X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=fs%2Fxfs%2Fxfs_itable.c;h=62efab2f38392e845bf82fc8ba48825f3c016727;hb=e09f98606dcc156de1146c209d45a0d6d5f51c3f;hp=32247b6bfee73cca08e5d41a38e1ff485e72a2be;hpb=b12dd34298cf0cff9f337f667045335140873039;p=safe%2Fjmp%2Flinux-2.6 diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 32247b6..62efab2 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -24,14 +24,12 @@ #include "xfs_trans.h" #include "xfs_sb.h" #include "xfs_ag.h" -#include "xfs_dir.h" #include "xfs_dir2.h" #include "xfs_dmapi.h" #include "xfs_mount.h" #include "xfs_bmap_btree.h" #include "xfs_alloc_btree.h" #include "xfs_ialloc_btree.h" -#include "xfs_dir_sf.h" #include "xfs_dir2_sf.h" #include "xfs_attr_sf.h" #include "xfs_dinode.h" @@ -41,10 +39,15 @@ #include "xfs_error.h" #include "xfs_btree.h" -#ifndef HAVE_USERACC -#define useracc(ubuffer, size, flags, foo) (0) -#define unuseracc(ubuffer, size, flags) -#endif +STATIC int +xfs_internal_inum( + xfs_mount_t *mp, + xfs_ino_t ino) +{ + return (ino == mp->m_sb.sb_rbmino || ino == mp->m_sb.sb_rsumino || + (xfs_sb_version_hasquota(&mp->m_sb) && + (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino))); +} STATIC int xfs_bulkstat_one_iget( @@ -54,27 +57,23 @@ xfs_bulkstat_one_iget( xfs_bstat_t *buf, /* return buffer */ int *stat) /* BULKSTAT_RV_... */ { - xfs_dinode_core_t *dic; /* dinode core info pointer */ + xfs_icdinode_t *dic; /* dinode core info pointer */ xfs_inode_t *ip; /* incore inode pointer */ - vnode_t *vp; + struct inode *inode; int error; - error = xfs_iget(mp, NULL, ino, 0, XFS_ILOCK_SHARED, &ip, bno); + error = xfs_iget(mp, NULL, ino, + XFS_IGET_BULKSTAT, XFS_ILOCK_SHARED, &ip, bno); if (error) { *stat = BULKSTAT_RV_NOTHING; return error; } ASSERT(ip != NULL); - ASSERT(ip->i_blkno != (xfs_daddr_t)0); - if (ip->i_d.di_mode == 0) { - *stat = BULKSTAT_RV_NOTHING; - error = XFS_ERROR(ENOENT); - goto out_iput; - } + ASSERT(ip->i_imap.im_blkno != 0); - vp = XFS_ITOV(ip); dic = &ip->i_d; + inode = VFS_I(ip); /* xfs_iget returns the following without needing * further change. @@ -86,11 +85,19 @@ xfs_bulkstat_one_iget( buf->bs_uid = dic->di_uid; buf->bs_gid = dic->di_gid; buf->bs_size = dic->di_size; - vn_atime_to_bstime(vp, &buf->bs_atime); - buf->bs_mtime.tv_sec = dic->di_mtime.t_sec; - buf->bs_mtime.tv_nsec = dic->di_mtime.t_nsec; - buf->bs_ctime.tv_sec = dic->di_ctime.t_sec; - buf->bs_ctime.tv_nsec = dic->di_ctime.t_nsec; + + /* + * We need to read the timestamps from the Linux inode because + * the VFS keeps writing directly into the inode structure instead + * of telling us about the updates. + */ + buf->bs_atime.tv_sec = inode->i_atime.tv_sec; + buf->bs_atime.tv_nsec = inode->i_atime.tv_nsec; + buf->bs_mtime.tv_sec = inode->i_mtime.tv_sec; + buf->bs_mtime.tv_nsec = inode->i_mtime.tv_nsec; + buf->bs_ctime.tv_sec = inode->i_ctime.tv_sec; + buf->bs_ctime.tv_nsec = inode->i_ctime.tv_nsec; + buf->bs_xflags = xfs_ip2xflags(ip); buf->bs_extsize = dic->di_extsize << mp->m_sb.sb_blocklog; buf->bs_extents = dic->di_nextents; @@ -120,22 +127,17 @@ xfs_bulkstat_one_iget( break; } - out_iput: xfs_iput(ip, XFS_ILOCK_SHARED); return error; } -STATIC int +STATIC void xfs_bulkstat_one_dinode( xfs_mount_t *mp, /* mount point for filesystem */ xfs_ino_t ino, /* inode number to get data for */ - xfs_dinode_t *dip, /* dinode inode pointer */ + xfs_dinode_t *dic, /* dinode inode pointer */ xfs_bstat_t *buf) /* return buffer */ { - xfs_dinode_core_t *dic; /* dinode core info pointer */ - - dic = &dip->di_core; - /* * The inode format changed when we moved the link count and * made it 32 bits long. If this is an old format inode, @@ -147,37 +149,37 @@ xfs_bulkstat_one_dinode( * the new format. We don't change the version number so that we * can distinguish this from a real new format inode. */ - if (INT_GET(dic->di_version, ARCH_CONVERT) == XFS_DINODE_VERSION_1) { - buf->bs_nlink = INT_GET(dic->di_onlink, ARCH_CONVERT); + if (dic->di_version == 1) { + buf->bs_nlink = be16_to_cpu(dic->di_onlink); buf->bs_projid = 0; } else { - buf->bs_nlink = INT_GET(dic->di_nlink, ARCH_CONVERT); - buf->bs_projid = INT_GET(dic->di_projid, ARCH_CONVERT); + buf->bs_nlink = be32_to_cpu(dic->di_nlink); + buf->bs_projid = be16_to_cpu(dic->di_projid); } buf->bs_ino = ino; - buf->bs_mode = INT_GET(dic->di_mode, ARCH_CONVERT); - buf->bs_uid = INT_GET(dic->di_uid, ARCH_CONVERT); - buf->bs_gid = INT_GET(dic->di_gid, ARCH_CONVERT); - buf->bs_size = INT_GET(dic->di_size, ARCH_CONVERT); - buf->bs_atime.tv_sec = INT_GET(dic->di_atime.t_sec, ARCH_CONVERT); - buf->bs_atime.tv_nsec = INT_GET(dic->di_atime.t_nsec, ARCH_CONVERT); - buf->bs_mtime.tv_sec = INT_GET(dic->di_mtime.t_sec, ARCH_CONVERT); - buf->bs_mtime.tv_nsec = INT_GET(dic->di_mtime.t_nsec, ARCH_CONVERT); - buf->bs_ctime.tv_sec = INT_GET(dic->di_ctime.t_sec, ARCH_CONVERT); - buf->bs_ctime.tv_nsec = INT_GET(dic->di_ctime.t_nsec, ARCH_CONVERT); + buf->bs_mode = be16_to_cpu(dic->di_mode); + buf->bs_uid = be32_to_cpu(dic->di_uid); + buf->bs_gid = be32_to_cpu(dic->di_gid); + buf->bs_size = be64_to_cpu(dic->di_size); + buf->bs_atime.tv_sec = be32_to_cpu(dic->di_atime.t_sec); + buf->bs_atime.tv_nsec = be32_to_cpu(dic->di_atime.t_nsec); + buf->bs_mtime.tv_sec = be32_to_cpu(dic->di_mtime.t_sec); + buf->bs_mtime.tv_nsec = be32_to_cpu(dic->di_mtime.t_nsec); + buf->bs_ctime.tv_sec = be32_to_cpu(dic->di_ctime.t_sec); + buf->bs_ctime.tv_nsec = be32_to_cpu(dic->di_ctime.t_nsec); buf->bs_xflags = xfs_dic2xflags(dic); - buf->bs_extsize = INT_GET(dic->di_extsize, ARCH_CONVERT) << mp->m_sb.sb_blocklog; - buf->bs_extents = INT_GET(dic->di_nextents, ARCH_CONVERT); - buf->bs_gen = INT_GET(dic->di_gen, ARCH_CONVERT); + buf->bs_extsize = be32_to_cpu(dic->di_extsize) << mp->m_sb.sb_blocklog; + buf->bs_extents = be32_to_cpu(dic->di_nextents); + buf->bs_gen = be32_to_cpu(dic->di_gen); memset(buf->bs_pad, 0, sizeof(buf->bs_pad)); - buf->bs_dmevmask = INT_GET(dic->di_dmevmask, ARCH_CONVERT); - buf->bs_dmstate = INT_GET(dic->di_dmstate, ARCH_CONVERT); - buf->bs_aextents = INT_GET(dic->di_anextents, ARCH_CONVERT); + buf->bs_dmevmask = be32_to_cpu(dic->di_dmevmask); + buf->bs_dmstate = be16_to_cpu(dic->di_dmstate); + buf->bs_aextents = be16_to_cpu(dic->di_anextents); - switch (INT_GET(dic->di_format, ARCH_CONVERT)) { + switch (dic->di_format) { case XFS_DINODE_FMT_DEV: - buf->bs_rdev = INT_GET(dip->di_u.di_dev, ARCH_CONVERT); + buf->bs_rdev = xfs_dinode_get_rdev(dic); buf->bs_blksize = BLKDEV_IOSIZE; buf->bs_blocks = 0; break; @@ -191,10 +193,25 @@ xfs_bulkstat_one_dinode( case XFS_DINODE_FMT_BTREE: buf->bs_rdev = 0; buf->bs_blksize = mp->m_sb.sb_blocksize; - buf->bs_blocks = INT_GET(dic->di_nblocks, ARCH_CONVERT); + buf->bs_blocks = be64_to_cpu(dic->di_nblocks); break; } +} +/* Return 0 on success or positive error */ +STATIC int +xfs_bulkstat_one_fmt( + void __user *ubuffer, + int ubsize, + int *ubused, + const xfs_bstat_t *buffer) +{ + if (ubsize < sizeof(*buffer)) + return XFS_ERROR(ENOMEM); + if (copy_to_user(ubuffer, buffer, sizeof(*buffer))) + return XFS_ERROR(EFAULT); + if (ubused) + *ubused = sizeof(*buffer); return 0; } @@ -202,13 +219,13 @@ xfs_bulkstat_one_dinode( * Return stat information for one inode. * Return 0 if ok, else errno. */ -int /* error status */ -xfs_bulkstat_one( +int /* error status */ +xfs_bulkstat_one_int( xfs_mount_t *mp, /* mount point for filesystem */ xfs_ino_t ino, /* inode number to get data for */ void __user *buffer, /* buffer to place output in */ int ubsize, /* size of buffer */ - void *private_data, /* my private data */ + bulkstat_one_fmt_pf formatter, /* formatter, copy to user */ xfs_daddr_t bno, /* starting bno of inode cluster */ int *ubused, /* bytes used by me */ void *dibuff, /* on-disk inode buffer */ @@ -219,17 +236,10 @@ xfs_bulkstat_one( xfs_dinode_t *dip; /* dinode inode pointer */ dip = (xfs_dinode_t *)dibuff; + *stat = BULKSTAT_RV_NOTHING; - if (!buffer || ino == mp->m_sb.sb_rbmino || ino == mp->m_sb.sb_rsumino || - (XFS_SB_VERSION_HASQUOTA(&mp->m_sb) && - (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino))) { - *stat = BULKSTAT_RV_NOTHING; + if (!buffer || xfs_internal_inum(mp, ino)) return XFS_ERROR(EINVAL); - } - if (ubsize < sizeof(*buf)) { - *stat = BULKSTAT_RV_NOTHING; - return XFS_ERROR(ENOMEM); - } buf = kmem_alloc(sizeof(*buf), KM_SLEEP); @@ -244,21 +254,83 @@ xfs_bulkstat_one( xfs_bulkstat_one_dinode(mp, ino, dip, buf); } - if (copy_to_user(buffer, buf, sizeof(*buf))) { - *stat = BULKSTAT_RV_NOTHING; - error = EFAULT; + error = formatter(buffer, ubsize, ubused, buf); + if (error) goto out_free; - } *stat = BULKSTAT_RV_DIDONE; - if (ubused) - *ubused = sizeof(*buf); out_free: - kmem_free(buf, sizeof(*buf)); + kmem_free(buf); return error; } +int +xfs_bulkstat_one( + xfs_mount_t *mp, /* mount point for filesystem */ + xfs_ino_t ino, /* inode number to get data for */ + void __user *buffer, /* buffer to place output in */ + int ubsize, /* size of buffer */ + void *private_data, /* my private data */ + xfs_daddr_t bno, /* starting bno of inode cluster */ + int *ubused, /* bytes used by me */ + void *dibuff, /* on-disk inode buffer */ + int *stat) /* BULKSTAT_RV_... */ +{ + return xfs_bulkstat_one_int(mp, ino, buffer, ubsize, + xfs_bulkstat_one_fmt, bno, + ubused, dibuff, stat); +} + +/* + * Test to see whether we can use the ondisk inode directly, based + * on the given bulkstat flags, filling in dipp accordingly. + * Returns zero if the inode is dodgey. + */ +STATIC int +xfs_bulkstat_use_dinode( + xfs_mount_t *mp, + int flags, + xfs_buf_t *bp, + int clustidx, + xfs_dinode_t **dipp) +{ + xfs_dinode_t *dip; + unsigned int aformat; + + *dipp = NULL; + if (!bp || (flags & BULKSTAT_FG_IGET)) + return 1; + dip = (xfs_dinode_t *) + xfs_buf_offset(bp, clustidx << mp->m_sb.sb_inodelog); + /* + * Check the buffer containing the on-disk inode for di_mode == 0. + * This is to prevent xfs_bulkstat from picking up just reclaimed + * inodes that have their in-core state initialized but not flushed + * to disk yet. This is a temporary hack that would require a proper + * fix in the future. + */ + if (be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC || + !XFS_DINODE_GOOD_VERSION(dip->di_version) || + !dip->di_mode) + return 0; + if (flags & BULKSTAT_FG_QUICK) { + *dipp = dip; + return 1; + } + /* BULKSTAT_FG_INLINE: if attr fork is local, or not there, use it */ + aformat = dip->di_aformat; + if ((XFS_DFORK_Q(dip) == 0) || + (aformat == XFS_DINODE_FMT_LOCAL) || + (aformat == XFS_DINODE_FMT_EXTENTS && !dip->di_anextents)) { + *dipp = dip; + return 1; + } + return 1; +} + +#define XFS_BULKSTAT_UBLEFT(ubleft) ((ubleft) >= statstruct_size) + /* * Return stat information in bulk (by-inode) for the filesystem. */ @@ -272,7 +344,7 @@ xfs_bulkstat( size_t statstruct_size, /* sizeof struct filling */ char __user *ubuffer, /* buffer with inode stats */ int flags, /* defined in xfs_itable.h */ - int *done) /* 1 if there're more stats to get */ + int *done) /* 1 if there are more stats to get */ { xfs_agblock_t agbno=0;/* allocation group block number */ xfs_buf_t *agbp; /* agi header buffer */ @@ -286,16 +358,14 @@ xfs_bulkstat( int end_of_ag; /* set if we've seen the ag end */ int error; /* error code */ int fmterror;/* bulkstat formatter result */ - __int32_t gcnt; /* current btree rec's count */ - xfs_inofree_t gfree; /* current btree rec's free mask */ - xfs_agino_t gino; /* current btree rec's start inode */ int i; /* loop index */ int icount; /* count of inodes good in irbuf */ + size_t irbsize; /* size of irec buffer in bytes */ xfs_ino_t ino; /* inode number (filesystem) */ - xfs_inobt_rec_t *irbp; /* current irec buffer pointer */ - xfs_inobt_rec_t *irbuf; /* start of irec buffer */ - xfs_inobt_rec_t *irbufend; /* end of good irec buffer entries */ - xfs_ino_t lastino=0; /* last inode number returned */ + xfs_inobt_rec_incore_t *irbp; /* current irec buffer pointer */ + xfs_inobt_rec_incore_t *irbuf; /* start of irec buffer */ + xfs_inobt_rec_incore_t *irbufend; /* end of good irec buffer entries */ + xfs_ino_t lastino; /* last inode number returned */ int nbcluster; /* # of blocks in a cluster */ int nicluster; /* # of inodes in a cluster */ int nimask; /* mask for inode clusters */ @@ -309,12 +379,12 @@ xfs_bulkstat( int ubused; /* bytes used by formatter */ xfs_buf_t *bp; /* ptr to on-disk inode cluster buf */ xfs_dinode_t *dip; /* ptr into bp for specific inode */ - xfs_inode_t *ip; /* ptr to in-core inode struct */ /* * Get the last inode value, see if there's nothing to do. */ ino = (xfs_ino_t)*lastinop; + lastino = ino; dip = NULL; agno = XFS_INO_TO_AGNO(mp, ino); agino = XFS_INO_TO_AGINO(mp, ino); @@ -324,6 +394,9 @@ xfs_bulkstat( *ubcountp = 0; return 0; } + if (!ubcountp || *ubcountp <= 0) { + return EINVAL; + } ubcount = *ubcountp; /* statstruct's */ ubleft = ubcount * statstruct_size; /* bytes */ *ubcountp = ubelem = 0; @@ -335,28 +408,17 @@ xfs_bulkstat( (XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog); nimask = ~(nicluster - 1); nbcluster = nicluster >> mp->m_sb.sb_inopblog; - /* - * Lock down the user's buffer. If a buffer was not sent, as in the case - * disk quota code calls here, we skip this. - */ - if (ubuffer && - (error = useracc(ubuffer, ubcount * statstruct_size, - (B_READ|B_PHYS), NULL))) { - return error; - } - /* - * Allocate a page-sized buffer for inode btree records. - * We could try allocating something smaller, but for normal - * calls we'll always (potentially) need the whole page. - */ - irbuf = kmem_alloc(NBPC, KM_SLEEP); - nirbuf = NBPC / sizeof(*irbuf); + irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4, + KM_SLEEP | KM_MAYFAIL | KM_LARGE); + nirbuf = irbsize / sizeof(*irbuf); + /* * Loop over the allocation groups, starting from the last * inode returned; 0 means start of the allocation group. */ rval = 0; - while (ubleft >= statstruct_size && agno < mp->m_sb.sb_agcount) { + while (XFS_BULKSTAT_UBLEFT(ubleft) && agno < mp->m_sb.sb_agcount) { + cond_resched(); bp = NULL; down_read(&mp->m_peraglock); error = xfs_ialloc_read_agi(mp, NULL, agno, &agbp); @@ -373,8 +435,7 @@ xfs_bulkstat( /* * Allocate and initialize a btree cursor for ialloc btree. */ - cur = xfs_btree_init_cursor(mp, NULL, agbp, agno, XFS_BTNUM_INO, - (xfs_inode_t *)0, 0); + cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno); irbp = irbuf; irbufend = irbuf + nirbuf; end_of_ag = 0; @@ -383,40 +444,43 @@ xfs_bulkstat( * we need to get the remainder of the chunk we're in. */ if (agino > 0) { + xfs_inobt_rec_incore_t r; + /* * Lookup the inode chunk that this inode lives in. */ - error = xfs_inobt_lookup_le(cur, agino, 0, 0, &tmp); + error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, + &tmp); if (!error && /* no I/O error */ tmp && /* lookup succeeded */ /* got the record, should always work */ - !(error = xfs_inobt_get_rec(cur, &gino, &gcnt, - &gfree, &i)) && + !(error = xfs_inobt_get_rec(cur, &r, &i)) && i == 1 && /* this is the right chunk */ - agino < gino + XFS_INODES_PER_CHUNK && + agino < r.ir_startino + XFS_INODES_PER_CHUNK && /* lastino was not last in chunk */ - (chunkidx = agino - gino + 1) < + (chunkidx = agino - r.ir_startino + 1) < XFS_INODES_PER_CHUNK && /* there are some left allocated */ - XFS_INOBT_MASKN(chunkidx, - XFS_INODES_PER_CHUNK - chunkidx) & ~gfree) { + xfs_inobt_maskn(chunkidx, + XFS_INODES_PER_CHUNK - chunkidx) & + ~r.ir_free) { /* * Grab the chunk record. Mark all the * uninteresting inodes (because they're * before our start point) free. */ for (i = 0; i < chunkidx; i++) { - if (XFS_INOBT_MASK(i) & ~gfree) - gcnt++; + if (XFS_INOBT_MASK(i) & ~r.ir_free) + r.ir_freecount++; } - gfree |= XFS_INOBT_MASKN(0, chunkidx); - INT_SET(irbp->ir_startino, ARCH_CONVERT, gino); - INT_SET(irbp->ir_freecount, ARCH_CONVERT, gcnt); - INT_SET(irbp->ir_free, ARCH_CONVERT, gfree); + r.ir_free |= xfs_inobt_maskn(0, chunkidx); + irbp->ir_startino = r.ir_startino; + irbp->ir_freecount = r.ir_freecount; + irbp->ir_free = r.ir_free; irbp++; - agino = gino + XFS_INODES_PER_CHUNK; - icount = XFS_INODES_PER_CHUNK - gcnt; + agino = r.ir_startino + XFS_INODES_PER_CHUNK; + icount = XFS_INODES_PER_CHUNK - r.ir_freecount; } else { /* * If any of those tests failed, bump the @@ -429,12 +493,12 @@ xfs_bulkstat( * In any case, increment to the next record. */ if (!error) - error = xfs_inobt_increment(cur, 0, &tmp); + error = xfs_btree_increment(cur, 0, &tmp); } else { /* * Start of ag. Lookup the first inode chunk. */ - error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &tmp); + error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &tmp); icount = 0; } /* @@ -442,6 +506,8 @@ xfs_bulkstat( * until we run out of inodes or space in the buffer. */ while (irbp < irbufend && icount < ubcount) { + xfs_inobt_rec_incore_t r; + /* * Loop as long as we're unable to read the * inode btree. @@ -451,35 +517,57 @@ xfs_bulkstat( if (XFS_AGINO_TO_AGBNO(mp, agino) >= be32_to_cpu(agi->agi_length)) break; - error = xfs_inobt_lookup_ge(cur, agino, 0, 0, - &tmp); + error = xfs_inobt_lookup(cur, agino, + XFS_LOOKUP_GE, &tmp); + cond_resched(); } /* * If ran off the end of the ag either with an error, * or the normal way, set end and stop collecting. */ - if (error || - (error = xfs_inobt_get_rec(cur, &gino, &gcnt, - &gfree, &i)) || - i == 0) { + if (error) { end_of_ag = 1; break; } + + error = xfs_inobt_get_rec(cur, &r, &i); + if (error || i == 0) { + end_of_ag = 1; + break; + } + /* * If this chunk has any allocated inodes, save it. + * Also start read-ahead now for this chunk. */ - if (gcnt < XFS_INODES_PER_CHUNK) { - INT_SET(irbp->ir_startino, ARCH_CONVERT, gino); - INT_SET(irbp->ir_freecount, ARCH_CONVERT, gcnt); - INT_SET(irbp->ir_free, ARCH_CONVERT, gfree); + if (r.ir_freecount < XFS_INODES_PER_CHUNK) { + /* + * Loop over all clusters in the next chunk. + * Do a readahead if there are any allocated + * inodes in that cluster. + */ + agbno = XFS_AGINO_TO_AGBNO(mp, r.ir_startino); + for (chunkidx = 0; + chunkidx < XFS_INODES_PER_CHUNK; + chunkidx += nicluster, + agbno += nbcluster) { + if (xfs_inobt_maskn(chunkidx, nicluster) + & ~r.ir_free) + xfs_btree_reada_bufs(mp, agno, + agbno, nbcluster); + } + irbp->ir_startino = r.ir_startino; + irbp->ir_freecount = r.ir_freecount; + irbp->ir_free = r.ir_free; irbp++; - icount += XFS_INODES_PER_CHUNK - gcnt; + icount += XFS_INODES_PER_CHUNK - r.ir_freecount; } /* * Set agino to after this chunk and bump the cursor. */ - agino = gino + XFS_INODES_PER_CHUNK; - error = xfs_inobt_increment(cur, 0, &tmp); + agino = r.ir_startino + XFS_INODES_PER_CHUNK; + error = xfs_btree_increment(cur, 0, &tmp); + cond_resched(); } /* * Drop the btree buffers and the agi buffer. @@ -493,35 +581,13 @@ xfs_bulkstat( */ irbufend = irbp; for (irbp = irbuf; - irbp < irbufend && ubleft >= statstruct_size; irbp++) { - /* - * Read-ahead the next chunk's worth of inodes. - */ - if (&irbp[1] < irbufend) { - /* - * Loop over all clusters in the next chunk. - * Do a readahead if there are any allocated - * inodes in that cluster. - */ - for (agbno = XFS_AGINO_TO_AGBNO(mp, - INT_GET(irbp[1].ir_startino, ARCH_CONVERT)), - chunkidx = 0; - chunkidx < XFS_INODES_PER_CHUNK; - chunkidx += nicluster, - agbno += nbcluster) { - if (XFS_INOBT_MASKN(chunkidx, - nicluster) & - ~(INT_GET(irbp[1].ir_free, ARCH_CONVERT))) - xfs_btree_reada_bufs(mp, agno, - agbno, nbcluster); - } - } + irbp < irbufend && XFS_BULKSTAT_UBLEFT(ubleft); irbp++) { /* * Now process this chunk of inodes. */ - for (agino = INT_GET(irbp->ir_startino, ARCH_CONVERT), chunkidx = 0, clustidx = 0; - ubleft > 0 && - INT_GET(irbp->ir_freecount, ARCH_CONVERT) < XFS_INODES_PER_CHUNK; + for (agino = irbp->ir_startino, chunkidx = clustidx = 0; + XFS_BULKSTAT_UBLEFT(ubleft) && + irbp->ir_freecount < XFS_INODES_PER_CHUNK; chunkidx++, clustidx++, agino++) { ASSERT(chunkidx < XFS_INODES_PER_CHUNK); /* @@ -529,7 +595,7 @@ xfs_bulkstat( * first inode of the cluster. * * Careful with clustidx. There can be - * multple clusters per chunk, a single + * multiple clusters per chunk, a single * cluster per chunk or a cluster that has * inodes represented from several different * chunks (if blocksize is large). @@ -541,11 +607,14 @@ xfs_bulkstat( */ if ((chunkidx & (nicluster - 1)) == 0) { agbno = XFS_AGINO_TO_AGBNO(mp, - INT_GET(irbp->ir_startino, ARCH_CONVERT)) + + irbp->ir_startino) + ((chunkidx & nimask) >> mp->m_sb.sb_inopblog); - if (flags & BULKSTAT_FG_QUICK) { + if (flags & (BULKSTAT_FG_QUICK | + BULKSTAT_FG_INLINE)) { + int offset; + ino = XFS_AGINO_TO_INO(mp, agno, agino); bno = XFS_AGB_TO_DADDR(mp, agno, @@ -554,19 +623,15 @@ xfs_bulkstat( /* * Get the inode cluster buffer */ - ASSERT(xfs_inode_zone != NULL); - ip = kmem_zone_zalloc(xfs_inode_zone, - KM_SLEEP); - ip->i_ino = ino; - ip->i_mount = mp; if (bp) xfs_buf_relse(bp); - error = xfs_itobp(mp, NULL, ip, - &dip, &bp, bno, - XFS_IMAP_BULKSTAT); + + error = xfs_inotobp(mp, NULL, ino, &dip, + &bp, &offset, + XFS_IGET_BULKSTAT); + if (!error) - clustidx = ip->i_boffset / mp->m_sb.sb_inodesize; - kmem_zone_free(xfs_inode_zone, ip); + clustidx = offset / mp->m_sb.sb_inodesize; if (XFS_TEST_ERROR(error != 0, mp, XFS_ERRTAG_BULKSTAT_READ_CHUNK, XFS_RANDOM_BULKSTAT_READ_CHUNK)) { @@ -577,33 +642,41 @@ xfs_bulkstat( } } } + ino = XFS_AGINO_TO_INO(mp, agno, agino); + bno = XFS_AGB_TO_DADDR(mp, agno, agbno); /* * Skip if this inode is free. */ - if (XFS_INOBT_MASK(chunkidx) & INT_GET(irbp->ir_free, ARCH_CONVERT)) + if (XFS_INOBT_MASK(chunkidx) & irbp->ir_free) { + lastino = ino; continue; + } /* * Count used inodes as free so we can tell * when the chunk is used up. */ - INT_MOD(irbp->ir_freecount, ARCH_CONVERT, +1); - ino = XFS_AGINO_TO_INO(mp, agno, agino); - bno = XFS_AGB_TO_DADDR(mp, agno, agbno); - if (flags & BULKSTAT_FG_QUICK) { - dip = (xfs_dinode_t *)xfs_buf_offset(bp, - (clustidx << mp->m_sb.sb_inodelog)); - - if (INT_GET(dip->di_core.di_magic, ARCH_CONVERT) - != XFS_DINODE_MAGIC - || !XFS_DINODE_GOOD_VERSION( - INT_GET(dip->di_core.di_version, ARCH_CONVERT))) - continue; + irbp->ir_freecount++; + if (!xfs_bulkstat_use_dinode(mp, flags, bp, + clustidx, &dip)) { + lastino = ino; + continue; + } + /* + * If we need to do an iget, cannot hold bp. + * Drop it, until starting the next cluster. + */ + if ((flags & BULKSTAT_FG_INLINE) && !dip) { + if (bp) + xfs_buf_relse(bp); + bp = NULL; } /* * Get the inode and fill in a single buffer. * BULKSTAT_FG_QUICK uses dip to fill it in. * BULKSTAT_FG_IGET uses igets. + * BULKSTAT_FG_INLINE uses dip if we have an + * inline attr fork, else igets. * See: xfs_bulkstat_one & xfs_dm_bulkstat_one. * This is also used to count inodes/blks, etc * in xfs_qm_quotacheck. @@ -613,8 +686,13 @@ xfs_bulkstat( ubleft, private_data, bno, &ubused, dip, &fmterror); if (fmterror == BULKSTAT_RV_NOTHING) { - if (error == ENOMEM) + if (error && error != ENOENT && + error != EINVAL) { ubleft = 0; + rval = error; + break; + } + lastino = ino; continue; } if (fmterror == BULKSTAT_RV_GIVEUP) { @@ -629,6 +707,8 @@ xfs_bulkstat( ubelem++; lastino = ino; } + + cond_resched(); } if (bp) @@ -637,7 +717,7 @@ xfs_bulkstat( /* * Set up for the next loop iteration. */ - if (ubleft > 0) { + if (XFS_BULKSTAT_UBLEFT(ubleft)) { if (end_of_ag) { agno++; agino = 0; @@ -649,10 +729,13 @@ xfs_bulkstat( /* * Done, we're either out of filesystem or space to put the data. */ - kmem_free(irbuf, NBPC); - if (ubuffer) - unuseracc(ubuffer, ubcount * statstruct_size, (B_READ|B_PHYS)); + kmem_free(irbuf); *ubcountp = ubelem; + /* + * Found some inodes, return them now and return the error next time. + */ + if (ubelem) + rval = 0; if (agno >= mp->m_sb.sb_agcount) { /* * If we ran out of filesystem, mark lastino as off @@ -676,7 +759,7 @@ xfs_bulkstat_single( xfs_mount_t *mp, /* mount point for filesystem */ xfs_ino_t *lastinop, /* inode to return */ char __user *buffer, /* buffer with inode stats */ - int *done) /* 1 if there're more stats to get */ + int *done) /* 1 if there are more stats to get */ { int count; /* count value for bulkstat call */ int error; /* return value */ @@ -716,6 +799,19 @@ xfs_bulkstat_single( return 0; } +int +xfs_inumbers_fmt( + void __user *ubuffer, /* buffer to write to */ + const xfs_inogrp_t *buffer, /* buffer to read from */ + long count, /* # of elements to read */ + long *written) /* # of bytes written */ +{ + if (copy_to_user(ubuffer, buffer, count * sizeof(*buffer))) + return -EFAULT; + *written = count * sizeof(*buffer); + return 0; +} + /* * Return inode number table for the filesystem. */ @@ -724,7 +820,8 @@ xfs_inumbers( xfs_mount_t *mp, /* mount point for filesystem */ xfs_ino_t *lastino, /* last inode returned */ int *count, /* size of buffer/count returned */ - xfs_inogrp_t __user *ubuffer)/* buffer with inode descriptions */ + void __user *ubuffer,/* buffer with inode descriptions */ + inumbers_fmt_pf formatter) { xfs_buf_t *agbp; xfs_agino_t agino; @@ -734,9 +831,7 @@ xfs_inumbers( int bufidx; xfs_btree_cur_t *cur; int error; - __int32_t gcnt; - xfs_inofree_t gfree; - xfs_agino_t gino; + xfs_inobt_rec_incore_t r; int i; xfs_ino_t ino; int left; @@ -747,7 +842,7 @@ xfs_inumbers( agino = XFS_INO_TO_AGINO(mp, ino); left = *count; *count = 0; - bcount = MIN(left, (int)(NBPP / sizeof(*buffer))); + bcount = MIN(left, (int)(PAGE_SIZE / sizeof(*buffer))); buffer = kmem_alloc(bcount * sizeof(*buffer), KM_SLEEP); error = bufidx = 0; cur = NULL; @@ -768,16 +863,16 @@ xfs_inumbers( agino = 0; continue; } - cur = xfs_btree_init_cursor(mp, NULL, agbp, agno, - XFS_BTNUM_INO, (xfs_inode_t *)0, 0); - error = xfs_inobt_lookup_ge(cur, agino, 0, 0, &tmp); + cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno); + error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_GE, + &tmp); if (error) { xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); cur = NULL; xfs_buf_relse(agbp); agbp = NULL; /* - * Move up the the last inode in the current + * Move up the last inode in the current * chunk. The lookup_ge will always get * us the first inode in the next chunk. */ @@ -785,9 +880,8 @@ xfs_inumbers( continue; } } - if ((error = xfs_inobt_get_rec(cur, &gino, &gcnt, &gfree, - &i)) || - i == 0) { + error = xfs_inobt_get_rec(cur, &r, &i); + if (error || i == 0) { xfs_buf_relse(agbp); agbp = NULL; xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); @@ -796,24 +890,26 @@ xfs_inumbers( agino = 0; continue; } - agino = gino + XFS_INODES_PER_CHUNK - 1; - buffer[bufidx].xi_startino = XFS_AGINO_TO_INO(mp, agno, gino); - buffer[bufidx].xi_alloccount = XFS_INODES_PER_CHUNK - gcnt; - buffer[bufidx].xi_allocmask = ~gfree; + agino = r.ir_startino + XFS_INODES_PER_CHUNK - 1; + buffer[bufidx].xi_startino = + XFS_AGINO_TO_INO(mp, agno, r.ir_startino); + buffer[bufidx].xi_alloccount = + XFS_INODES_PER_CHUNK - r.ir_freecount; + buffer[bufidx].xi_allocmask = ~r.ir_free; bufidx++; left--; if (bufidx == bcount) { - if (copy_to_user(ubuffer, buffer, - bufidx * sizeof(*buffer))) { + long written; + if (formatter(ubuffer, buffer, bufidx, &written)) { error = XFS_ERROR(EFAULT); break; } - ubuffer += bufidx; + ubuffer += written; *count += bufidx; bufidx = 0; } if (left) { - error = xfs_inobt_increment(cur, 0, &tmp); + error = xfs_btree_increment(cur, 0, &tmp); if (error) { xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); cur = NULL; @@ -830,15 +926,15 @@ xfs_inumbers( } if (!error) { if (bufidx) { - if (copy_to_user(ubuffer, buffer, - bufidx * sizeof(*buffer))) + long written; + if (formatter(ubuffer, buffer, bufidx, &written)) error = XFS_ERROR(EFAULT); else *count += bufidx; } *lastino = XFS_AGINO_TO_INO(mp, agno, agino); } - kmem_free(buffer, bcount * sizeof(*buffer)); + kmem_free(buffer); if (cur) xfs_btree_del_cursor(cur, (error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR));