Return to sys.ufs.t CVS log

Up to [CSRG BSD Unix] / 43BSDReno / share / doc / smm / 13.kchanges

Annotation of 43BSDReno/share/doc/smm/13.kchanges/sys.ufs.t, revision 1.1

1.1     ! root        1: .\" Copyright (c) 1986 Regents of the University of California.
        !             2: .\" All rights reserved.  The Berkeley software License Agreement
        !             3: .\" specifies the terms and conditions for redistribution.
        !             4: .\"
        !             5: .\"    @(#)sys.ufs.t   1.5 (Berkeley) 4/11/86
        !             6: .\"
        !             7: .NH
        !             8: Changes in the filesystem
        !             9: .PP
        !            10: The major change in the filesystem was the addition of a name translation
        !            11: cache.
        !            12: A table of recent name-to-inode translations is maintained by \fInamei\fP,
        !            13: and used as a lookaside cache when translating each component of each
        !            14: file pathname.
        !            15: Each \fInamecache\fP entry contains the parent directory's device and inode,
        !            16: the length of the name, and the name itself, and is hashed on the name.
        !            17: It also contains a pointer to the inode for the file whose name it contains.
        !            18: Unlike most inode pointers, which hold a ``hard'' reference
        !            19: by incrementing the reference count,
        !            20: the name cache holds a ``soft'' reference, a pointer to an inode
        !            21: that may be reused.
        !            22: In order to validate the inode from a name cache reference,
        !            23: each inode is assigned a unique ``capability'' when it is brought
        !            24: into memory.
        !            25: When the inode entry is reused for another file,
        !            26: or when the name of the file is changed,
        !            27: this capability is changed.
        !            28: This allows the inode cache to be handled normally,
        !            29: releasing inodes at the head of the LRU list without regard for name
        !            30: cache references,
        !            31: and allows multiple names for the same inode to be in the cache simultaneously
        !            32: without complicating the invalidation procedure.
        !            33: An additional feature of this scheme is that when opening
        !            34: a file, it is possible to determine whether the file was previously open.
        !            35: This is useful when beginning execution of a file, to check whether
        !            36: the file might be open for writing, and for similar situations.
        !            37: .PP
        !            38: Other changes that are visible throughout the filesystem
        !            39: include greater use of the ILOCK and IUNLOCK macros rather than the
        !            40: subroutine equivalents.
        !            41: The inode times are updated on each \fIirele\fP, not only when
        !            42: the reference count reaches zero,
        !            43: if the IACC, IUPD or ICHG flags are set.
        !            44: This is accomplished with the ITIMES macro;
        !            45: the inode is marked as modified with the new IMOD flag,
        !            46: that causes it to be written to disk when released, or on the next sync.
        !            47: .PP
        !            48: The remainder of this section describes the filesystem changes that are
        !            49: localized to individual files.
        !            50: .XP ufs_alloc.c
        !            51: The algorithm for extending file fragments was changed
        !            52: to take advantage of the observation that fragments that were once extended
        !            53: were frequently extended again, that is, that the file was being written
        !            54: in fragments.
        !            55: Therefore, the first time a given fragment is allocated,
        !            56: a best-fit strategy is used.
        !            57: Thereafter, when this fragment is to be extended,
        !            58: a full-sized block is allocated, the fragment removed from it,
        !            59: and the remainder freed for use in subsequent expansion.
        !            60: As this policy may result in increased fragmentation,
        !            61: it is not used when the filesystem becomes excessively
        !            62: fragmented (i.e. when the number of free fragments falls to 2%
        !            63: of the minfree value);
        !            64: the policy is stored in the superblock and may be changed with \fItunefs\fP.
        !            65: The \fIfserr\fP routine was converted to use \fIlog\fP rather than \fIprintf\fP.
        !            66: .XP ufs_bio.c
        !            67: I/O operations traced now include the size where relevant.
        !            68: .XP ufs_inode.c
        !            69: The size of the buffer hash table was increased substantially
        !            70: and changed to a power of two to allow the modulus to be computed with a mask
        !            71: operation.
        !            72: \fIIget\fP invalidates the capability in each inode that is flushed
        !            73: from the inode cache for reuse.
        !            74: The new \fIigrab\fP routine is used instead of \fIiget\fP
        !            75: when fetching an inode from a name cache reference;
        !            76: it waits for the inode to be unlocked if necessary,
        !            77: and removes it from the free list if it was free.
        !            78: The caller must check that the inode is still valid after the \fIigrab\fP.
        !            79: A bug was fixed in \fIitrunc\fP that allowed old contents to creep back into
        !            80: a file.
        !            81: When truncating to a location within a block,
        !            82: \fIitrunc\fP must clear the remainder of the block.
        !            83: Otherwise, if the file is extended by seeking past the end of file
        !            84: and then writing, the old contents reappear.
        !            85: .\" \fIItrunc\fP also waits for 
        !            86: .XP ufs_mount.c
        !            87: The \fImount\fP system call was modified to return different error numbers
