fix rotational layout code to work with disks with nsect % bsize != 0

[unix-history] / usr / src / sys / ufs / ffs / fs.h

diff --git a/usr/src/sys/ufs/ffs/fs.h b/usr/src/sys/ufs/ffs/fs.h
index d87b42b..00107aa 100644 (file)
--- a/usr/src/sys/ufs/ffs/fs.h
+++ b/usr/src/sys/ufs/ffs/fs.h
@@ -1,6 +1,6 @@
 /* Copyright (c) 1981 Regents of the University of California */
 
 /* Copyright (c) 1981 Regents of the University of California */
 

-/*     fs.h    1.4     %G%     */
+/*     fs.h    1.11    %G%     */

 
 /*
  * Each disk drive contains some number of file systems.
 
 /*
  * Each disk drive contains some number of file systems.


@@ -14,206 +14,344 @@
  * super-block data does not change, so the copies need not be
  * referenced further unless disaster strikes.
  *
  * super-block data does not change, so the copies need not be
  * referenced further unless disaster strikes.
  *

- * For file system fs and a cylinder group number cg:
- *     [BBLOCK]        Boot sector and bad block information
- *     [SBLOCK]        Super-block
- *     [CBLOCK]        Cylinder group block
- *     [IBLOCK..IBLOCK+fs.fs_ipg/INOPB)
- *                     Inode blocks
- *     [IBLOCK+fs.fs_ipg/INOPB..fs.fs_fpg/FRAG)
- *                     Data blocks
- * The beginning of data blocks for cg in fs is also given by cgdmin(cg,fs).
- */
-#define        BBLOCK  ((daddr_t)(0*FRAG))
-#define        SBLOCK  ((daddr_t)(1*FRAG))
-#define        CBLOCK  ((daddr_t)(2*FRAG))
-#define        IBLOCK  ((daddr_t)(3*FRAG))
-
-/*
- * Addresses stored in inodes are capable of addressing fragments of `blocks.'
- * File system blocks of size BSIZE can be broken into FRAG pieces,
- * each of which is addressible; these pieces may be sectors, or some
- * multiple of a sector size (e.g. 1k byte units).
+ * For file system fs, the offsets of the various blocks of interest
+ * are given in the super block as:
+ *     [fs->fs_bblkno]         Boot sector
+ *     [fs->fs_sblkno]         Super-block
+ *     [fs->fs_cblkno]         Cylinder group block
+ *     [fs->fs_iblkno]         Inode blocks
+ *     [fs->fs_dblkno]         Data blocks
+ * The beginning of cylinder group cg in fs, is given by
+ * the ``cgbase(cg, fs)'' macro.

  *
  *

- * Large files consist of exclusively large (BSIZE) data blocks.  To avoid
- * undue fragmentation, the last data block of a small file may be
- * allocated as only as many pieces
- * of a large block as are necessary.  The file system format retains
- * only a single pointer to such a fragment, which is a piece of a single
- * BSIZE block which has been divided.  The size of such a fragment is
- * determinable from information in the inode.
+ * The first boot and super blocks are given in absolute disk addresses.
+ */
+#define BBSIZE         1024
+#define SBSIZE         8192
+#define        BBLOCK          ((daddr_t)(0))
+#define        SBLOCK          ((daddr_t)(BBLOCK + BBSIZE / DEV_BSIZE))
+
+/*
+ * Addresses stored in inodes are capable of addressing fragments
+ * of `blocks'. File system blocks of at most size MAXBSIZE can 
+ * be optionally broken into 2, 4, or 8 pieces, each of which is
+ * addressible; these pieces may be DEV_BSIZE, or some multiple of
+ * a DEV_BSIZE unit.
+ *
+ * Large files consist of exclusively large data blocks.  To avoid
+ * undue wasted disk space, the last data block of a small file may be
+ * allocated as only as many fragments of a large block as are
+ * necessary.  The file system format retains only a single pointer
+ * to such a fragment, which is a piece of a single large block that
+ * has been divided.  The size of such a fragment is determinable from
+ * information in the inode, using the ``blksize(fs, ip, lbn)'' macro.

