 */
typedef enum {
        CD_DISABLED_NOMEM = -2,
        CD_DISABLED_TOOBIG,
        CD_DISABLED,
        CD_ENABLED
} cachedir_t;

/*
 * Large Files: Note we use the inline functions load_double, store_double
 * to load and store the long long values of i_size. Therefore the
 * address of i_size must be eight byte aligned. Kmem_alloc of incore
 * inode structure makes sure that the structure is 8-byte aligned.
 * XX64 - reorder this structure?
 */
typedef struct inode {
        struct  inode *i_chain[2];      /* must be first */
        struct inode *i_freef;  /* free list forward - must be before i_ic */
        struct inode *i_freeb;  /* free list back - must be before i_ic */
        struct  icommon i_ic;   /* Must be here */
        struct  vnode *i_vnode; /* vnode associated with this inode */
        struct  vnode *i_devvp; /* vnode for block I/O */
        dev_t   i_dev;          /* device where inode resides */
        ino_t   i_number;       /* i number, 1-to-1 with device address */
        off_t   i_diroff;       /* offset in dir, where we found last entry */
                                /* just a hint - no locking needed */
        struct ufsvfs *i_ufsvfs; /* incore fs associated with inode */
        struct  dquot *i_dquot; /* quota structure controlling this file */
        krwlock_t i_rwlock;     /* serializes write/setattr requests */
        krwlock_t i_contents;   /* protects (most of) inode contents */
        kmutex_t i_tlock;       /* protects time fields, i_flag */
        offset_t i_nextr;       /*                                      */
                                /* next byte read offset (read-ahead)   */
                                /*   No lock required                   */
                                /*                                      */
        uint_t  i_flag;         /* inode flags */
        uint_t  i_seq;          /* modification sequence number */
        cachedir_t i_cachedir;  /* Cache this directory on next lookup */
                                /* - no locking needed  */
        long    i_mapcnt;       /* mappings to file pages */
        int     *i_map;         /* block list for the corresponding file */
        dev_t   i_rdev;         /* INCORE rdev from i_oldrdev by ufs_iget */
        size_t  i_delaylen;     /* delayed writes, units=bytes */
        offset_t i_delayoff;    /* where we started delaying */
        offset_t i_nextrio;     /* where to start the next clust */
        long    i_writes;       /* number of outstanding bytes in write q */
        kcondvar_t i_wrcv;      /* sleep/wakeup for write throttle */
        offset_t i_doff;        /* dinode byte offset in file system */
        si_t *i_ufs_acl;        /* pointer to acl entry */
        dcanchor_t i_danchor;   /* directory cache anchor */
        kthread_t *i_writer;    /* thread which is in window in wrip() */
} inode_t;

struct dinode {
        union {
                struct  icommon di_icom;
                char    di_size[128];
        } di_un;
};

#define i_mode          i_ic.ic_smode
#define i_nlink         i_ic.ic_nlink
#define i_uid           i_ic.ic_uid
#define i_gid           i_ic.ic_gid
#define i_smode         i_ic.ic_smode
#define i_suid          i_ic.ic_suid
#define i_sgid          i_ic.ic_sgid

#define i_size          i_ic.ic_lsize
#define i_db            i_ic.ic_db
#define i_ib            i_ic.ic_ib

#define i_atime         i_ic.ic_atime
#define i_mtime         i_ic.ic_mtime
#define i_ctime         i_ic.ic_ctime

#define i_shadow        i_ic.ic_shadow
#define i_oeftflag      i_ic.ic_oeftflag
#define i_blocks        i_ic.ic_blocks
#define i_cflags        i_ic.ic_flags
#ifdef _LITTLE_ENDIAN
/*
 * Originally done on x86, but carried on to all other little
 * architectures, which provides for file system compatibility.
 */
#define i_ordev         i_ic.ic_db[1]   /* USL SVR4 compatibility */
#else
#define i_ordev         i_ic.ic_db[0]   /* was i_oldrdev */
#endif
#define i_gen           i_ic.ic_gen
#define i_forw          i_chain[0]
#define i_back          i_chain[1]

/* EFT transition aids - obsolete */
#define oEFT_MAGIC      0x90909090
#define di_oeftflag     di_ic.ic_oeftflag

#define di_ic           di_un.di_icom
#define di_mode         di_ic.ic_smode
#define di_nlink        di_ic.ic_nlink
#define di_uid          di_ic.ic_uid
#define di_gid          di_ic.ic_gid
#define di_smode        di_ic.ic_smode
#define di_suid         di_ic.ic_suid
#define di_sgid         di_ic.ic_sgid