        !            88: for different types of errors.
        !            89: \fIMount\fP now examines the superblock more carefully
        !            90: before using size field it contains as the amount to copy into a new buffer.
        !            91: If a mount fails for a reason other than the device already being
        !            92: mounted, the device is closed again.
        !            93: When performing the name lookup for the mount point,
        !            94: \fImount\fP must prevent the name translation from being left
        !            95: in the name cache;
        !            96: \fIumount\fP must flush all name translations for the device.
        !            97: A bug in \fIgetmdev\fP caused an inode to remain locked
        !            98: if the specified device was not a block special file; this has been fixed.
        !            99: .XP ufs_namei.c
        !           100: This file was previously called ufs_nami.c.
        !           101: The \fInamei\fP function has a new calling convention
        !           102: with its arguments, associated context, and side effects
        !           103: encapsulated in a single structure.
        !           104: It has been extensively modified to implement the name cache
        !           105: and to cache directory offsets for each process.
        !           106: It may now return ENAMETOOLONG when appropriate,
        !           107: and returns EINVAL if the 8th bit is set on one of the pathname
        !           108: characters.
        !           109: Directories may be foreshortened if the last one or more blocks
        !           110: contain no entries;
        !           111: this is done when files are being created, as the entire directory 
        !           112: must already be searched.
        !           113: An entry is provided for invalidating the entire name cache
        !           114: when the 32-bit prototype for capabilities wraps around.
        !           115: This is expected to happen after 13 months of operation,
        !           116: assuming 100 name lookups per second, all of which miss the cache.
        !           117: .XP
        !           118: A change in filesystem semantics is the introduction
        !           119: of ``sticky'' directories.
        !           120: If the ISVTX (sticky text) bit is set in the mode of a directory,
        !           121: files may only be removed from that directory by the owner of the file,
        !           122: the owner of the directory, or the superuser.
        !           123: This is enforced by \fInamei\fP when the lookup operation is DELETE. 
        !           124: .XP ufs_subr.c
        !           125: The strategy for \fIsyncip\fP, the internal routine implementing \fIfsync\fP,
        !           126: has been modified for large files (those larger than half of the buffer
        !           127: cache).
        !           128: For large files all modified buffers for the device are written out.
        !           129: The old algorithm could run for a very long time on a very large file,
        !           130: that might not actually have many data blocks.
        !           131: The \fIupdate\fP routine now saves some work by calling \fIiupdate\fP
        !           132: only for modified inodes.
        !           133: The C replacements for the special VAX instructions have been collected
        !           134: in this file.
        !           135: .XP ufs_syscalls.c
        !           136: When doing an open with flags O_CREAT and O_EXCL (create only if the file
        !           137: did not exist), it is now considered to be an error if the target exists
        !           138: and is a symbolic link, even if the symbolic link refers to a nonexistent
        !           139: file.
        !           140: This behavior is desirable for reasons of security
        !           141: in programs that create files with predictable names.
        !           142: \fIRename\fP follows the policy of \fInamei\fP in disallowing removal
        !           143: of the target of a rename if the target directory is ``sticky''
        !           144: and the user is not the owner of the target or the target directory.
        !           145: A serious bug in the open code which allowed directories and other unwritable
        !           146: files to be truncated has been corrected.
        !           147: Interrupted opens no longer lose file descriptors.
        !           148: The \fIlseek\fP call returns an ESPIPE error when seeking on sockets
        !           149: (including pipes) for backward compatibility.
        !           150: The error returned from \fIreadlink\fP when reading something other than
        !           151: a symbolic link was changed from ENXIO to EINVAL.
        !           152: Several calls that previously failed silently on read-only filesystems
        !           153: (\fIchmod\fP, \fIchown\fP, \fIfchmod\fP, \fIfchown\fP and \fIutimes\fP)
        !           154: now return EROFS.
        !           155: The \fIrename\fP code was reworked to avoid several races
        !           156: and to invalidate the name cache.
        !           157: It marks a directory being renamed with IRENAME
        !           158: to avoid races due to concurrent renames of the same directory.
        !           159: \fIMkdir\fP now sets the size of all new directories to DIRBLKSIZE.
        !           160: \fIRmdir\fP purges the name cache of entries for the removed directory.
        !           161: .XP ufs_xxx.c
        !           162: The routines \fIuchar\fP and \fIschar\fP are no longer used
        !           163: and have been removed.
        !           164: .XP quota_kern.c
        !           165: The quota hash size was changed to a power of 2 so that the modulus could
        !           166: be computed with a mask.
        !           167: .XP quota_ufs.c
        !           168: If a user has run out of warnings and had the hard limit enforced
        !           169: while logged in,
        !           170: but has then brought his allocation below the hard limit,
        !           171: the quota system reverts to enforcing the soft limit,
        !           172: and resets the warning count;
        !           173: users previously were required to log out and in again to
        !           174: get this affect.

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.