  *
  * The file system records space availability at the fragment level;
  * to determine block availability, aligned fragments are examined.
  *
  *
  * The file system records space availability at the fragment level;
  * to determine block availability, aligned fragments are examined.
  *

+ * The root inode is the root of the file system.
+ * Inode 0 can't be used for normal purposes and
+ * historically bad blocks were linked to inode 1,
+ * thus the root inode is 2. (inode 1 is no longer used for
+ * this purpose, however numerous dump tapes make this
+ * assumption, so we are stuck with it)
+ * The lost+found directory is given the next available
+ * inode when it is created by ``mkfs''.
+ */
+#define        ROOTINO         ((ino_t)2)      /* i number of all roots */
+#define LOSTFOUNDINO   (ROOTINO + 1)
+
+/*
+ * MINFREE gives the minimum acceptable percentage of file system
+ * blocks which may be free. If the freelist drops below this level
+ * only the superuser may continue to allocate blocks. This may
+ * be set to 0 if no reserve of free blocks is deemed necessary,
+ * however severe performance degredations will be observed if the
+ * file system is run at greater than 90% full; thus the default
+ * value of fs_minfree is 10%.
+ *
+ * Empirically the best trade-off between block fragmentation and
+ * overall disk utilization at a loading of 90% comes with a
+ * fragmentation of 4, thus the default fragment size is a fourth
+ * of the block size.
+ */
+#define MINFREE                10
+#define DESFRAG                4
+
+/*
+ * Under current technology, most 300MB disks have 32 sectors and
+ * 19 tracks, thus these are the defaults used for fs_nsect and 
+ * fs_ntrak respectively.
+ */
+#define DFLNSECT       32
+#define DFLNTRAK       19
+
+/*
+ * Cylinder group related limits.
+ *

  * For each cylinder we keep track of the availability of blocks at different
  * rotational positions, so that we can lay out the data to be picked
  * up with minimum rotational latency.  NRPOS is the number of rotational
  * positions which we distinguish.  With NRPOS 8 the resolution of our
  * summary information is 2ms for a typical 3600 rpm drive.
  * For each cylinder we keep track of the availability of blocks at different
  * rotational positions, so that we can lay out the data to be picked
  * up with minimum rotational latency.  NRPOS is the number of rotational
  * positions which we distinguish.  With NRPOS 8 the resolution of our
  * summary information is 2ms for a typical 3600 rpm drive.

+ *
+ * ROTDELAY gives the minimum number of milliseconds to initiate
+ * another disk transfer on the same cylinder. It is used in
+ * determining the rotationally optimal layout for disk blocks
+ * within a file; the default of fs_rotdelay is 2ms.
+ */
+#define        NRPOS           8       /* number distinct rotational positions */
+#define ROTDELAY       2
+
+/*
+ * Each file system has a number of inodes statically allocated.
+ * We allocate one inode slot per NBPI bytes, expecting this
+ * to be far more than we will ever need.
+ *
+ * MAXIPG bounds the number of inodes per cylinder group, and
+ * is needed only to keep the structure simpler by having the
+ * only a single variable size element (the free bit map).
+ *
+ * N.B.: MAXIPG must be a multiple of INOPB(fs).
+ */
+#define        NBPI            2048
+#define        MAXIPG          2048    /* max number inodes/cyl group */
+
+/*
+ * MINBSIZE is the smallest allowable block size.
+ * In order to insure that it is possible to create files of size
+ * 2^32 with only two levels of indirection, MINBSIZE is set to 4096.
+ * MINBSIZE must be big enough to hold a cylinder group block,
+ * thus changes to (struct cg) must keep its size within MINBSIZE.
+ * MAXCPG is limited only to dimension an array in (struct cg);
+ * it can be made larger as long as that structures size remains
+ * within the bounds dictated by MINBSIZE.
+ * Note that super blocks are always of size MAXBSIZE,
+ * and that MAXBSIZE must be >= MINBSIZE.

  */
  */

-#define        NRPOS   8               /* number distinct rotational positions */
+#define MINBSIZE       4096
+#define        DESCPG          16      /* desired fs_cpg */
+#define        MAXCPG          32      /* maximum fs_cpg */

 
 /*
 
 /*

- * Information per cylinder group summarized in blocks allocated
+ * Per cylinder group information; summarized in blocks allocated

  * from first cylinder group data blocks.  These blocks have to be
  * from first cylinder group data blocks.  These blocks have to be

- * specially noticed in mkfs, icheck and fsck.
+ * read in from fs_csaddr (size fs_cssize) in addition to the
+ * super block.
+ *
+ * N.B. sizeof(struct csum) must be a power of two in order for
+ * the ``fs_cs'' macro to work (see below).

