!C99Shell v. 1.0 pre-release build #13!

Software: Apache/2.0.54 (Unix) mod_perl/1.99_09 Perl/v5.8.0 mod_ssl/2.0.54 OpenSSL/0.9.7l DAV/2 FrontPage/5.0.2.2635 PHP/4.4.0 mod_gzip/2.0.26.1a 

uname -a: Linux snow.he.net 4.4.276-v2-mono-1 #1 SMP Wed Jul 21 11:21:17 PDT 2021 i686 

uid=99(nobody) gid=98(nobody) groups=98(nobody) 

Safe-mode: OFF (not secure)

/usr/src/linux-2.4.18-xfs-1.1/fs/xfs/   drwxr-xr-x
Free 318.33 GB of 458.09 GB (69.49%)
Home    Back    Forward    UPDIR    Refresh    Search    Buffer    Encoder    Tools    Proc.    FTP brute    Sec.    SQL    PHP-code    Update    Feedback    Self remove    Logout    


Viewing file:     xfs_dquot_item.c (18.55 KB)      -rw-r--r--
Select action/file-type:
(+) | (+) | (+) | Code (+) | Session (+) | (+) | SDB (+) | (+) | (+) | (+) | (+) | (+) |
/*
 * Copyright (c) 2000 Silicon Graphics, Inc.  All Rights Reserved.
 * 
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of version 2 of the GNU General Public License as
 * published by the Free Software Foundation.
 * 
 * This program is distributed in the hope that it would be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 * 
 * Further, this software is distributed without any warranty that it is
 * free of the rightful claim of any third person regarding infringement
 * or the like.  Any license provided herein, whether implied or
 * otherwise, applies only to this software file.  Patent licenses, if
 * any, provided herein do not apply to combinations of this program with
 * other software, or any other product whatsoever.
 * 
 * You should have received a copy of the GNU General Public License along
 * with this program; if not, write the Free Software Foundation, Inc., 59
 * Temple Place - Suite 330, Boston MA 02111-1307, USA.
 * 
 * Contact information: Silicon Graphics, Inc., 1600 Amphitheatre Pkwy,
 * Mountain View, CA  94043, or:
 * 
 * http://www.sgi.com 
 * 
 * For further information regarding this notice, see: 
 * 
 * http://oss.sgi.com/projects/GenInfo/SGIGPLNoticeExplan/
 */

#include <xfs.h>
#include <xfs_quota_priv.h>


/*
 * returns the number of iovecs needed to log the given dquot item.
 */
/* ARGSUSED */
STATIC uint
xfs_qm_dquot_logitem_size(
    xfs_dq_logitem_t    *logitem)
{
    /* 
     * we need only two iovecs, one for the format, one for the real thing 
     */
    return (2);
}

/*
 * fills in the vector of log iovecs for the given dquot log item.
 */
STATIC void
xfs_qm_dquot_logitem_format(
    xfs_dq_logitem_t    *logitem,
    xfs_log_iovec_t        *logvec)
{
    ASSERT(logitem);
    ASSERT(logitem->qli_dquot);

    logvec->i_addr = (xfs_caddr_t)&logitem->qli_format;
    logvec->i_len  = sizeof(xfs_dq_logformat_t);
    logvec++;
    logvec->i_addr = (xfs_caddr_t)&logitem->qli_dquot->q_core;
    logvec->i_len  = sizeof(xfs_disk_dquot_t);

    ASSERT(2 == logitem->qli_item.li_desc->lid_size);
    logitem->qli_format.qlf_size = 2;
    
}

/*
 * Increment the pin count of the given dquot.
 * This value is protected by pinlock spinlock in the xQM structure.
 */
STATIC void
xfs_qm_dquot_logitem_pin(
    xfs_dq_logitem_t *logitem)
{
    int     s;
    xfs_dquot_t *dqp;
    
    dqp = logitem->qli_dquot;
    ASSERT(XFS_DQ_IS_LOCKED(dqp));
    s = XFS_DQ_PINLOCK(dqp);
    dqp->q_pincount++;
    XFS_DQ_PINUNLOCK(dqp, s);
}