#define di_size         di_ic.ic_lsize
#define di_db           di_ic.ic_db
#define di_ib           di_ic.ic_ib

#define di_atime        di_ic.ic_atime
#define di_mtime        di_ic.ic_mtime
#define di_ctime        di_ic.ic_ctime
#define di_cflags       di_ic.ic_flags

#ifdef _LITTLE_ENDIAN
#define di_ordev        di_ic.ic_db[1]
#else
#define di_ordev        di_ic.ic_db[0]
#endif
#define di_shadow       di_ic.ic_shadow
#define di_blocks       di_ic.ic_blocks
#define di_gen          di_ic.ic_gen

/* flags */
#define IUPD            0x0001          /* file has been modified */
#define IACC            0x0002          /* inode access time to be updated */
#define IMOD            0x0004          /* inode has been modified */
#define ICHG            0x0008          /* inode has been changed */
#define INOACC          0x0010          /* no access time update in getpage */
#define IMODTIME        0x0020          /* mod time already set */
#define IREF            0x0040          /* inode is being referenced */
#define ISYNC           0x0080          /* do all allocation synchronously */
#define IFASTSYMLNK     0x0100          /* fast symbolic link */
#define IMODACC         0x0200          /* only access time changed; */
                                        /*   filesystem won't become active */
#define IATTCHG         0x0400          /* only size/blocks have changed */
#define IBDWRITE        0x0800          /* the inode has been scheduled for */
                                        /* write operation asynchronously */
#define ISTALE          0x1000          /* inode couldn't be read from disk */
#define IDEL            0x2000          /* inode is being deleted */
#define IDIRECTIO       0x4000          /* attempt directio */
#define ISEQ            0x8000          /* deferred i_seq increase */
#define IJUNKIQ         0x10000         /* on junk idle queue */
#define IQUIET          0x20000         /* No file system full messages */

/* cflags */
#define IXATTR          0x0001          /* extended attribute */
#define IFALLOCATE      0x0002          /* fallocate'd file */
#define ICOMPRESS       0x0004          /* compressed for dcfs - see */
                                        /*   `ufs_ioctl()`_FIO_COMPRESSED */

/* modes */
#define IFMT            0170000         /* type of file */
#define IFIFO           0010000         /* named pipe (fifo) */
#define IFCHR           0020000         /* character special */
#define IFDIR           0040000         /* directory */
#define IFBLK           0060000         /* block special */
#define IFREG           0100000         /* regular */
#define IFLNK           0120000         /* symbolic link */
#define IFSHAD          0130000         /* shadow indode */
#define IFSOCK          0140000         /* socket */
#define IFATTRDIR       0160000         /* Attribute directory */

#define ISUID           04000           /* set user id on execution */
#define ISGID           02000           /* set group id on execution */
#define ISVTX           01000           /* save swapped text even after use */
#define IREAD           0400            /* read, write, execute permissions */
#define IWRITE          0200
#define IEXEC           0100

/* specify how the inode info is written in ufs_syncip() */
#define I_SYNC          1               /* wait for the inode written to disk */
#define I_DSYNC         2               /* wait for the inode written to disk */
                                        /* only if IATTCHG is set */
#define I_ASYNC         0               /* don't wait for the inode written */

/* flags passed to ufs_itrunc(), indirtrunc(), and free() */
#define I_FREE  0x00000001              /* inode is being freed */
#define I_DIR   0x00000002              /* inode is a directory */
#define I_IBLK  0x00000004              /* indirect block */
#define I_CHEAP 0x00000008              /* cheap free */
#define I_SHAD  0x00000010              /* inode is a shadow inode */
#define I_QUOTA 0x00000020              /* quota file */
#define I_NOCANCEL      0x40            /* Don't cancel these fragments */
#define I_ACCT  0x00000080              /* Update ufsvfs' unreclaimed_blocks */