  */
 struct csum {
  */
 struct csum {

-       short   cs_ndir;        /* number of directories */
-       short   cs_nbfree;      /* number of free blocks */
-       short   cs_nifree;      /* number of free inodes */
-       short   cs_xx;          /* for later use... */
+       long    cs_ndir;        /* number of directories */
+       long    cs_nbfree;      /* number of free blocks */
+       long    cs_nifree;      /* number of free inodes */
+       long    cs_nffree;      /* number of free frags */

 };
 };

-#define        cssize(fs)      ((fs)->fs_ncg*sizeof(struct csum))
-#define        csaddr(fs)      (cgdmin(0, fs))

 
 

-/*
- * Each file system has a number of inodes statically allocated.
- * We allocate one inode slot per NBPI data bytes, expecting this
- * to be far more than we will ever need.  Actually, the directory
- * structure has inode numbers kept in 16 bits, so no more than
- * 65K inodes are possible, and this usually cuts off well above
- * the number suggested by NBPI.
- * 
- * THE DIRECTORY STRUCTURE SHOULD BE CHANGED SOON TO ALLOW
- * LARGER INODE NUMBERS (SEE DIR.H).
- */
-#define        NBPI    2048
-
-#define        DESCPG  16                      /* desired fs_cpg */
-#define        MAXCPG  32                      /* maximum fs_cpg */
-/* MAXCPG is limited only to dimension an array in (struct cg); */
-/* it can be made larger as long as that structures size remains sane. */
- 

 /*
  * Super block for a file system.
 /*
  * Super block for a file system.

- *
- * The super block is nominally located at disk block 1 although
- * this is naive due to bad blocks.  Inode 0 can't be used for normal
- * purposes, thus the root inode is inode 1.

  */
  */

-#define        ROOTINO ((ino_t)1)      /* i number of all roots */
-

 #define        FS_MAGIC        0x110854
 struct fs
 {
        long    fs_magic;               /* magic number */
 #define        FS_MAGIC        0x110854
 struct fs
 {
        long    fs_magic;               /* magic number */

-       daddr_t fs_sblkno;              /* offset of super-block in filesys */
+       daddr_t fs_bblkno;              /* abs addr of boot-block in filesys */
+       daddr_t fs_sblkno;              /* abs addr of super-block in filesys */
+       daddr_t fs_cblkno;              /* offset of cyl-block in filesys */
+       daddr_t fs_iblkno;              /* offset of inode-blocks in filesys */
+       daddr_t fs_dblkno;              /* offset of data-blocks in filesys */

        time_t  fs_time;                /* last time written */
        time_t  fs_time;                /* last time written */

-       daddr_t fs_size;                /* number of blocks in fs */
-       short   fs_ncg;                 /* number of cylinder groups */
+       long    fs_size;                /* number of blocks in fs */
+       long    fs_dsize;               /* number of data blocks in fs */
+       long    fs_ncg;                 /* number of cylinder groups */
+       long    fs_bsize;               /* size of basic blocks in fs */
+       long    fs_fsize;               /* size of frag blocks in fs */
+       short   fs_frag;                /* number of frags in a block in fs */
+       short   fs_minfree;             /* minimum percentage of free blocks */
+       short   fs_rotdelay;            /* num of ms for optimal next block */
+       short   fs_rps;                 /* disk revolutions per second */

 /* sizes determined by number of cylinder groups and their sizes */
 /* sizes determined by number of cylinder groups and their sizes */

-       short   fs_cssize;              /* size of cyl grp summary area */
-       short   fs_cgsize;              /* cylinder group size */
+       daddr_t fs_csaddr;              /* blk addr of cyl grp summary area */
+       long    fs_cssize;              /* size of cyl grp summary area */
+       long    fs_cgsize;              /* cylinder group size */

 /* these fields should be derived from the hardware */
        short   fs_ntrak;               /* tracks per cylinder */
        short   fs_nsect;               /* sectors per track */
 /* these fields should be derived from the hardware */
        short   fs_ntrak;               /* tracks per cylinder */
        short   fs_nsect;               /* sectors per track */

