freebsd-dev/bin/pax/tables.h
Kris Kennaway 46be34b902 Various spelling/formatting changes.
Submitted by: Philippe Charnier <charnier@xp11.frmug.org>
1999-05-08 10:22:15 +00:00

174 lines
7.7 KiB
C

/*-
* Copyright (c) 1992 Keith Muller.
* Copyright (c) 1992, 1993
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Keith Muller of the University of California, San Diego.
*
* 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. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. 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.
*
* @(#)tables.h 8.1 (Berkeley) 5/31/93
* $Id: tables.h,v 1.5 1997/02/22 14:04:44 peter Exp $
*/
/*
* data structures and constants used by the different databases kept by pax
*/
/*
* Hash Table Sizes MUST BE PRIME, if set too small performance suffers.
* Probably safe to expect 500000 inodes per tape. Assuming good key
* distribution (inodes) chains of under 50 long (worse case) is ok.
*/
#define L_TAB_SZ 2503 /* hard link hash table size */
#define F_TAB_SZ 50503 /* file time hash table size */
#define N_TAB_SZ 541 /* interactive rename hash table */
#define D_TAB_SZ 317 /* unique device mapping table */
#define A_TAB_SZ 317 /* ftree dir access time reset table */
#define MAXKEYLEN 64 /* max number of chars for hash */
/*
* file hard link structure (hashed by dev/ino and chained) used to find the
* hard links in a file system or with some archive formats (cpio)
*/
typedef struct hrdlnk {
char *name; /* name of first file seen with this ino/dev */
dev_t dev; /* files device number */
ino_t ino; /* files inode number */
u_long nlink; /* expected link count */
struct hrdlnk *fow;
} HRDLNK;
/*
* Archive write update file time table (the -u, -C flag), hashed by filename.
* Filenames are stored in a scratch file at seek offset into the file. The
* file time (mod time) and the file name length (for a quick check) are
* stored in a hash table node. We were forced to use a scratch file because
* with -u, the mtime for every node in the archive must always be available
* to compare against (and this data can get REALLY large with big archives).
* By being careful to read only when we have a good chance of a match, the
* performance loss is not measurable (and the size of the archive we can
* handle is greatly increased).
*/
typedef struct ftm {
int namelen; /* file name length */
time_t mtime; /* files last modification time */
off_t seek; /* location in scratch file */
struct ftm *fow;
} FTM;
/*
* Interactive rename table (-i flag), hashed by orig filename.
* We assume this will not be a large table as this mapping data can only be
* obtained through interactive input by the user. Nobody is going to type in
* changes for 500000 files? We use chaining to resolve collisions.
*/
typedef struct namt {
char *oname; /* old name */
char *nname; /* new name typed in by the user */
struct namt *fow;
} NAMT;
/*
* Unique device mapping tables. Some protocols (e.g. cpio) require that the
* <c_dev,c_ino> pair will uniquely identify a file in an archive unless they
* are links to the same file. Appending to archives can break this. For those
* protocols that have this requirement we map c_dev to a unique value not seen
* in the archive when we append. We also try to handle inode truncation with
* this table. (When the inode field in the archive header are too small, we
* remap the dev on writes to remove accidental collisions).
*
* The list is hashed by device number using chain collision resolution. Off of
* each DEVT are linked the various remaps for this device based on those bits
* in the inode which were truncated. For example if we are just remapping to
* avoid a device number during an update append, off the DEVT we would have
* only a single DLIST that has a truncation id of 0 (no inode bits were
* stripped for this device so far). When we spot inode truncation we create
* a new mapping based on the set of bits in the inode which were stripped off.
* so if the top four bits of the inode are stripped and they have a pattern of
* 0110...... (where . are those bits not truncated) we would have a mapping
* assigned for all inodes that has the same 0110.... pattern (with this dev
* number of course). This keeps the mapping sparse and should be able to store
* close to the limit of files which can be represented by the optimal
* combination of dev and inode bits, and without creating a fouled up archive.
* Note we also remap truncated devs in the same way (an exercise for the
* dedicated reader; always wanted to say that...:)
*/
typedef struct devt {
dev_t dev; /* the orig device number we now have to map */
struct devt *fow; /* new device map list */
struct dlist *list; /* map list based on inode truncation bits */
} DEVT;
typedef struct dlist {
ino_t trunc_bits; /* truncation pattern for a specific map */
dev_t dev; /* the new device id we use */
struct dlist *fow;
} DLIST;
/*
* ftree directory access time reset table. When we are done with with a
* subtree we reset the access and mod time of the directory when the tflag is
* set. Not really explicitly specified in the pax spec, but easy and fast to
* do (and this may have even been intended in the spec, it is not clear).
* table is hashed by inode with chaining.
*/
typedef struct atdir {
char *name; /* name of directory to reset */
dev_t dev; /* dev and inode for fast lookup */
ino_t ino;
time_t mtime; /* access and mod time to reset to */
time_t atime;
struct atdir *fow;
} ATDIR;
/*
* created directory time and mode storage entry. After pax is finished during
* extraction or copy, we must reset directory access modes and times that
* may have been modified after creation (they no longer have the specified
* times and/or modes). We must reset time in the reverse order of creation,
* because entries are added from the top of the file tree to the bottom.
* We MUST reset times from leaf to root (it will not work the other
* direction). Entries are recorded into a spool file to make reverse
* reading faster.
*/
typedef struct dirdata {
int nlen; /* length of the directory name (includes \0) */
off_t npos; /* position in file where this dir name starts */
mode_t mode; /* file mode to restore */
time_t mtime; /* mtime to set */
time_t atime; /* atime to set */
int frc_mode; /* do we force mode settings? */
} DIRDATA;