/*      $NetBSD: ext2fs_balloc.c,v 1.43 2020/09/05 16:30:13 riastradh Exp $     */

/*
* Copyright (c) 1982, 1986, 1989, 1993
*      The Regents of the University of California.  All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
*    may be used to endorse or promote products derived from this software
*    without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
*      @(#)ffs_balloc.c        8.4 (Berkeley) 9/23/93
* Modified for ext2fs by Manuel Bouyer.
*/

/*
* Copyright (c) 1997 Manuel Bouyer.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
*      @(#)ffs_balloc.c        8.4 (Berkeley) 9/23/93
* Modified for ext2fs by Manuel Bouyer.
*/

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: ext2fs_balloc.c,v 1.43 2020/09/05 16:30:13 riastradh Exp $");

#if defined(_KERNEL_OPT)
#include "opt_uvmhist.h"
#endif

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/buf.h>
#include <sys/proc.h>
#include <sys/file.h>
#include <sys/vnode.h>
#include <sys/mount.h>
#include <sys/kauth.h>

#ifdef UVMHIST
#include <uvm/uvm.h>
#endif
#include <uvm/uvm_stat.h>

#include <ufs/ufs/inode.h>
#include <ufs/ufs/ufs_extern.h>

#include <ufs/ext2fs/ext2fs.h>
#include <ufs/ext2fs/ext2fs_extern.h>

/*
* Balloc defines the structure of file system storage
* by allocating the physical blocks on a device given
* the inode and the logical block number in a file.
*/
int
ext2fs_balloc(struct inode *ip, daddr_t bn, int size,
   kauth_cred_t cred, struct buf **bpp, int flags)
{
       struct m_ext2fs *fs;
       daddr_t nb;
       struct buf *bp, *nbp;
       struct vnode *vp = ITOV(ip);
       struct indir indirs[EXT2FS_NIADDR + 2];
       daddr_t newb, lbn, pref;
       int32_t *bap;   /* XXX ondisk32 */
       int num, i, error;
       u_int deallocated;
       daddr_t *blkp, *allocblk, allociblk[EXT2FS_NIADDR + 1];
       int32_t *allocib;       /* XXX ondisk32 */
       int unwindidx = -1;
       UVMHIST_FUNC("ext2fs_balloc"); UVMHIST_CALLED(ubchist);

       UVMHIST_LOG(ubchist, "bn 0x%x", bn, 0, 0, 0);

       if (bpp != NULL) {
               *bpp = NULL;
       }
       if (bn < 0)
               return EFBIG;
       fs = ip->i_e2fs;
       lbn = bn;

       /*
        * The first EXT2FS_NDADDR blocks are direct blocks
        */
       if (bn < EXT2FS_NDADDR) {
               /* XXX ondisk32 */
               nb = fs2h32(ip->i_e2fs_blocks[bn]);
               if (nb != 0) {

                       /*
                        * the block is already allocated, just read it.
                        */

                       if (bpp != NULL) {
                               error = bread(vp, bn, fs->e2fs_bsize,
                                             B_MODIFY, &bp);
                               if (error) {
                                       return error;
                               }
                               *bpp = bp;
                       }
                       return 0;
               }

