The files:
nfsclient.c
makefile
RUNME
api/makefile
api/nfsapi.c
api/nfsapi.h
api/tester.c
myrpc/awake.c
myrpc/awake.h
myrpc/makefile
myrpc/mount.h
myrpc/mount_clnt.c
myrpc/mount_svc.c
myrpc/mount_xdr.c
myrpc/nfs_prot.h
myrpc/nfs_prot_clnt.c
myrpc/nfs_prot_svc.c
myrpc/nfs_prot_xdr.c

An NFS client library

This is a simple NFS client library that allows an application to read files stored on an NFS server. The library exports the following functions:

int NFSmount(const char *host, const char *path)

Logically mounts the tree of an exported file system. It must be called before any other function, and all successive function calls are relative to this location.

FILE * NFSopen(char *filename)

Opens a file at the given pathname in the mounted location. The file handle that it returns is not compatible with a regular file handle, and client code must take care to use it only with this API's functions.

int NFSread(void *ptr, int size, FILE * stream)

Reads the next size bytes from the given file, starting at the file's logical position, into the character array ptr. The return value is the number of bytes read; -1 on error.

int NFSclose(FILE * stream)

Closes the given file.


This has two pieces, the NFSClient piece and the API. The NFSClient must be running for programs using the API to operate. This is a Solaris program, and I don't believe the RPC calls work under other UNIXes.

To make the NFSClient, either cd myrpc; make; cd ..; make or just execute RUNME, which will do this automatically.

NFSClient is run with no arguments, and listens on UNIX socket /tmp/smNFS. Sorry, yes, this is hardcoded.

The API is found in the api directory. You must make it separately. With my compiler, it works only when called from C, not C++. Your mileage may vary.

To run your own code, compile with api/nfsapi.o and include api/nfsapi.h.

To try my test program, compile api/tester.c. Usage is Tester hostname directory filename readsize.


The NFS client handles only one connection at a time, but no connection takes very long.

Read requests must be for under 8000 bytes. This has to do with packet size. You don't want to know.

Once 256 files are open simultaneously -- by all applications, since the client does not discriminate between requests in any way -- file handles begin to be overwritten. The client prints an error.

If the client has problems opening sockets it quits gracefully, including returning a message over the socket to the application. The exception is if it is given a bad hostname to mount, in which case it just responds with failure rather than quitting.

If the formatting of the code looks messed up, it's because the code was written half on a Mac (tab = 4 spaces).