/*
 * Decrement the pin count of the given dquot, and wake up
 * anyone in xfs_dqwait_unpin() if the count goes to 0.  The
 * dquot must have been previously pinned with a call to xfs_dqpin().
 */
STATIC void
xfs_qm_dquot_logitem_unpin(
    xfs_dq_logitem_t *logitem)
{
    int     s;
    xfs_dquot_t *dqp;

    dqp = logitem->qli_dquot;
    ASSERT(dqp->q_pincount > 0);
    s = XFS_DQ_PINLOCK(dqp);
    dqp->q_pincount--;
    if (dqp->q_pincount == 0) {
        sv_broadcast(&dqp->q_pinwait);
    }
    XFS_DQ_PINUNLOCK(dqp, s);
}

/* ARGSUSED */
STATIC void
xfs_qm_dquot_logitem_unpin_remove(
    xfs_dq_logitem_t *logitem,
    xfs_trans_t     *tp)
{
    xfs_qm_dquot_logitem_unpin(logitem);
}

/*
 * Given the logitem, this writes the corresponding dquot entry to disk
 * asynchronously. This is called with the dquot entry securely locked;
 * we simply get xfs_qm_dqflush() to do the work, and unlock the dquot
 * at the end.
 */
STATIC void
xfs_qm_dquot_logitem_push(
    xfs_dq_logitem_t    *logitem)
{
    xfs_dquot_t    *dqp;

    dqp = logitem->qli_dquot;
    
    ASSERT(XFS_DQ_IS_LOCKED(dqp));
    ASSERT(XFS_DQ_IS_FLUSH_LOCKED(dqp));  
     
        /*
         * Since we were able to lock the dquot's flush lock and
         * we found it on the AIL, the dquot must be dirty.  This
         * is because the dquot is removed from the AIL while still
         * holding the flush lock in xfs_dqflush_done().  Thus, if
         * we found it in the AIL and were able to obtain the flush
         * lock without sleeping, then there must not have been
         * anyone in the process of flushing the dquot.
         */
    xfs_qm_dqflush(dqp, XFS_B_DELWRI);
    xfs_dqunlock(dqp);
}

/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_qm_dquot_logitem_committed(
    xfs_dq_logitem_t    *l,
    xfs_lsn_t        lsn)
{
    /*
     * We always re-log the entire dquot when it becomes dirty,
     * so, the latest copy _is_ the only one that matters.
     */
    return (lsn);
}


/*
 * This is called to wait for the given dquot to be unpinned.
 * Most of these pin/unpin routines are plagiarized from inode code.
 */
void
xfs_qm_dqunpin_wait(
    xfs_dquot_t    *dqp)
{
    int             s;
    
    ASSERT(XFS_DQ_IS_LOCKED(dqp));
    if (dqp->q_pincount == 0) {
        return;
    }

    /*
     * Give the log a push so we don't wait here too long.
     */
    xfs_log_force(dqp->q_mount, (xfs_lsn_t)0, XFS_LOG_FORCE);
    s = XFS_DQ_PINLOCK(dqp);
    if (dqp->q_pincount == 0) {
        XFS_DQ_PINUNLOCK(dqp, s);
        return;
    }
    sv_wait(&(dqp->q_pinwait), PINOD,
        &(XFS_DQ_TO_QINF(dqp)->qi_pinlock), s);
}

/*
 * This is called when IOP_TRYLOCK returns XFS_ITEM_PUSHBUF to indicate that
 * the dquot is locked by us, but the flush lock isn't. So, here we are
 * going to see if the relevant dquot buffer is incore, waiting on DELWRI.
 * If so, we want to push it out to help us take this item off the AIL as soon
 * as possible.
 *
 * We must not be holding the AIL_LOCK at this point. Calling incore() to
 * search the buffercache can be a time consuming thing, and AIL_LOCK is a
 * spinlock.
 */
STATIC void
xfs_qm_dquot_logitem_pushbuf(
        xfs_dq_logitem_t    *qip)
{
    xfs_dquot_t    *dqp;
    xfs_mount_t    *mp;
    xfs_buf_t     *bp;
    uint        dopush;

    dqp = qip->qli_dquot;
    ASSERT(XFS_DQ_IS_LOCKED(dqp));

