brintos

brintos / linux-shallow public Read only

0
0
Text · 10.1 KiB · 2bbf68b Raw
278 lines · plain
1.. SPDX-License-Identifier: GPL-2.02 3=======================================4v9fs: Plan 9 Resource Sharing for Linux5=======================================6 7About8=====9 10v9fs is a Unix implementation of the Plan 9 9p remote filesystem protocol.11 12This software was originally developed by Ron Minnich <rminnich@sandia.gov>13and Maya Gokhale.  Additional development by Greg Watson14<gwatson@lanl.gov> and most recently Eric Van Hensbergen15<ericvh@gmail.com>, Latchesar Ionkov <lucho@ionkov.net> and Russ Cox16<rsc@swtch.com>.17 18The best detailed explanation of the Linux implementation and applications of19the 9p client is available in the form of a USENIX paper:20 21   https://www.usenix.org/events/usenix05/tech/freenix/hensbergen.html22 23Other applications are described in the following papers:24 25	* XCPU & Clustering26	  http://xcpu.org/papers/xcpu-talk.pdf27	* KVMFS: control file system for KVM28	  http://xcpu.org/papers/kvmfs.pdf29	* CellFS: A New Programming Model for the Cell BE30	  http://xcpu.org/papers/cellfs-talk.pdf31	* PROSE I/O: Using 9p to enable Application Partitions32	  http://plan9.escet.urjc.es/iwp9/cready/PROSE_iwp9_2006.pdf33	* VirtFS: A Virtualization Aware File System pass-through34	  https://kernel.org/doc/ols/2010/ols2010-pages-109-120.pdf35 36Usage37=====38 39For remote file server::40 41	mount -t 9p 10.10.1.2 /mnt/942 43For Plan 9 From User Space applications (http://swtch.com/plan9)::44 45	mount -t 9p `namespace`/acme /mnt/9 -o trans=unix,uname=$USER46 47For server running on QEMU host with virtio transport::48 49	mount -t 9p -o trans=virtio <mount_tag> /mnt/950 51where mount_tag is the tag generated by the server to each of the exported52mount points. Each 9P export is seen by the client as a virtio device with an53associated "mount_tag" property. Available mount tags can be54seen by reading /sys/bus/virtio/drivers/9pnet_virtio/virtio<n>/mount_tag files.55 56USBG Usage57==========58 59To mount a 9p FS on a USB Host accessible via the gadget at runtime::60 61	mount -t 9p -o trans=usbg,aname=/path/to/fs <device> /mnt/962 63To mount a 9p FS on a USB Host accessible via the gadget as root filesystem::64 65	root=<device> rootfstype=9p rootflags=trans=usbg,cache=loose,uname=root,access=0,dfltuid=0,dfltgid=0,aname=/path/to/rootfs66 67where <device> is the tag associated by the usb gadget transport.68It is defined by the configfs instance name.69 70USBG Example71============72 73The USB host exports a filesystem, while the gadget on the USB device74side makes it mountable.75 76Diod (9pfs server) and the forwarder are on the development host, where77the root filesystem is actually stored. The gadget is initialized during78boot (or later) on the embedded board. Then the forwarder will find it79on the USB bus and start forwarding requests.80 81In this case the 9p requests come from the device and are handled by the82host. The reason is that USB device ports are normally not available on83PCs, so a connection in the other direction would not work.84 85When using the usbg transport, for now there is no native usb host86service capable to handle the requests from the gadget driver. For87this we have to use the extra python tool p9_fwd.py from tools/usb.88 89Just start the 9pfs capable network server like diod/nfs-ganesha e.g.::90 91        $ diod -f -n -d 0 -S -l 0.0.0.0:9999 -e $PWD92 93Optionaly scan your bus if there are more then one usbg gadgets to find their path::94 95        $ python $kernel_dir/tools/usb/p9_fwd.py list96 97        Bus | Addr | Manufacturer     | Product          | ID        | Path98        --- | ---- | ---------------- | ---------------- | --------- | ----99          2 |   67 | unknown          | unknown          | 1d6b:0109 | 2-1.1.2100          2 |   68 | unknown          | unknown          | 1d6b:0109 | 2-1.1.3101 102Then start the python transport::103 104        $ python $kernel_dir/tools/usb/p9_fwd.py --path 2-1.1.2 connect -p 9999105 106After that the gadget driver can be used as described above.107 108One use-case is to use it as an alternative to NFS root booting during109the development of embedded Linux devices.110 111Options112=======113 114  ============= ===============================================================115  trans=name	select an alternative transport.  Valid options are116  		currently:117 118			========  ============================================119			unix 	  specifying a named pipe mount point120			tcp	  specifying a normal TCP/IP connection121			fd   	  used passed file descriptors for connection122                                  (see rfdno and wfdno)123			virtio	  connect to the next virtio channel available124				  (from QEMU with trans_virtio module)125			rdma	  connect to a specified RDMA channel126			usbg	  connect to a specified usb gadget channel127			========  ============================================128 129  uname=name	user name to attempt mount as on the remote server.  The130  		server may override or ignore this value.  Certain user131		names may require authentication.132 133  aname=name	aname specifies the file tree to access when the server is134  		offering several exported file systems.135 136  cache=mode	specifies a caching policy.  