/*
 * If ufs_dircheckforname() fails to find an entry with the given name,
 * this "slot" structure holds state for ufs_direnter_*() as to where
 * there is space to put an entry with that name.
 * If ufs_dircheckforname() finds an entry with the given name, this structure
 * holds state for ufs_dirrename() and ufs_dirremove() as to where the
 * entry is. "status" indicates what ufs_dircheckforname() found:
 *      NONE            name not found, large enough free slot not found,
 *      FOUND           name not found, large enough free slot found
 *      EXIST           name found
 * If ufs_dircheckforname() fails due to an error, this structure is not
 * filled in.
 *
 * After ufs_dircheckforname() succeeds the values are:
 *      status  offset          size            fbp, ep
 *      ------  ------          ----            -------
 *      NONE    end of dir      needed          not valid
 *      FOUND   start of entry  of ent          both valid if fbp != NULL
 *      EXIST   start of entry  of prev ent     valid
 *
 * "endoff" is set to 0 if the an entry with the given name is found, or if no
 * free slot could be found or made; this means that the directory should not
 * be truncated.  If the entry was found, the search terminates so
 * ufs_dircheckforname() didn't find out where the last valid entry in the
 * directory was, so it doesn't know where to cut the directory off; if no free
 * slot could be found or made, the directory has to be extended to make room
 * for the new entry, so there's nothing to cut off.
 * Otherwise, "endoff" is set to the larger of the offset of the last
 * non-empty entry in the directory, or the offset at which the new entry will
 * be placed, whichever is larger.  This is used by ufs_diraddentry(); if a new
 * entry is to be added to the directory, any complete directory blocks at the
 * end of the directory that contain no non-empty entries are lopped off the
 * end, thus shrinking the directory dynamically.
 */
typedef enum {NONE, FOUND, EXIST} slotstat_t;
struct ufs_slot {
        struct  direct *ep;     /* pointer to slot */
        struct  fbuf *fbp;      /* dir buf where slot is */
        off_t   offset;         /* offset of area with free space */
        off_t   endoff;         /* last useful location found in search */
        slotstat_t status;      /* status of slot */
        int     size;           /* size of area at slotoffset */
        int     cached;         /* cached directory */
};

/*
 * Statistics on inodes
 * Not protected by locks
 */
struct instats {
        kstat_named_t in_size;          /* current cache size */
        kstat_named_t in_maxsize;       /* maximum cache size */
        kstat_named_t in_hits;          /* cache hits */
        kstat_named_t in_misses;        /* cache misses */
        kstat_named_t in_malloc;        /* kmem_alloce'd */
        kstat_named_t in_mfree;         /* kmem_free'd */
        kstat_named_t in_maxreached;    /* Largest size reached by cache */
        kstat_named_t in_frfront;       /* # put at front of freelist */
        kstat_named_t in_frback;        /* # put at back of freelist */
        kstat_named_t in_qfree;         /* q's to delete thread */
        kstat_named_t in_scan;          /* # inodes scanned */
        kstat_named_t in_tidles;        /* # inodes idled by idle thread */
        kstat_named_t in_lidles;        /* # inodes idled by ufs_lookup */
        kstat_named_t in_vidles;        /* # inodes idled by ufs_vget */
        kstat_named_t in_kcalloc;       /* # inodes kmem_cache_alloced */
        kstat_named_t in_kcfree;        /* # inodes kmem_cache_freed */
        kstat_named_t in_poc;           /* # push-on-close's */
};

#ifdef _KERNEL

/*
 * Extended attributes
 */

#define XATTR_DIR_NAME  "/@/"
extern int      ufs_ninode;             /* high-water mark for inode cache */

extern struct vnodeops *ufs_vnodeops;   /* vnode operations for ufs */
extern const struct fs_operation_def ufs_vnodeops_template[];

/*
 * Convert between inode pointers and vnode pointers
 */
#define VTOI(VP)        ((struct inode *)(VP)->v_data)
#define ITOV(IP)        ((struct vnode *)(IP)->i_vnode)

/*
 * convert to fs
 */
#define ITOF(IP)        ((struct fs *)(IP)->i_fs)

/*
 * Convert between vnode types and inode formats
 */
extern enum vtype       iftovt_tab[];

#ifdef notneeded

/* Look at sys/mode.h and os/vnode.c */

extern int              vttoif_tab[];

#endif

/*
 * Mark an inode with the current (unique) timestamp.
 * (Note that UFS's concept of time only keeps 32 bits of seconds
 * in the on-disk format).
 */
struct timeval32 iuniqtime;
extern kmutex_t ufs_iuniqtime_lock;

#define ITIMES_NOLOCK(ip) ufs_itimes_nolock(ip)

#define ITIMES(ip) { \
        mutex_enter(&(ip)->i_tlock); \
        ITIMES_NOLOCK(ip); \
        mutex_exit(&(ip)->i_tlock); \
}