    /*
     * The qli_pushbuf_flag keeps others from
     * trying to duplicate our effort.
     */
    ASSERT(qip->qli_pushbuf_flag != 0);
    ASSERT(qip->qli_push_owner == get_thread_id());
    
    /*
     * If flushlock isn't locked anymore, chances are that the
     * inode flush completed and the inode was taken off the AIL.
     * So, just get out.
     */
    if ((valusema(&(dqp->q_flock)) > 0)  ||
        ((qip->qli_item.li_flags & XFS_LI_IN_AIL) == 0)) {
        qip->qli_pushbuf_flag = 0;
        xfs_dqunlock(dqp);
            return;
    }
    mp = dqp->q_mount;
    bp = xfs_incore(mp->m_ddev_targ, qip->qli_format.qlf_blkno, 
            XFS_QI_DQCHUNKLEN(mp),
            XFS_INCORE_TRYLOCK);
    if (bp != NULL) {
        if (XFS_BUF_ISDELAYWRITE(bp)) {
            dopush = ((qip->qli_item.li_flags & XFS_LI_IN_AIL) && 
                  (valusema(&(dqp->q_flock)) <= 0));
            qip->qli_pushbuf_flag = 0;
            xfs_dqunlock(dqp);

            if (XFS_BUF_ISPINNED(bp)) {
                xfs_log_force(mp, (xfs_lsn_t)0,
                          XFS_LOG_FORCE);
            }
            if (dopush) {
#ifdef XFSRACEDEBUG
                delay_for_intr();
                delay(300);
#endif                
                xfs_bawrite(mp, bp);
            } else {
                xfs_buf_relse(bp);
            }
        } else {
            qip->qli_pushbuf_flag = 0;
            xfs_dqunlock(dqp);        
            xfs_buf_relse(bp);
        }
        return;
    }

    qip->qli_pushbuf_flag = 0;
    xfs_dqunlock(dqp);    
}

/*
 * This is called to attempt to lock the dquot associated with this
 * dquot log item.  Don't sleep on the dquot lock or the flush lock.
 * If the flush lock is already held, indicating that the dquot has
 * been or is in the process of being flushed, then see if we can
 * find the dquot's buffer in the buffer cache without sleeping.  If
 * we can and it is marked delayed write, then we want to send it out.
 * We delay doing so until the push routine, though, to avoid sleeping
 * in any device strategy routines.
 */
STATIC uint
xfs_qm_dquot_logitem_trylock(
        xfs_dq_logitem_t    *qip)
{
    xfs_dquot_t            *dqp;
    uint            retval;

    dqp = qip->qli_dquot;
    if (dqp->q_pincount > 0)
        return (XFS_ITEM_PINNED);
    
    if (! xfs_qm_dqlock_nowait(dqp))
        return (XFS_ITEM_LOCKED);
    
    retval = XFS_ITEM_SUCCESS;
    if (! xfs_qm_dqflock_nowait(dqp)) {
        /* 
                 * The dquot is already being flushed.  It may have been
                 * flushed delayed write, however, and we don't want to
                 * get stuck waiting for that to complete.  So, we want to check
                 * to see if we can lock the dquot's buffer without sleeping.
                 * If we can and it is marked for delayed write, then we
                 * hold it and send it out from the push routine.  We don't
                 * want to do that now since we might sleep in the device
                 * strategy routine.  We also don't want to grab the buffer lock
         * here because we'd like not to call into the buffer cache
         * while holding the AIL_LOCK.
         * Make sure to only return PUSHBUF if we set pushbuf_flag
         * ourselves.  If someone else is doing it then we don't 
         * want to go to the push routine and duplicate their efforts.
                 */
                if (qip->qli_pushbuf_flag == 0) {
            qip->qli_pushbuf_flag = 1;
            ASSERT(qip->qli_format.qlf_blkno == dqp->q_blkno);
#ifdef DEBUG
            qip->qli_push_owner = get_thread_id();
#endif
            /*
             * The dquot is left locked.
             */
            retval = XFS_ITEM_PUSHBUF;
        } else {
            retval = XFS_ITEM_FLUSHING;
            xfs_dqunlock_nonotify(dqp);
        }
    }
    
