cachepc-linux

Fork of AMDESE/linux with modifications for CachePC side-channel attack
git clone https://git.sinitax.com/sinitax/cachepc-linux
Log | Files | Refs | README | LICENSE | sfeed.txt

dlmfs.rst (4553B)


      1.. SPDX-License-Identifier: GPL-2.0
      2.. include:: <isonum.txt>
      3
      4=====
      5DLMFS
      6=====
      7
      8A minimal DLM userspace interface implemented via a virtual file
      9system.
     10
     11dlmfs is built with OCFS2 as it requires most of its infrastructure.
     12
     13:Project web page:    http://ocfs2.wiki.kernel.org
     14:Tools web page:      https://github.com/markfasheh/ocfs2-tools
     15:OCFS2 mailing lists: https://oss.oracle.com/projects/ocfs2/mailman/
     16
     17All code copyright 2005 Oracle except when otherwise noted.
     18
     19Credits
     20=======
     21
     22Some code taken from ramfs which is Copyright |copy| 2000 Linus Torvalds
     23and Transmeta Corp.
     24
     25Mark Fasheh <mark.fasheh@oracle.com>
     26
     27Caveats
     28=======
     29- Right now it only works with the OCFS2 DLM, though support for other
     30  DLM implementations should not be a major issue.
     31
     32Mount options
     33=============
     34None
     35
     36Usage
     37=====
     38
     39If you're just interested in OCFS2, then please see ocfs2.txt. The
     40rest of this document will be geared towards those who want to use
     41dlmfs for easy to setup and easy to use clustered locking in
     42userspace.
     43
     44Setup
     45=====
     46
     47dlmfs requires that the OCFS2 cluster infrastructure be in
     48place. Please download ocfs2-tools from the above url and configure a
     49cluster.
     50
     51You'll want to start heartbeating on a volume which all the nodes in
     52your lockspace can access. The easiest way to do this is via
     53ocfs2_hb_ctl (distributed with ocfs2-tools). Right now it requires
     54that an OCFS2 file system be in place so that it can automatically
     55find its heartbeat area, though it will eventually support heartbeat
     56against raw disks.
     57
     58Please see the ocfs2_hb_ctl and mkfs.ocfs2 manual pages distributed
     59with ocfs2-tools.
     60
     61Once you're heartbeating, DLM lock 'domains' can be easily created /
     62destroyed and locks within them accessed.
     63
     64Locking
     65=======
     66
     67Users may access dlmfs via standard file system calls, or they can use
     68'libo2dlm' (distributed with ocfs2-tools) which abstracts the file
     69system calls and presents a more traditional locking api.
     70
     71dlmfs handles lock caching automatically for the user, so a lock
     72request for an already acquired lock will not generate another DLM
     73call. Userspace programs are assumed to handle their own local
     74locking.
     75
     76Two levels of locks are supported - Shared Read, and Exclusive.
     77Also supported is a Trylock operation.
     78
     79For information on the libo2dlm interface, please see o2dlm.h,
     80distributed with ocfs2-tools.
     81
     82Lock value blocks can be read and written to a resource via read(2)
     83and write(2) against the fd obtained via your open(2) call. The
     84maximum currently supported LVB length is 64 bytes (though that is an
     85OCFS2 DLM limitation). Through this mechanism, users of dlmfs can share
     86small amounts of data amongst their nodes.
     87
     88mkdir(2) signals dlmfs to join a domain (which will have the same name
     89as the resulting directory)
     90
     91rmdir(2) signals dlmfs to leave the domain
     92
     93Locks for a given domain are represented by regular inodes inside the
     94domain directory.  Locking against them is done via the open(2) system
     95call.
     96
     97The open(2) call will not return until your lock has been granted or
     98an error has occurred, unless it has been instructed to do a trylock
     99operation. If the lock succeeds, you'll get an fd.
    100
    101open(2) with O_CREAT to ensure the resource inode is created - dlmfs does
    102not automatically create inodes for existing lock resources.
    103
    104============  ===========================
    105Open Flag     Lock Request Type
    106============  ===========================
    107O_RDONLY      Shared Read
    108O_RDWR        Exclusive
    109============  ===========================
    110
    111
    112============  ===========================
    113Open Flag     Resulting Locking Behavior
    114============  ===========================
    115O_NONBLOCK    Trylock operation
    116============  ===========================
    117
    118You must provide exactly one of O_RDONLY or O_RDWR.
    119
    120If O_NONBLOCK is also provided and the trylock operation was valid but
    121could not lock the resource then open(2) will return ETXTBUSY.
    122
    123close(2) drops the lock associated with your fd.
    124
    125Modes passed to mkdir(2) or open(2) are adhered to locally. Chown is
    126supported locally as well. This means you can use them to restrict
    127access to the resources via dlmfs on your local node only.
    128
    129The resource LVB may be read from the fd in either Shared Read or
    130Exclusive modes via the read(2) system call. It can be written via
    131write(2) only when open in Exclusive mode.
    132
    133Once written, an LVB will be visible to other nodes who obtain Read
    134Only or higher level locks on the resource.
    135
    136See Also
    137========
    138http://opendlm.sourceforge.net/cvsmirror/opendlm/docs/dlmbook_final.pdf
    139
    140For more information on the VMS distributed locking API.