/*
 * The following interfaces are used to do atomic loads and stores
 * of an inode's i_size, which is a long long data type.
 *
 * For LP64, we just to a load or a store - atomicity and alignment
 * are 8-byte guaranteed.  For x86 there are no such instructions,
 * so we grab i_contents as reader to get the size; we already hold
 * it as writer when we're setting the size.
 */

#ifdef _LP64

#define UFS_GET_ISIZE(resultp, ip)      *(resultp) = (ip)->i_size
#define UFS_SET_ISIZE(value, ip)        (ip)->i_size = (value)

#else   /* _LP64 */

#define UFS_GET_ISIZE(resultp, ip)                              \
        {                                                       \
                rw_enter(&(ip)->i_contents, RW_READER);         \
                *(resultp) = (ip)->i_size;                      \
                rw_exit(&(ip)->i_contents);                     \
        }
#define UFS_SET_ISIZE(value, ip)                                \
        {                                                       \
                ASSERT(RW_WRITE_HELD(&(ip)->i_contents));       \
                (ip)->i_size = (value);                         \
        }

#endif  /* _LP64 */

/*
 * Allocate the specified block in the inode
 * and make sure any in-core pages are initialized.
 */
#define BMAPALLOC(ip, off, size, cr) \
        bmap_write((ip), (u_offset_t)(off), (size), BI_NORMAL, NULL, cr)

#define ESAME   (-1)            /* trying to rename linked files (special) */

#define UFS_HOLE        (daddr32_t)-1   /* value used when no block allocated */

/*
 * enums
 */

/* direnter ops */
enum de_op { DE_CREATE, DE_MKDIR, DE_LINK, DE_RENAME, DE_SYMLINK, DE_ATTRDIR};

/* dirremove ops */
enum dr_op { DR_REMOVE, DR_RMDIR, DR_RENAME };

/*
 * block initialization type for bmap_write
 *
 * BI_NORMAL - allocate and zero fill pages in memory
 * BI_ALLOC_ONLY - only allocate the block, do not zero out pages in mem
 * BI_FALLOCATE - allocate only, do not zero out pages, and store as negative
 *                block number in inode block list
 */
enum bi_type { BI_NORMAL, BI_ALLOC_ONLY, BI_FALLOCATE };

/*
 * This overlays the fid structure (see vfs.h)
 *
 * LP64 note: we use int32_t instead of ino_t since UFS does not use
 * inode numbers larger than 32-bits and ufid's are passed to NFS
 * which expects them to not grow in size beyond 10 bytes (12 including
 * the length).
 */
struct ufid {
        ushort_t ufid_len;
        ushort_t ufid_flags;
        int32_t ufid_ino;
        int32_t ufid_gen;
};

/*
 * each ufs thread (see ufs_thread.c) is managed by this struct
 */
struct ufs_q {
        union uq_head {
                void            *_uq_generic;   /* first entry on q */
                struct inode    *_uq_i;
                ufs_failure_t   *_uq_uf;
        } _uq_head;
        int             uq_ne;          /* # of entries/failures found */
        int             uq_lowat;       /* thread runs when ne == lowat */
        int             uq_hiwat;       /* synchronous idle if ne >= hiwat */
        ushort_t        uq_flags;       /* flags (see below) */
        kcondvar_t      uq_cv;          /* for sleep/wakeup */
        kthread_id_t    uq_threadp;     /* thread managing this q */
        kmutex_t        uq_mutex;       /* protects this struct */
};

#define uq_head         _uq_head._uq_generic
#define uq_ihead        _uq_head._uq_i
#define uq_ufhead       _uq_head._uq_uf

/*
 * uq_flags
 */
#define UQ_EXIT         (0x0001)        /* q server exits at its convenience */
#define UQ_WAIT         (0x0002)        /* thread is waiting on q server */
#define UQ_SUSPEND      (0x0004)        /* request for suspension */
#define UQ_SUSPENDED    (0x0008)        /* thread has suspended itself */

/*
 * When logging is enabled, statvfs must account for blocks and files that
 * may be on the delete queue.  Protected by ufsvfsp->vfs_delete.uq_mutex
 */
struct ufs_delq_info {
        u_offset_t      delq_unreclaimed_blocks;
        ulong_t         delq_unreclaimed_files;
};


