Return to README CVS log

Up to [Research Unix] / researchv10no / netfs

Annotation of researchv10no/netfs/README, revision 1.1.1.1

1.1       root        1: This is the source to various network file system pieces.
                      2: There are protocol-specific parts (interface libraries and support
                      3: programs), and there are servers.
                      4: 
                      5: serv/* are server sources.  In principal, they may be compiled
                      6: with any protocol library; in practice, serv/makefile uses -lnetb,
                      7: so by default the servers speak that protocol.
                      8: Some of the servers are
                      9: zarf   present the local file system
                     10: ffs    interpret a 4BSD file system
                     11: f6     interpret a Fifth or Sixth Edition Unix file system
                     12: f7     interpret a Seventh Edition Unix file system
                     13: d7     interpret a Seventh Edition dump tape (present as a file system)
                     14: f11    interpret a DEC FILES-11 ODS 1 version 2 file system
                     15: All but zarf are read-only file systems for now.
                     16: Only zarf is known to be portable; some of the others were written as
                     17: quick hacks, and need some tidying up.
                     18: 
                     19: libnetb/* makes /usr/lib/libnetb.a, the protocol library for
                     20: netb (the Research Unix file system protocol).  There are two
                     21: ancillary programs: runfs starts a server by hand, mounted
                     22: on some local file and connected through a pipe; setup is the
                     23: daemon that maintains a handful of file systems, restarting them
                     24: when they break.  Both runfs and setup are specific to Research
                     25: Unix; the library is meant to run anywhere.
                     26: 
                     27: Some of the files in libnetb aren't really netb-specific.  Some
                     28: day, when there are more protocols supported, there will be a
                     29: separate directory with the protocol-independent parts, and life
                     30: will be considerably more confusing.
                     31: 
                     32: Here are some hints on making things work on a new system:
                     33: 
                     34: Bring along libnetb/*.[ch] libnetb/makefile serv/*.[ch] serv/makefile
                     35: (You probably don't want all the servers; to take just the ordinary
                     36: file system server, bring zarf.c except.c *dir.c zarf.h)
                     37: Deposit them somewhere in directories libnetb and serv.
                     38: serv/makefile and the server programs expect to find <rf.h> in /usr/include
                     39: and libnetb in /usr/lib.  If you aren't planning to install them there,
                     40: add -I../libnetb to CFLAGS in serv/makefile, and change L from -lnetb
                     41: to ../libnetb/libnetb.a (or wherever you put those files).
                     42: You probably also want to edit serv/makefile to make only zarf,
                     43: and libnetb/makefile to make only libnetb.a; deleting the `all'
                     44: lines should suffice.
                     45: 
                     46: Whenever possible, adjust things by changing the makefile to use different
                     47: source files, rather than by editing the sources themselves.  There are
                     48: almost no #ifdefs in the programs, and even those shouldn't be.  Generally
                     49: there are different source files containing routines that vary among systems:
                     50: a couple of directory-reading routines, for example.
                     51: 
                     52: You should have to adjust two particular things (and no others, I hope):
                     53: The routine that reads directories, and the routine that determines who called.
                     54: The former is mandatory; the latter can be put off unless you grant different
                     55: rights to different clients.
                     56: 
                     57: Directories:
                     58: As shipped, zarf reads directories using a research-specific system
                     59: call; this probably isn't what you want.  To use the standard `portable'
                     60: directory-reading routines, change resdir.o to libdir.o in serv/makefile.
                     61: If you're on a 4BSD system, you probably want bsdlibdir.o instead, which
                     62: differs only in a couple of names.  If the directory routines are in a
                     63: separate library, add the library to the line that makes zarf.
                     64: 
                     65: If you think (or measure) that the library routines are too slow, write
                     66: your own, but it probably isn't worth the bother.
                     67: 
                     68: Who called:
                     69: As shipped, libnetb uses a routine that cracks one of a couple of
                     70: environment variables set up by present and historic research connection
                     71: servers and Datakit managers.  zarf cares who called so it can print it
                     72: in the logfile, and to compare with entries in the except file.  If you
                     73: don't care, leave it be; at worst it will leave the caller's name empty.
                     74: If you do care, edit libnetb/makefile as follows:
                     75: 
                     76: If you're on a 4BSD-socket-like system, use bsdwhocalled.o and bsdgetpeer.o
                     77: instead of reswhocalled.o and callsource.o.
                     78: 
                     79: If you're on a system with 4BSD sockets for TCP/IP and the research connection
                     80: server for Datakit, use mixwhocalled.o instead of reswhocalled.o, keep
                     81: callsource.o, and add bsdwhocalled.o.
                     82: 
                     83: Certain System-V-with-Berkeley-extensions systems have sockets but put
                     84: Berkeley headers and libraries in funny places.  If bsdgetpeer.c won't
                     85: compile, add -I/usr/include/bsd (or whatever) to CFLAGS in the makefile;
                     86: if zarf won't load because getpeername is missing, add -lbsd (or whatever)
                     87: to the makefile line that makes zarf.
                     88: 
                     89: Miscellany:
                     90: If the target system doesn't have makedir and rmdir system calls, add
                     91: makedir.o to the line that makes zarf in serv/makefile.
                     92: If the C library on the target doesn't have strchr and strrchr, it
                     93: probably has index and rindex instead; add `-Dstrchr=index -Dstrrchr=rindex'
                     94: to CFLAGS in both makefiles.
                     95: 
                     96: There is one #if hack: on 64-bit machines, the size of a file is often
                     97: 64 bits long, but the netb protocol only allows 32-bit sizes.  There is
                     98: a little compensating code in libnetb/funcs.c.
                     99: 
                    100: 
                    101: Once you have all the makefiles right,
                    102: cd libnetb; make
                    103: cd ../serv; make zarf
                    104: 
                    105: Unlike earlier programs of the same name, zarf has no built-in filenames.
                    106: It prints logging info on the standard error file; it expects arguments
                    107: of the form `-e except' to point it at the except file.  Arrange for
                    108: network calls to invoke zarf with output redirected and with the appropriate
                    109: arguments.  If necessary, write a shell script, or perhaps a tiny C program,
                    110: that does the setup.  On research machines there is a shell script; see
                    111: serv/zarf.go.  Edit it to your taste.
                    112: 
                    113: For zarf to offer full permissions to the client, it must run as the super-
                    114: user.  It is probably a bad idea to make it set-userid to root; use network
                    115: services instead.  On TCP/IP server machines, the preferred way is to put
                    116: zarf.go in a known place, and arrange for TCP port 400 to invoke zarf.go
                    117: as the super-user, e.g. with the following line in /etc/inetd.conf:
                    118: zarf   stream  tcp     nowait  root    /usr/netb/zarf.go       /usr/netb/zarf.go

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