    ASSERT(qip->qli_item.li_flags & XFS_LI_IN_AIL);
    return (retval);
}


/*
 * Unlock the dquot associated with the log item.
 * Clear the fields of the dquot and dquot log item that
 * are specific to the current transaction.  If the
 * hold flags is set, do not unlock the dquot.  
 */
STATIC void
xfs_qm_dquot_logitem_unlock(
        xfs_dq_logitem_t    *ql)
{
    xfs_dquot_t    *dqp;

    ASSERT(ql != NULL);
    dqp = ql->qli_dquot;
    ASSERT(XFS_DQ_IS_LOCKED(dqp));

    /*
         * Clear the transaction pointer in the dquot
         */
    dqp->q_transp = NULL;

    /*
     * dquots are never 'held' from getting unlocked at the end of
     * a transaction.  Their locking and unlocking is hidden inside the
     * transaction layer, within trans_commit. Hence, no LI_HOLD flag
     * for the logitem.
     */
    xfs_dqunlock(dqp);
}


/*
 * The transaction with the dquot locked has aborted.  The dquot
 * must not be dirty within the transaction.  We simply unlock just
 * as if the transaction had been cancelled.
 */
STATIC void
xfs_qm_dquot_logitem_abort(
        xfs_dq_logitem_t    *ql)
{
        xfs_qm_dquot_logitem_unlock(ql);
}

/*
 * this needs to stamp an lsn into the dquot, I think.
 * rpc's that look at user dquot's would then have to
 * push on the dependency recorded in the dquot
 */
/* ARGSUSED */
STATIC void
xfs_qm_dquot_logitem_committing(
    xfs_dq_logitem_t    *l,
    xfs_lsn_t        lsn)
{
    return;
}


/*
 * This is the ops vector for dquots
 */
struct xfs_item_ops xfs_dquot_item_ops = {
    iop_size:    (uint(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_size,
    iop_format:    (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
                    xfs_qm_dquot_logitem_format,
    iop_pin:    (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_pin,
    iop_unpin:    (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_unpin,
    iop_unpin_remove: (void(*)(xfs_log_item_t*, xfs_trans_t*))
                    xfs_qm_dquot_logitem_unpin_remove,
    iop_trylock:    (uint(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_trylock,
    iop_unlock:    (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_unlock,
    iop_committed:    (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
                    xfs_qm_dquot_logitem_committed,
    iop_push:    (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_push,
    iop_abort:    (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_abort,
    iop_pushbuf:    (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_pushbuf,
    iop_committing:    (void(*)(xfs_log_item_t*, xfs_lsn_t))
                    xfs_qm_dquot_logitem_committing
};

/*
 * Initialize the dquot log item for a newly allocated dquot.
 * The dquot isn't locked at this point, but it isn't on any of the lists
 * either, so we don't care.
 */
void
xfs_qm_dquot_logitem_init(
    struct xfs_dquot *dqp)
{
    xfs_dq_logitem_t  *lp;
    lp = &dqp->q_logitem;

    lp->qli_item.li_type = XFS_LI_DQUOT;
    lp->qli_item.li_ops = &xfs_dquot_item_ops;
    lp->qli_item.li_mountp = dqp->q_mount;
    lp->qli_dquot = dqp;
    lp->qli_format.qlf_type = XFS_LI_DQUOT;
    lp->qli_format.qlf_id = INT_GET(dqp->q_core.d_id, ARCH_CONVERT);
    lp->qli_format.qlf_blkno = dqp->q_blkno;
    lp->qli_format.qlf_len = 1;
    /*
     * This is just the offset of this dquot within its buffer
     * (which is currently 1 FSB and probably won't change).
     * Hence 32 bits for this offset should be just fine. 
     * Alternatively, we can store (bufoffset / sizeof(xfs_dqblk_t))
     * here, and recompute it at recovery time.
     */
    lp->qli_format.qlf_boffset = (__uint32_t)dqp->q_bufoffset;
}