-       short   fs_spc;                 /* sectors per cylinder */
+       long    fs_spc;                 /* sectors per cylinder */

 /* this comes from the disk driver partitioning */
 /* this comes from the disk driver partitioning */

-       short   fs_ncyl;                /* cylinders in file system */
+       long    fs_ncyl;                /* cylinders in file system */

 /* these fields can be computed from the others */
        short   fs_cpg;                 /* cylinders per group */
 /* these fields can be computed from the others */
        short   fs_cpg;                 /* cylinders per group */

-       short   fs_fpg;                 /* blocks per group*FRAG */

        short   fs_ipg;                 /* inodes per group */
        short   fs_ipg;                 /* inodes per group */

-/* these fields must be re-computed after crashes */
-       daddr_t fs_nffree;              /* total free fragments */
-       daddr_t fs_nbfree;              /* free data blocks */
-       ino_t   fs_nifree;              /* total free inodes */
+       long    fs_fpg;                 /* blocks per group * fs_frag */
+/* this data must be re-computed after crashes */
+       struct  csum fs_cstotal;        /* cylinder summary information */

 /* these fields are cleared at mount time */
        char    fs_fmod;                /* super block modified flag */
        char    fs_ronly;               /* mounted read-only flag */
 /* these fields are cleared at mount time */
        char    fs_fmod;                /* super block modified flag */
        char    fs_ronly;               /* mounted read-only flag */

-       char    fs_fsmnt[32];           /* name mounted on */
+       char    fs_fsmnt[34];           /* name mounted on */

 /* these fields retain the current block allocation info */
 /* these fields retain the current block allocation info */

-       short   fs_cgrotor;             /* last cg searched */
+       long    fs_cgrotor;             /* last cg searched */

        struct  csum *fs_csp[NBUF];     /* list of fs_cs info buffers */
        struct  csum *fs_csp[NBUF];     /* list of fs_cs info buffers */

-       short   fs_postbl[NRPOS];       /* head of blocks for each rotation */
-       short   fs_rotbl[1];            /* list of blocks for each rotation */
+       short   fs_cpc;                 /* cyl per cycle in postbl */
+       short   fs_postbl[MAXCPG][NRPOS];/* head of blocks for each rotation */
+       u_char  fs_rotbl[1];            /* list of blocks for each rotation */
+/* actually longer */
+};
+
+/*
+ * convert cylinder group to base address of its global summary info.
+ *
+ * N.B. This macro assumes that sizeof(struct csum) is a power of two.
+ */
+#define fs_cs(fs, indx) \
+       fs_csp[(indx) / ((fs)->fs_bsize / sizeof(struct csum))] \
+       [(indx) % ((fs)->fs_bsize / sizeof(struct csum))]
+
+/*
+ * MAXBPC bounds the size of the rotational layout tables and
+ * is limited by the fact that the super block is of size SBSIZE.
+ * The size of these tables is INVERSELY proportional to the block
+ * size of the file system. It is aggravated by sector sizes that
+ * are not powers of two, as this increases the number of cylinders
+ * included before the rotational pattern repeats (fs_cpc).
+ * Its size is derived from the number of bytes remaining in (struct fs)
+ */
+#define        MAXBPC  (SBSIZE - sizeof (struct fs))
+
+/*
+ * Cylinder group block for a file system.
+ */
+#define        CG_MAGIC        0x092752
+struct cg {
+       long    cg_magic;               /* magic number */
+       time_t  cg_time;                /* time last written */
+       long    cg_cgx;                 /* we are the cgx'th cylinder group */
+       short   cg_ncyl;                /* number of cyl's this cg */
+       short   cg_niblk;               /* number of inode blocks this cg */
+       long    cg_ndblk;               /* number of data blocks this cg */
+       struct  csum cg_cs;             /* cylinder summary information */
+       long    cg_rotor;               /* position of last used block */
+       long    cg_frotor;              /* position of last used frag */
+       long    cg_irotor;              /* position of last used inode */
+       long    cg_frsum[MAXFRAG];      /* counts of available frags */
+       short   cg_b[MAXCPG][NRPOS];    /* positions of free blocks */
+       char    cg_iused[MAXIPG/NBBY];  /* used inode map */
+       char    cg_free[1];             /* free block map */