/*
 * global idle queues
 * The queues are sized dynamically in proportion to ufs_ninode
 * which, unless overridden, scales with the amount of memory.
 * The idle queue is halved whenever it hits the low water mark
 * (1/4 of ufs_ninode), but can burst to sizes much larger. The number
 * of hash queues is currently maintained to give on average IQHASHQLEN
 * entries when the idle queue is at the low water mark.
 * Note, we do not need to search along the hash queues, but use them
 * in order to batch together geographically local inodes to allow
 * their updates (via the log or buffer cache) to require less disk seeks.
 * This gives an incredible performance boost for logging and a boost for
 * non logging file systems.
 */
typedef struct {
        inode_t *i_chain[2];    /* must match inode_t, but unused */
        inode_t *i_freef;       /* must match inode_t, idle list forward */
        inode_t *i_freeb;       /* must match inode_t, idle list back  */
} iqhead_t;

extern struct ufs_q ufs_idle_q;         /* used by global ufs idle thread */
extern iqhead_t *ufs_junk_iq;           /* junk idle queues */
extern iqhead_t *ufs_useful_iq;         /* useful idle queues */
extern int ufs_njunk_iq;                /* number of entries in junk iq */
extern int ufs_nuseful_iq;              /* number of entries in useful iq */
extern int ufs_niqhash;                 /* number of iq hash qs - power of 2 */
extern int ufs_iqhashmask;              /* iq hash mask = ufs_niqhash - 1 */

#define IQHASHQLEN 32                   /* see comments above */
#define INOCGSHIFT 7                    /* 128 inodes per cylinder group */
#define IQHASH(ip) (((ip)->i_number >> INOCGSHIFT) & ufs_iqhashmask)
#define IQNEXT(i) ((i) + 1) & ufs_iqhashmask /* next idle queue */

extern struct ufs_q     ufs_hlock;      /* used by global ufs hlock thread */

/*
 * vfs_lfflags flags
 */
#define UFS_LARGEFILES  ((ushort_t)0x1) /* set if mount allows largefiles */

/*
 * vfs_dfritime flags
 */
#define UFS_DFRATIME    0x1             /* deferred access time */

/*
 * UFS VFS private data.
 *
 * UFS file system instances may be linked on several lists.
 *
 * -    The vfs_next field chains together every extant ufs instance; this
 *      list is rooted at ufs_instances and should be used in preference to
 *      the overall vfs list (which is properly the province of the generic
 *      file system code, not of file system implementations).  This same list
 *      link is used during forcible unmounts to chain together instances that
 *      can't yet be completely dismantled,
 *
 * -    The vfs_wnext field is used within ufs_update to form a work list of
 *      UFS instances to be synced out.
 */
typedef struct ufsvfs {
        struct vfs      *vfs_vfs;       /* back link                    */
        struct ufsvfs   *vfs_next;      /* instance list link           */
        struct ufsvfs   *vfs_wnext;     /* work list link               */
        struct vnode    *vfs_root;      /* root vnode                   */
        struct buf      *vfs_bufp;      /* buffer containing superblock */
        struct vnode    *vfs_devvp;     /* block device vnode           */
        ushort_t        vfs_lfflags;    /* Large files (set by mount)   */
        ushort_t        vfs_qflags;     /* QUOTA: filesystem flags      */
        struct inode    *vfs_qinod;     /* QUOTA: pointer to quota file */
        uint_t          vfs_btimelimit; /* QUOTA: block time limit      */
        uint_t          vfs_ftimelimit; /* QUOTA: file time limit       */
        krwlock_t       vfs_dqrwlock;   /* QUOTA: protects quota fields */
        /*
         * some fs local threads
         */
        struct ufs_q    vfs_delete;     /* delayed inode delete */
        struct ufs_q    vfs_reclaim;    /* reclaim open, deleted files */

        /*
         * This is copied from the super block at mount time.
         */
        int             vfs_nrpos;      /* # rotational positions */
        /*
         * This lock protects cg's and super block pointed at by
         * vfs_bufp->b_fs.  Locks contents of fs and cg's and contents
         * of vfs_dio.
         */
        kmutex_t        vfs_lock;
        struct ulockfs  vfs_ulockfs;    /* ufs lockfs support */
        uint_t          vfs_dio;        /* delayed io (_FIODIO) */
        uint_t          vfs_nointr;     /* disallow lockfs interrupts */
        uint_t          vfs_nosetsec;   /* disallow ufs_setsecattr */
        uint_t          vfs_syncdir;    /* synchronous local directory ops */
        uint_t          vfs_dontblock;  /* don't block on forced umount */