By default, no caches are used.137		The mode can be specified as a bitmask or by using one of the138		preexisting common 'shortcuts'.139		The bitmask is described below: (unspecified bits are reserved)140 141			==========  ====================================================142			0b00000000  all caches disabled, mmap disabled143			0b00000001  file caches enabled144			0b00000010  meta-data caches enabled145			0b00000100  writeback behavior (as opposed to writethrough)146			0b00001000  loose caches (no explicit consistency with server)147			0b10000000  fscache enabled for persistent caching148			==========  ====================================================149 150		The current shortcuts and their associated bitmask are:151 152			=========   ====================================================153			none        0b00000000 (no caching)154			readahead   0b00000001 (only read-ahead file caching)155			mmap        0b00000101 (read-ahead + writeback file cache)156			loose       0b00001111 (non-coherent file and meta-data caches)157			fscache     0b10001111 (persistent loose cache)158			=========   ====================================================159 160		NOTE: only these shortcuts are tested modes of operation at the161		moment, so using other combinations of bit-patterns is not162		known to work.  Work on better cache support is in progress.163 164		IMPORTANT: loose caches (and by extension at the moment fscache)165		do not necessarily validate cached values on the server.  In other166		words changes on the server are not guaranteed to be reflected167		on the client system.  Only use this mode of operation if you168		have an exclusive mount and the server will modify the filesystem169		underneath you.170 171  debug=n	specifies debug level.  The debug level is a bitmask.172 173			=====   ================================174			0x01    display verbose error messages175			0x02    developer debug (DEBUG_CURRENT)176			0x04    display 9p trace177			0x08    display VFS trace178			0x10    display Marshalling debug179			0x20    display RPC debug180			0x40    display transport debug181			0x80    display allocation debug182			0x100   display protocol message debug183			0x200   display Fid debug184			0x400   display packet debug185			0x800   display fscache tracing debug186			=====   ================================187 188  rfdno=n	the file descriptor for reading with trans=fd189 190  wfdno=n	the file descriptor for writing with trans=fd191 192  msize=n	the number of bytes to use for 9p packet payload193 194  port=n	port to connect to on the remote server195 196  noextend	force legacy mode (no 9p2000.u or 9p2000.L semantics)197 198  version=name	Select 9P protocol version. Valid options are:199 200			========        ==============================201			9p2000          Legacy mode (same as noextend)202			9p2000.u        Use 9P2000.u protocol203			9p2000.L        Use 9P2000.L protocol204			========        ==============================205 206  dfltuid	attempt to mount as a particular uid207 208  dfltgid	attempt to mount with a particular gid209 210  afid		security channel - used by Plan 9 authentication protocols211 212  nodevmap	do not map special files - represent them as normal files.213  		This can be used to share devices/named pipes/sockets between214		hosts.  This functionality will be expanded in later versions.215 216  directio	bypass page cache on all read/write operations217 218  ignoreqv	ignore qid.version==0 as a marker to ignore cache219 220  noxattr	do not offer xattr functions on this mount.221 222  access	there are four access modes.223			user224				if a user tries to access a file on v9fs225			        filesystem for the first time, v9fs sends an226			        attach command (Tattach) for that user.227				This is the default mode.228			<uid>229				allows only user with uid=<uid> to access230				the files on the mounted filesystem231			any232				v9fs does single attach and performs all233				operations as one user234			clien235				 ACL based access check on the 9p client236			         side for access validation237 238  cachetag	cache tag to use the specified persistent cache.239		cache tags for existing cache sessions can be listed at240		/sys/fs/9p/caches. (applies only to cache=fscache)241  ============= ===============================================================242 243Behavior244========245 246This section aims at describing 9p 'quirks' that can be different247from a local filesystem behaviors.248 249 - Setting O_NONBLOCK on a file will make client reads return as early250   as the server returns some data instead of trying to fill the read251   buffer with the requested amount of bytes or end of file is reached.252 253Resources254=========255 256Protocol specifications are maintained on github:257http://ericvh.github.com/9p-rfc/258 2599p client and server implementations are listed on260http://9p.cat-v.org/implementations261 262A 9p2000.L server is being developed by LLNL and can be found263at http://code.google.com/p/diod/264 265There are user and developer mailing lists available through the v9fs project266on sourceforge (http://sourceforge.net/projects/v9fs).267 268News and other information is maintained on a Wiki.269(http://sf.net/apps/mediawiki/v9fs/index.php).270 271Bug reports are best issued via the mailing list.272 273For more information on the Plan 9 Operating System check out274http://plan9.bell-labs.com/plan9275 276For information on Plan 9 from User Space (Plan 9 applications and libraries277ported to Linux/BSD/OSX/etc) check out https://9fans.github.io/plan9port/278