 /* actually longer */
 };
 /* actually longer */
 };

-#define fs_cs(indx) fs_csp[(indx) / (BSIZE / sizeof(struct csum))] \
-                         [(indx) % (BSIZE / sizeof(struct csum))]
+
+/*
+ * MAXBPG bounds the number of blocks of data per cylinder group,
+ * and is limited by the fact that cylinder groups are at most one block.
+ * Its size is derived from the size of blocks and the (struct cg) size,
+ * by the number of remaining bits.
+ */
+#define        MAXBPG(fs) \
+       (NBBY * ((fs)->fs_bsize - (sizeof (struct cg))) / (fs)->fs_frag)
+
+/*
+ * Turn file system block numbers into disk block addresses.
+ * This maps file system blocks to device size blocks.
+ */
+#define fsbtodb(fs, b) ((b) * ((fs)->fs_fsize / DEV_BSIZE))
+#define        dbtofsb(fs, b)  ((b) / ((fs)->fs_fsize / DEV_BSIZE))

 
 /*
  * Cylinder group macros to locate things in cylinder groups.
 
 /*
  * Cylinder group macros to locate things in cylinder groups.

+ *
+ * cylinder group to disk block address of spare boot block
+ * and super block
+ * Note that these are in absolute addresses, and can NOT
+ * in general be expressable in terms of file system addresses.

  */
  */

+#define        cgbblock(c,fs)  (fsbtodb(fs, cgbase(c,fs)) + (fs)->fs_bblkno)
+#define        cgsblock(c,fs)  (fsbtodb(fs, cgbase(c,fs)) + (fs)->fs_sblkno)

 
 

-/* cylinder group to disk block at very beginning */
-#define        cgbase(c,fs)    ((daddr_t)((fs)->fs_fpg*(c)))
+/*
+ * cylinder group to disk block at very beginning
+ */
+#define        cgbase(c,fs)    ((daddr_t)((fs)->fs_fpg * (c)))

 
 

-/* cylinder group to spare super block address */
-#define        cgsblock(c,fs)  \
-       (cgbase(c,fs) + SBLOCK)
+/*
+ * convert cylinder group to index of its cg block
+ */
+#define        cgtod(c,fs)     (cgbase(c,fs) + (fs)->fs_cblkno)

 
 

-/* convert cylinder group to index of its cg block */
-#define        cgtod(c,fs)     \
-       (cgbase(c,fs) + CBLOCK)
+/*
+ * give address of first inode block in cylinder group
+ */
+#define        cgimin(c,fs)    (cgbase(c,fs) + (fs)->fs_iblkno)

 
 

-/* give address of first inode block in cylinder group */
-#define        cgimin(c,fs)    \
-       (cgbase(c,fs) + IBLOCK)
+/*
+ * give address of first data block in cylinder group
+ */
+#define        cgdmin(c,fs)    (cgbase(c,fs) + (fs)->fs_dblkno)

 
 

-/* give address of first data block in cylinder group */
-#define        cgdmin(c,fs)    (cgimin(c,fs) + (fs)->fs_ipg / INOPF)
+/*
+ * turn inode number into cylinder group number
+ */
+#define        itog(x,fs)      ((x) / (fs)->fs_ipg)

 
 

-/* turn inode number into cylinder group number */
-#define        itog(x,fs)      ((x)/(fs)->fs_ipg)
+/*
+ * turn inode number into file system block address
+ */
+#define        itod(x,fs) \
+       ((daddr_t)(cgimin(itog(x,fs),fs) + \
+       (x) % (fs)->fs_ipg / INOPB(fs) * (fs)->fs_frag))

 
 

-/* turn inode number into disk block address */
-#define        itod(x,fs)      (cgimin(itog(x,fs),fs)+FRAG*((x)%(fs)->fs_ipg/INOPB))
+/*
+ * turn inode number into file system block offset
+ */
+#define        itoo(x,fs)      ((x) % INOPB(fs))

 
 

-/* turn inode number into disk block offset */
-#define        itoo(x)         ((x)%INOPB)
+/*
+ * give cylinder group number for a file system block
+ */
+#define        dtog(d,fs)      ((d) / (fs)->fs_fpg)

 
 