/*------------------  QUOTAOFF LOG ITEMS  -------------------*/

/*
 * This returns the number of iovecs needed to log the given quotaoff item.
 * We only need 1 iovec for an quotaoff item.  It just logs the
 * quotaoff_log_format structure.
 */
/*ARGSUSED*/
STATIC uint
xfs_qm_qoff_logitem_size(xfs_qoff_logitem_t *qf)
{
    return (1);
}

/*
 * This is called to fill in the vector of log iovecs for the
 * given quotaoff log item. We use only 1 iovec, and we point that
 * at the quotaoff_log_format structure embedded in the quotaoff item.
 * It is at this point that we assert that all of the extent
 * slots in the quotaoff item have been filled.
 */
STATIC void
xfs_qm_qoff_logitem_format(xfs_qoff_logitem_t    *qf,
               xfs_log_iovec_t    *log_vector)
{
    ASSERT(qf->qql_format.qf_type == XFS_LI_QUOTAOFF);

    log_vector->i_addr = (xfs_caddr_t)&(qf->qql_format);
    log_vector->i_len = sizeof(xfs_qoff_logitem_t);
    qf->qql_format.qf_size = 1;
}
    

/*
 * Pinning has no meaning for an quotaoff item, so just return.
 */
/*ARGSUSED*/
STATIC void
xfs_qm_qoff_logitem_pin(xfs_qoff_logitem_t *qf)
{
    return;
}


/*
 * Since pinning has no meaning for an quotaoff item, unpinning does
 * not either.
 */
/*ARGSUSED*/
STATIC void
xfs_qm_qoff_logitem_unpin(xfs_qoff_logitem_t *qf)
{
    return;
}

/*ARGSUSED*/
STATIC void
xfs_qm_qoff_logitem_unpin_remove(xfs_qoff_logitem_t *qf, xfs_trans_t *tp)
{
    return;
}

/*
 * Quotaoff items have no locking, so just return success.
 */
/*ARGSUSED*/
STATIC uint
xfs_qm_qoff_logitem_trylock(xfs_qoff_logitem_t *qf)
{
    return XFS_ITEM_LOCKED;
}

/*
 * Quotaoff items have no locking or pushing, so return failure
 * so that the caller doesn't bother with us.
 */
/*ARGSUSED*/
STATIC void
xfs_qm_qoff_logitem_unlock(xfs_qoff_logitem_t *qf)
{
    return;
}

/*
 * The quotaoff-start-item is logged only once and cannot be moved in the log,
 * so simply return the lsn at which it's been logged.  
 */
/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_qm_qoff_logitem_committed(xfs_qoff_logitem_t *qf, xfs_lsn_t lsn)
{
    return (lsn);
}

/*
 * The transaction of which this QUOTAOFF is a part has been aborted.
 * Just clean up after ourselves.
 * Shouldn't this never happen in the case of qoffend logitems? XXX
 */
STATIC void
xfs_qm_qoff_logitem_abort(xfs_qoff_logitem_t *qf)
{
    kmem_free(qf, sizeof(xfs_qoff_logitem_t));
}

/*
 * There isn't much you can do to push on an quotaoff item.  It is simply
 * stuck waiting for the log to be flushed to disk.
 */
/*ARGSUSED*/
STATIC void
xfs_qm_qoff_logitem_push(xfs_qoff_logitem_t *qf)
{
    return;
}


/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_qm_qoffend_logitem_committed(
    xfs_qoff_logitem_t *qfe, 
    xfs_lsn_t lsn)
{
    xfs_qoff_logitem_t     *qfs;
    SPLDECL(s);
    
    qfs = qfe->qql_start_lip;
    AIL_LOCK(qfs->qql_item.li_mountp,s);
    /*
     * Delete the qoff-start logitem from the AIL.
     * xfs_trans_delete_ail() drops the AIL lock.
     */
    xfs_trans_delete_ail(qfs->qql_item.li_mountp, (xfs_log_item_t *)qfs, s);
    kmem_free(qfs, sizeof(xfs_qoff_logitem_t));
    kmem_free(qfe, sizeof(xfs_qoff_logitem_t));
    return (xfs_lsn_t)-1;
}