        /*
         * trans (logging ufs) stuff
         */
        uint_t          vfs_domatamap;  /* set if matamap enabled */
        ulong_t         vfs_maxacl;     /* transaction stuff - max acl size */
        ulong_t         vfs_dirsize;    /* logspace for directory creation */
        ulong_t         vfs_avgbfree;   /* average free blks in cg (blkpref) */
        /*
         * Some useful constants
         */
        int     vfs_nindirshift;        /* calc. from fs_nindir */
        int     vfs_nindiroffset;       /* calc. from fs_ninidr */
        int     vfs_ioclustsz;          /* bytes in read/write cluster */
        int     vfs_iotransz;           /* max device i/o transfer size  */

        vfs_ufsfx_t     vfs_fsfx;       /* lock/fix-on-panic support */
        /*
         * More useful constants
         */
        int     vfs_minfrags;           /* calc. from fs_minfree */
        /*
         * Force DirectIO on all files
         */
        uint_t  vfs_forcedirectio;
        /*
         * Deferred inode time related fields
         */
        clock_t         vfs_iotstamp;   /* last I/O timestamp */
        uint_t          vfs_dfritime;   /* deferred inode time flags */
        /*
         * Some more useful info
         */
        dev_t           vfs_dev;        /* device mounted from */
        struct ml_unit  *vfs_log;       /* pointer to embedded log struct */
        uint_t          vfs_noatime;    /* disable inode atime updates */
        /*
         * snapshot stuff
         */
        void            *vfs_snapshot;  /* snapshot handle */
        /*
         *  Controls logging "file system full" messages to messages file
         */
        clock_t         vfs_lastwhinetime;

        int             vfs_nolog_si;   /* not logging summary info */
        int             vfs_validfs;    /* indicates mounted fs */

        /*
         * Additional information about vfs_delete above
         */
        struct ufs_delq_info vfs_delete_info; /* what's on the delete queue */
} ufsvfs_t;

#define vfs_fs  vfs_bufp->b_un.b_fs

/*
 * values for vfs_validfs
 */
#define UT_UNMOUNTED    0
#define UT_MOUNTED      1
#define UT_HLOCKING     2

/* inohsz is guaranteed to be a power of 2 */
#define INOHASH(ino)    (((int)ino) & (inohsz - 1))

#define ISFALLOCBLK(ip, bn)     \
        (((bn) < 0) && ((bn) % ip->i_fs->fs_frag == 0) && \
        ((ip)->i_cflags & IFALLOCATE && (bn) != UFS_HOLE))

union ihead {
        union   ihead   *ih_head[2];
        struct  inode   *ih_chain[2];
};

extern  union   ihead   *ihead;
extern  kmutex_t        *ih_lock;
extern  int     *ih_ne;
extern  int     inohsz;

extern  clock_t ufs_iowait;

#endif /* _KERNEL */

/*
 * ufs function prototypes
 */
#if defined(_KERNEL) && !defined(_BOOT)

extern  void    ufs_iinit(void);
extern  int     ufs_iget(struct vfs *, ino_t, struct inode **, cred_t *);
extern  int     ufs_iget_alloced(struct vfs *, ino_t, struct inode **,
    cred_t *);
extern  void    ufs_reset_vnode(vnode_t *);
extern  void    ufs_iinactive(struct inode *);
extern  void    ufs_iupdat(struct inode *, int);
extern  int     ufs_rmidle(struct inode *);
extern  int     ufs_itrunc(struct inode *, u_offset_t, int, cred_t *);
extern  int     ufs_iaccess(struct inode *, int, cred_t *, int);
extern  int     rdip(struct inode *, struct uio *, int, struct cred *);
extern  int     wrip(struct inode *, struct uio *, int, struct cred *);

extern void     ufs_imark(struct inode *);
extern void     ufs_itimes_nolock(struct inode *);