               /*
                * allocate a new direct block.
                */

               error = ext2fs_alloc(ip, bn,
                   ext2fs_blkpref(ip, bn, bn, &ip->i_e2fs_blocks[0]),
                   cred, &newb);
               if (error)
                       return error;
               ip->i_e2fs_last_lblk = lbn;
               ip->i_e2fs_last_blk = newb;
               /* XXX ondisk32 */
               ip->i_e2fs_blocks[bn] = h2fs32((int32_t)newb);
               ip->i_flag |= IN_CHANGE | IN_UPDATE;
               if (bpp != NULL) {
                       bp = getblk(vp, bn, fs->e2fs_bsize, 0, 0);
                       bp->b_blkno = EXT2_FSBTODB(fs, newb);
                       if (flags & B_CLRBUF)
                               clrbuf(bp);
                       *bpp = bp;
               }
               return 0;
       }
       /*
        * Determine the number of levels of indirection.
        */
       pref = 0;
       if ((error = ufs_getlbns(vp, bn, indirs, &num)) != 0)
               return error;
#ifdef DIAGNOSTIC
       if (num < 1)
               panic("%s: ufs_getlbns returned indirect block\n", __func__);
#endif
       /*
        * Fetch the first indirect block allocating if necessary.
        */
       --num;
       /* XXX ondisk32 */
       nb = fs2h32(ip->i_e2fs_blocks[EXT2FS_NDADDR + indirs[0].in_off]);
       allocib = NULL;
       allocblk = allociblk;
       if (nb == 0) {
               pref = ext2fs_blkpref(ip, lbn, 0, (int32_t *)0);
               error = ext2fs_alloc(ip, lbn, pref, cred, &newb);
               if (error)
                       return error;
               nb = newb;
               *allocblk++ = nb;
               ip->i_e2fs_last_blk = newb;
               bp = getblk(vp, indirs[1].in_lbn, fs->e2fs_bsize, 0, 0);
               bp->b_blkno = EXT2_FSBTODB(fs, newb);
               clrbuf(bp);
               /*
                * Write synchronously so that indirect blocks
                * never point at garbage.
                */
               if ((error = bwrite(bp)) != 0)
                       goto fail;
               unwindidx = 0;
               allocib = &ip->i_e2fs_blocks[EXT2FS_NDADDR + indirs[0].in_off];
               /* XXX ondisk32 */
               *allocib = h2fs32((int32_t)newb);
               ip->i_flag |= IN_CHANGE | IN_UPDATE;
       }
       /*
        * Fetch through the indirect blocks, allocating as necessary.
        */
       for (i = 1;;) {
               error = bread(vp,
                   indirs[i].in_lbn, (int)fs->e2fs_bsize, 0, &bp);
               if (error) {
                       goto fail;
               }
               bap = (int32_t *)bp->b_data;    /* XXX ondisk32 */
               nb = fs2h32(bap[indirs[i].in_off]);
               if (i == num)
                       break;
               i++;
               if (nb != 0) {
                       brelse(bp, 0);
                       continue;
               }
               pref = ext2fs_blkpref(ip, lbn, 0, (int32_t *)0);
               error = ext2fs_alloc(ip, lbn, pref, cred, &newb);
               if (error) {
                       brelse(bp, 0);
                       goto fail;
               }
               nb = newb;
               *allocblk++ = nb;
               ip->i_e2fs_last_blk = newb;
               nbp = getblk(vp, indirs[i].in_lbn, fs->e2fs_bsize, 0, 0);
               nbp->b_blkno = EXT2_FSBTODB(fs, nb);
               clrbuf(nbp);
               /*
                * Write synchronously so that indirect blocks
                * never point at garbage.
                */
               if ((error = bwrite(nbp)) != 0) {
                       brelse(bp, 0);
                       goto fail;
               }
               if (unwindidx < 0)
                       unwindidx = i - 1;
               /* XXX ondisk32 */
               bap[indirs[i - 1].in_off] = h2fs32((int32_t)nb);
               /*
                * If required, write synchronously, otherwise use
                * delayed write.
                */
               if (flags & B_SYNC) {
                       bwrite(bp);
               } else {
                       bdwrite(bp);
               }
       }
       /*
        * Get the data block, allocating if necessary.
        */
       if (nb == 0) {
               pref = ext2fs_blkpref(ip, lbn, indirs[num].in_off, &bap[0]);
               error = ext2fs_alloc(ip, lbn, pref, cred, &newb);
               if (error) {
                       brelse(bp, 0);
                       goto fail;
               }
               nb = newb;
               *allocblk++ = nb;
               ip->i_e2fs_last_lblk = lbn;
               ip->i_e2fs_last_blk = newb;
               /* XXX ondisk32 */
               bap[indirs[num].in_off] = h2fs32((int32_t)nb);
               /*
                * If required, write synchronously, otherwise use
                * delayed write.
                */
               if (flags & B_SYNC) {
                       bwrite(bp);
               } else {
                       bdwrite(bp);
               }
               if (bpp != NULL) {
                       nbp = getblk(vp, lbn, fs->e2fs_bsize, 0, 0);
                       nbp->b_blkno = EXT2_FSBTODB(fs, nb);
                       if (flags & B_CLRBUF)
                               clrbuf(nbp);
                       *bpp = nbp;
               }
               return 0;
       }
       brelse(bp, 0);
       if (bpp != NULL) {
               if (flags & B_CLRBUF) {
                       error = bread(vp, lbn, (int)fs->e2fs_bsize,
                                     B_MODIFY, &nbp);
                       if (error) {
                               goto fail;
                       }
               } else {
                       nbp = getblk(vp, lbn, fs->e2fs_bsize, 0, 0);
                       nbp->b_blkno = EXT2_FSBTODB(fs, nb);
               }
               *bpp = nbp;
       }
       return 0;
fail:
       /*
        * If we have failed part way through block allocation, we
        * have to deallocate any indirect blocks that we have allocated.
        */
       for (deallocated = 0, blkp = allociblk; blkp < allocblk; blkp++) {
               ext2fs_blkfree(ip, *blkp);
               deallocated += fs->e2fs_bsize;
       }
       if (unwindidx >= 0) {
               if (unwindidx == 0) {
                       *allocib = 0;
               } else {
                       int r;