/*
 * XXX rcc - don't know quite what to do with this.  I think we can
 * just ignore it.  The only time that isn't the case is if we allow
 * the client to somehow see that quotas have been turned off in which
 * we can't allow that to get back until the quotaoff hits the disk.
 * So how would that happen?  Also, do we need different routines for
 * quotaoff start and quotaoff end?  I suspect the answer is yes but
 * to be sure, I need to look at the recovery code and see how quota off
 * recovery is handled (do we roll forward or back or do something else).
 * If we roll forwards or backwards, then we need two separate routines,
 * one that does nothing and one that stamps in the lsn that matters
 * (truly makes the quotaoff irrevocable).  If we do something else,
 * then maybe we don't need two.
 */
/* ARGSUSED */
STATIC void
xfs_qm_qoff_logitem_committing(xfs_qoff_logitem_t *qip, xfs_lsn_t commit_lsn)
{
    return;
}

/* ARGSUSED */
STATIC void
xfs_qm_qoffend_logitem_committing(xfs_qoff_logitem_t *qip, xfs_lsn_t commit_lsn)
{
    return;
}

struct xfs_item_ops xfs_qm_qoffend_logitem_ops = {
    iop_size:    (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_size,
    iop_format:    (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
                    xfs_qm_qoff_logitem_format,
    iop_pin:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
    iop_unpin:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unpin,
    iop_unpin_remove: (void(*)(xfs_log_item_t*,xfs_trans_t*))
                    xfs_qm_qoff_logitem_unpin_remove,
    iop_trylock:    (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
    iop_unlock:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
    iop_committed:    (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
                    xfs_qm_qoffend_logitem_committed,
    iop_push:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_push,
    iop_abort:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_abort,
    iop_pushbuf:    NULL,
    iop_committing:    (void(*)(xfs_log_item_t*, xfs_lsn_t))
                    xfs_qm_qoffend_logitem_committing
};

/*
 * This is the ops vector shared by all quotaoff-start log items.
 */
struct xfs_item_ops xfs_qm_qoff_logitem_ops = {
    iop_size:    (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_size,
    iop_format:    (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
                    xfs_qm_qoff_logitem_format,
    iop_pin:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
    iop_unpin:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unpin,
    iop_unpin_remove: (void(*)(xfs_log_item_t*,xfs_trans_t*))
                    xfs_qm_qoff_logitem_unpin_remove,
    iop_trylock:    (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
    iop_unlock:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
    iop_committed:    (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
                    xfs_qm_qoff_logitem_committed,
    iop_push:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_push,
    iop_abort:    (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_abort,
    iop_pushbuf:    NULL,
    iop_committing:    (void(*)(xfs_log_item_t*, xfs_lsn_t))
                    xfs_qm_qoff_logitem_committing
};

/*
 * Allocate and initialize an quotaoff item of the correct quota type(s).
 */
xfs_qoff_logitem_t *
xfs_qm_qoff_logitem_init(
    struct xfs_mount *mp, 
    xfs_qoff_logitem_t *start,
    uint flags)
{
    xfs_qoff_logitem_t    *qf;

    qf = (xfs_qoff_logitem_t*) kmem_zalloc(sizeof(xfs_qoff_logitem_t), KM_SLEEP);

    qf->qql_item.li_type = XFS_LI_QUOTAOFF;
    if (start) 
        qf->qql_item.li_ops = &xfs_qm_qoffend_logitem_ops;
    else
        qf->qql_item.li_ops = &xfs_qm_qoff_logitem_ops;
    qf->qql_item.li_mountp = mp;
    qf->qql_format.qf_type = XFS_LI_QUOTAOFF;
    qf->qql_format.qf_flags = flags;
    qf->qql_start_lip = start;
    return (qf);
}

:: Command execute ::

Enter:
 
Select:
 

:: Search ::
  - regexp 

:: Upload ::
 
[ Read-Only ]

:: Make Dir ::
 
[ Read-Only ]
:: Make File ::
 
[ Read-Only ]

:: Go Dir ::
 
:: Go File ::
 

--[ c99shell v. 1.0 pre-release build #13 powered by Captain Crunch Security Team | http://ccteam.ru | Generation time: 0.0225 ]--