extern  int     ufs_diraccess(struct inode *, int, struct cred *);
extern  int     ufs_dirlook(struct inode *, char *, struct inode **,
    cred_t *, int, int);
extern  int     ufs_direnter_cm(struct inode *, char *, enum de_op,
    struct vattr *, struct inode **, cred_t *, int);
extern  int     ufs_direnter_lr(struct inode *, char *, enum de_op,
    struct inode *, struct inode *, cred_t *);
extern  int     ufs_dircheckpath(ino_t, struct inode *, struct inode *,
    struct cred *);
extern  int     ufs_dirmakeinode(struct inode *, struct inode **,
    struct vattr *, enum de_op, cred_t *);
extern  int     ufs_dirremove(struct inode *, char *, struct inode *,
    vnode_t *, enum dr_op, cred_t *);
extern  int     ufs_dircheckforname(struct inode *, char *, int,
    struct ufs_slot *, struct inode **, struct cred *, int);
extern  int     ufs_xattrdirempty(struct inode *, ino_t, cred_t *);
extern  int     blkatoff(struct inode *, off_t, char **, struct fbuf **);

extern  void    sbupdate(struct vfs *);

extern  int     ufs_ialloc(struct inode *, ino_t, mode_t, struct inode **,
    cred_t *);
extern  void    ufs_ifree(struct inode *, ino_t, mode_t);
extern  void    free(struct inode *, daddr_t, off_t, int);
extern  int     alloc(struct inode *, daddr_t, int, daddr_t *, cred_t *);
extern  int     realloccg(struct inode *, daddr_t, daddr_t, int, int,
    daddr_t *, cred_t *);
extern  int     ufs_allocsp(struct vnode *, struct flock64 *, cred_t *);
extern  int     ufs_freesp(struct vnode *, struct flock64 *, int, cred_t *);
extern  ino_t   dirpref(inode_t *);
extern  daddr_t blkpref(struct inode *, daddr_t, int, daddr32_t *);
extern  daddr_t contigpref(ufsvfs_t *, size_t, size_t);

extern  int     ufs_rdwri(enum uio_rw, int, struct inode *, caddr_t, ssize_t,
        offset_t, enum uio_seg, int *, cred_t *);

extern  int     bmap_read(struct inode *, u_offset_t, daddr_t *, int *);
extern  int     bmap_write(struct inode *, u_offset_t, int, enum bi_type,
    daddr_t *, struct cred *);
extern  int     bmap_has_holes(struct inode *);
extern  int     bmap_find(struct inode *, boolean_t, u_offset_t *);
extern  int     bmap_set_bn(struct vnode *, u_offset_t, daddr32_t);

extern  void    ufs_vfs_add(struct ufsvfs *);
extern  void    ufs_vfs_remove(struct ufsvfs *);

extern  void    ufs_sbwrite(struct ufsvfs *);
extern  void    ufs_update(int);
extern  int     ufs_getsummaryinfo(dev_t, struct ufsvfs *, struct fs *);
extern  int     ufs_putsummaryinfo(dev_t, struct ufsvfs *, struct fs *);
extern  int     ufs_syncip(struct inode *, int, int, top_t);
extern  int     ufs_sync_indir(struct inode *);
extern  int     ufs_indirblk_sync(struct inode *, offset_t);
extern  int     ufs_badblock(struct inode *, daddr_t);
extern  int     ufs_indir_badblock(struct inode *, daddr32_t *);
extern  void    ufs_notclean(struct ufsvfs *);
extern  void    ufs_checkclean(struct vfs *);
extern  int     isblock(struct fs *, uchar_t *, daddr_t);
extern  void    setblock(struct fs *, uchar_t *, daddr_t);
extern  void    clrblock(struct fs *, uchar_t *, daddr_t);
extern  int     isclrblock(struct fs *, uchar_t *, daddr_t);
extern  void    fragacct(struct fs *, int, int32_t *, int);
extern  int     skpc(char, uint_t, char *);
extern  int     ufs_fbwrite(struct fbuf *, struct inode *);
extern  int     ufs_fbiwrite(struct fbuf *, struct inode *, daddr_t, long);
extern  int     ufs_putapage(struct vnode *, struct page *, u_offset_t *,
                                size_t *, int, struct cred *);
extern inode_t  *ufs_alloc_inode(ufsvfs_t *, ino_t);
extern void     ufs_free_inode(inode_t *);

/*
 * special stuff
 */
extern  void    ufs_setreclaim(struct inode *);
extern  int     ufs_scan_inodes(int, int (*)(struct inode *, void *), void *,
                                struct ufsvfs *);
extern  int     ufs_sync_inode(struct inode *, void *);
extern  int     ufs_sticky_remove_access(struct inode *, struct inode *,
    struct cred *);
/*
 * quota
 */
extern  int     chkiq(struct ufsvfs *, int, struct inode *, uid_t, int,
                        struct cred *, char **errp, size_t *lenp);

/*
 * ufs thread stuff
 */
extern  void    ufs_thread_delete(struct vfs *);
extern  void    ufs_delete_drain(struct vfs *, int, int);
extern  void    ufs_delete(struct ufsvfs *, struct inode *, int);
extern  void    ufs_inode_cache_reclaim(void *);
extern  void    ufs_idle_drain(struct vfs *);
extern  void    ufs_idle_some(int);
extern  void    ufs_thread_idle(void);
extern  void    ufs_thread_reclaim(struct vfs *);
extern  void    ufs_thread_init(struct ufs_q *, int);
extern  void    ufs_thread_start(struct ufs_q *, void (*)(), struct vfs *);
extern  void    ufs_thread_exit(struct ufs_q *);
extern  void    ufs_thread_suspend(struct ufs_q *);
extern  void    ufs_thread_continue(struct ufs_q *);
extern  void    ufs_thread_hlock(void *);
extern  void    ufs_delete_init(struct ufsvfs *, int);
extern  void    ufs_delete_adjust_stats(struct ufsvfs *, struct statvfs64 *);
extern  void    ufs_delete_drain_wait(struct ufsvfs *, int);

/*
 * ufs lockfs stuff
 */
struct seg;
extern int ufs_reconcile_fs(struct vfs *, struct ufsvfs *, int);
extern int ufs_quiesce(struct ulockfs *);
extern int ufs_flush(struct vfs *);
extern int ufs_fiolfs(struct vnode *, struct lockfs *, int);
extern int ufs__fiolfs(struct vnode *, struct lockfs *, int, int);
extern int ufs_fiolfss(struct vnode *, struct lockfs *);
extern int ufs_fioffs(struct vnode *, char *, struct cred *);
extern int ufs_check_lockfs(struct ufsvfs *, struct ulockfs *, ulong_t);
extern int ufs_lockfs_begin(struct ufsvfs *, struct ulockfs **, ulong_t);
extern int ufs_lockfs_trybegin(struct ufsvfs *, struct ulockfs **, ulong_t);
extern int ufs_lockfs_begin_getpage(struct ufsvfs *, struct ulockfs **,
                struct seg *, int, uint_t *);
extern void ufs_lockfs_end(struct ulockfs *);
/*
 * ufs acl stuff
 */
extern int ufs_si_inherit(struct inode *, struct inode *, o_mode_t, cred_t *);
extern void si_cache_init(void);
extern int ufs_si_load(struct inode *, cred_t *);
extern void ufs_si_del(struct inode *);
extern int ufs_acl_access(struct inode *, int, cred_t *);
extern void ufs_si_cache_flush(dev_t);
extern int ufs_si_free(si_t *, struct vfs *, cred_t *);
extern int ufs_acl_setattr(struct inode *, struct vattr *, cred_t *);
extern int ufs_acl_get(struct inode *, vsecattr_t *, int, cred_t *);
extern int ufs_acl_set(struct inode *, vsecattr_t *, int, cred_t *);
/*
 * ufs directio stuff
 */
extern void ufs_directio_init();
extern int ufs_directio_write(struct inode *, uio_t *, int, int, cred_t *,
    int *);
extern int ufs_directio_read(struct inode *, uio_t *, cred_t *, int *);
#define DIRECTIO_FAILURE        (0)
#define DIRECTIO_SUCCESS        (1)

/*
 * ufs extensions for PXFS
 */

int ufs_rdwr_data(vnode_t *vp, u_offset_t offset, size_t len, fdbuffer_t *fdb,
    int flags, cred_t *cr);
int ufs_alloc_data(vnode_t *vp, u_offset_t offset, size_t *len, fdbuffer_t *fdb,
    int flags, cred_t *cr);

/*
 * prototypes to support the forced unmount
 */

void ufs_freeze(struct ulockfs *, struct lockfs *);
int ufs_thaw(struct vfs *, struct ufsvfs *, struct ulockfs *);

/*
 * extended attributes
 */

int ufs_xattrmkdir(inode_t *, inode_t **, int, struct cred *);
int ufs_xattr_getattrdir(vnode_t *, inode_t **, int, struct cred *);
void ufs_unhook_shadow(inode_t *, inode_t *);

#endif  /* defined(_KERNEL) && !defined(_BOOT) */

#ifdef  __cplusplus
}
#endif

#endif  /* _SYS_FS_UFS_INODE_H */