                       r = bread(vp, indirs[unwindidx].in_lbn,
                           (int)fs->e2fs_bsize, B_MODIFY, &bp);
                       if (r) {
                               panic("%s: Could not unwind indirect block, "
                                   "error %d", __func__, r);
                       } else {
                               bap = (int32_t *)bp->b_data; /* XXX ondisk32 */
                               bap[indirs[unwindidx].in_off] = 0;
                               if (flags & B_SYNC)
                                       bwrite(bp);
                               else
                                       bdwrite(bp);
                       }
               }
               for (i = unwindidx + 1; i <= num; i++) {
                       bp = getblk(vp, indirs[i].in_lbn, (int)fs->e2fs_bsize,
                           0, 0);
                       brelse(bp, BC_INVAL);
               }
       }
       if (deallocated) {
               ext2fs_setnblock(ip, ext2fs_nblock(ip) - btodb(deallocated));
               ip->i_e2fs_flags |= IN_CHANGE | IN_UPDATE;
       }
       return error;
}

int
ext2fs_gop_alloc(struct vnode *vp, off_t off, off_t len, int flags,
   kauth_cred_t cred)
{
       struct inode *ip = VTOI(vp);
       struct m_ext2fs *fs = ip->i_e2fs;
       int error, delta, bshift, bsize;
       UVMHIST_FUNC("ext2fs_gop_alloc"); UVMHIST_CALLED(ubchist);

       bshift = fs->e2fs_bshift;
       bsize = 1 << bshift;

       delta = off & (bsize - 1);
       off -= delta;
       len += delta;

       while (len > 0) {
               bsize = uimin(bsize, len);
               UVMHIST_LOG(ubchist, "off 0x%x len 0x%x bsize 0x%x",
                           off, len, bsize, 0);

               error = ext2fs_balloc(ip, ext2_lblkno(fs, off), bsize, cred,
                   NULL, flags);
               if (error) {
                       UVMHIST_LOG(ubchist, "error %d", error, 0, 0, 0);
                       return error;
               }

               /*
                * increase file size now, ext2fs_balloc() requires that
                * EOF be up-to-date before each call.
                */

               if (ext2fs_size(ip) < off + bsize) {
                       UVMHIST_LOG(ubchist, "old 0x%lx%8lx new 0x%lx%8lx",
                           /* Note that arguments are always cast to u_long. */
                                   ext2fs_size(ip) >> 32,
                                   ext2fs_size(ip) & 0xffffffff,
                                   (off + bsize) >> 32,
                                   (off + bsize) & 0xffffffff);
                       error = ext2fs_setsize(ip, off + bsize);
                       if (error) {
                               UVMHIST_LOG(ubchist, "error %d",
                                   error, 0, 0, 0);
                               return error;
                       }
               }

               off += bsize;
               len -= bsize;
       }
       return 0;
}