-/* give cylinder group number for a disk block */
-#define        dtog(d,fs)      ((d)/(fs)->fs_fpg)
+/*
+ * give cylinder group block number for a file system block
+ */
+#define        dtogd(d,fs)     ((d) % (fs)->fs_fpg)

 
 

-/* give cylinder group block number for a disk block */
-#define        dtogd(d,fs)     ((d)%(fs)->fs_fpg)
+/*
+ * compute the cylinder and rotational position of a cyl block addr
+ */
+#define cbtocylno(fs, bno) \
+       ((bno) * NSPF(fs) / (fs)->fs_spc)
+#define cbtorpos(fs, bno) \
+       ((bno) * NSPF(fs) % (fs)->fs_nsect * NRPOS / (fs)->fs_nsect)

 
 /*
 
 /*

- * Cylinder group related limits.
+ * determining the size of a file block in the file system

  */
  */

+#define blksize(fs, ip, lbn) \
+       (((lbn) >= NDADDR || (ip)->i_size >= ((lbn) + 1) * (fs)->fs_bsize) \
+               ? (fs)->fs_bsize \
+               : (roundup((ip)->i_size % (fs)->fs_bsize, (fs)->fs_fsize)))
+#define dblksize(fs, dip, lbn) \
+       (((lbn) >= NDADDR || (dip)->di_size >= ((lbn) + 1) * (fs)->fs_bsize) \
+               ? (fs)->fs_bsize \
+               : (roundup((dip)->di_size % (fs)->fs_bsize, (fs)->fs_fsize)))

 
 /*
 
 /*

- * MAXIPG bounds the number of inodes per cylinder group, and
- * is needed only to keep the structure simpler by having the
- * only a single variable size element (the free bit map).
- *
- * N.B.: MAXIPG must be a multiple of INOPB.
+ * number of disk sectors per block; assumes DEV_BSIZE byte sector size

  */
  */

-#define        MAXIPG  2048            /* max number inodes/cyl group */
+#define        NSPB(fs)        ((fs)->fs_bsize / DEV_BSIZE)
+#define        NSPF(fs)        ((fs)->fs_fsize / DEV_BSIZE)

 
 /*
 
 /*

- * MAXBPG bounds the number of blocks of data per cylinder group,
- * and is limited by the fact that cylinder groups are at most one block.
- * Its size is derived from the size of blocks and the (struct cg) size,
- * by the number of remaining bits.
+ * INOPB is the number of inodes in a secondary storage block

  */
  */

-#define        MAXBPG  (NBBY*(BSIZE-(sizeof (struct cg)))/FRAG)
+#define        INOPB(fs)       ((fs)->fs_bsize / sizeof (struct dinode))
+#define        INOPF(fs)       ((fs)->fs_fsize / sizeof (struct dinode))

 
 

-#define        CG_MAGIC        0x092752
-struct cg {
-       long    cg_magic;               /* magic number */
-       time_t  cg_time;                /* time last written */
-       short   cg_cgx;                 /* we are the cgx'th cylinder group */
-       short   cg_ncyl;                /* number of cyl's this cg */
-       short   cg_niblk;               /* number of inode blocks this cg */
-       short   cg_ndblk;               /* number of data blocks this cg */
-       short   cg_nifree;              /* free inodes */
-       short   cg_ndir;                /* allocated directories */
-       short   cg_nffree;              /* free block fragments */
-       short   cg_nbfree;              /* free blocks */
-       short   cg_rotor;               /* position of last used block */
-       short   cg_irotor;              /* position of last used inode */
-       short   cg_frotor;              /* position of last used frag */
-       short   cg_frsum[FRAG];         /* counts of available frags */
-       short   cg_b[MAXCPG][NRPOS];    /* positions of free blocks */
-       char    cg_iused[MAXIPG/NBBY];  /* used inode map */
-       char    cg_free[1];             /* free block map */
-/* actually longer */
-};
-#define        cgsize(fp)      (sizeof (struct cg) + ((fp)->fs_fpg+NBBY-1)/NBBY)
+/*
+ * NINDIR is the number of indirects in a file system block
+ */
+#define        NINDIR(fs)      ((fs)->fs_bsize / sizeof (daddr_t))

 
 #ifdef KERNEL
 struct fs *getfs();
 
 #ifdef KERNEL
 struct fs *getfs();