freebsd-skq/sys/rpc/xdr.h

371 lines
12 KiB
C
Raw Normal View History

Add the new kernel-mode NFS Lock Manager. To use it instead of the user-mode lock manager, build a kernel with the NFSLOCKD option and add '-k' to 'rpc_lockd_flags' in rc.conf. Highlights include: * Thread-safe kernel RPC client - many threads can use the same RPC client handle safely with replies being de-multiplexed at the socket upcall (typically driven directly by the NIC interrupt) and handed off to whichever thread matches the reply. For UDP sockets, many RPC clients can share the same socket. This allows the use of a single privileged UDP port number to talk to an arbitrary number of remote hosts. * Single-threaded kernel RPC server. Adding support for multi-threaded server would be relatively straightforward and would follow approximately the Solaris KPI. A single thread should be sufficient for the NLM since it should rarely block in normal operation. * Kernel mode NLM server supporting cancel requests and granted callbacks. I've tested the NLM server reasonably extensively - it passes both my own tests and the NFS Connectathon locking tests running on Solaris, Mac OS X and Ubuntu Linux. * Userland NLM client supported. While the NLM server doesn't have support for the local NFS client's locking needs, it does have to field async replies and granted callbacks from remote NLMs that the local client has contacted. We relay these replies to the userland rpc.lockd over a local domain RPC socket. * Robust deadlock detection for the local lock manager. In particular it will detect deadlocks caused by a lock request that covers more than one blocking request. As required by the NLM protocol, all deadlock detection happens synchronously - a user is guaranteed that if a lock request isn't rejected immediately, the lock will eventually be granted. The old system allowed for a 'deferred deadlock' condition where a blocked lock request could wake up and find that some other deadlock-causing lock owner had beaten them to the lock. * Since both local and remote locks are managed by the same kernel locking code, local and remote processes can safely use file locks for mutual exclusion. Local processes have no fairness advantage compared to remote processes when contending to lock a region that has just been unlocked - the local lock manager enforces a strict first-come first-served model for both local and remote lockers. Sponsored by: Isilon Systems PR: 95247 107555 115524 116679 MFC after: 2 weeks
2008-03-26 15:23:12 +00:00
/* $NetBSD: xdr.h,v 1.19 2000/07/17 05:00:45 matt Exp $ */
/*
* Sun RPC is a product of Sun Microsystems, Inc. and is provided for
* unrestricted use provided that this legend is included on all tape
* media and as a part of the software program in whole or part. Users
* may copy or modify Sun RPC without charge, but are not authorized
* to license or distribute it to anyone else except as part of a product or
* program developed by the user.
*
* SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE
* WARRANTIES OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE.
*
* Sun RPC is provided with no support and without any obligation on the
* part of Sun Microsystems, Inc. to assist in its use, correction,
* modification or enhancement.
*
* SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE
* INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC
* OR ANY PART THEREOF.
*
* In no event will Sun Microsystems, Inc. be liable for any lost revenue
* or profits or other special, indirect and consequential damages, even if
* Sun has been advised of the possibility of such damages.
*
* Sun Microsystems, Inc.
* 2550 Garcia Avenue
* Mountain View, California 94043
*
* from: @(#)xdr.h 1.19 87/04/22 SMI
* from: @(#)xdr.h 2.2 88/07/29 4.0 RPCSRC
* $FreeBSD$
*/
/*
* xdr.h, External Data Representation Serialization Routines.
*
* Copyright (C) 1984, Sun Microsystems, Inc.
*/
#ifndef _KRPC_XDR_H
#define _KRPC_XDR_H
#include <sys/cdefs.h>
/*
* XDR provides a conventional way for converting between C data
* types and an external bit-string representation. Library supplied
* routines provide for the conversion on built-in C data types. These
* routines and utility routines defined here are used to help implement
* a type encode/decode routine for each user-defined type.
*
* Each data type provides a single procedure which takes two arguments:
*
* bool_t
* xdrproc(xdrs, argresp)
* XDR *xdrs;
* <type> *argresp;
*
* xdrs is an instance of a XDR handle, to which or from which the data
* type is to be converted. argresp is a pointer to the structure to be
* converted. The XDR handle contains an operation field which indicates
* which of the operations (ENCODE, DECODE * or FREE) is to be performed.
*
* XDR_DECODE may allocate space if the pointer argresp is null. This
* data can be freed with the XDR_FREE operation.
*
* We write only one procedure per data type to make it easy
* to keep the encode and decode procedures for a data type consistent.
* In many cases the same code performs all operations on a user defined type,
* because all the hard work is done in the component type routines.
* decode as a series of calls on the nested data types.
*/
/*
* Xdr operations. XDR_ENCODE causes the type to be encoded into the
* stream. XDR_DECODE causes the type to be extracted from the stream.
* XDR_FREE can be used to release the space allocated by an XDR_DECODE
* request.
*/
enum xdr_op {
XDR_ENCODE=0,
XDR_DECODE=1,
XDR_FREE=2
};
/*
* This is the number of bytes per unit of external data.
*/
#define BYTES_PER_XDR_UNIT (4)
#define RNDUP(x) ((((x) + BYTES_PER_XDR_UNIT - 1) / BYTES_PER_XDR_UNIT) \
* BYTES_PER_XDR_UNIT)
/*
* The XDR handle.
* Contains operation which is being applied to the stream,
* an operations vector for the particular implementation (e.g. see xdr_mem.c),
* and two private fields for the use of the particular implementation.
*/
typedef struct __rpc_xdr {
enum xdr_op x_op; /* operation; fast additional param */
const struct xdr_ops {
/* get a long from underlying stream */
bool_t (*x_getlong)(struct __rpc_xdr *, long *);
/* put a long to " */
bool_t (*x_putlong)(struct __rpc_xdr *, const long *);
/* get some bytes from " */
bool_t (*x_getbytes)(struct __rpc_xdr *, char *, u_int);
/* put some bytes to " */
bool_t (*x_putbytes)(struct __rpc_xdr *, const char *, u_int);
/* returns bytes off from beginning */
u_int (*x_getpostn)(struct __rpc_xdr *);
/* lets you reposition the stream */
bool_t (*x_setpostn)(struct __rpc_xdr *, u_int);
/* buf quick ptr to buffered data */
int32_t *(*x_inline)(struct __rpc_xdr *, u_int);
/* free privates of this xdr_stream */
void (*x_destroy)(struct __rpc_xdr *);
bool_t (*x_control)(struct __rpc_xdr *, int, void *);
} *x_ops;
char * x_public; /* users' data */
void * x_private; /* pointer to private data */
char * x_base; /* private used for position info */
u_int x_handy; /* extra private word */
} XDR;
/*
* A xdrproc_t exists for each data type which is to be encoded or decoded.
*
* The second argument to the xdrproc_t is a pointer to an opaque pointer.
* The opaque pointer generally points to a structure of the data type
* to be decoded. If this pointer is 0, then the type routines should
* allocate dynamic storage of the appropriate size and return it.
*/
#ifdef _KERNEL
typedef bool_t (*xdrproc_t)(XDR *, void *, ...);
#else
/*
* XXX can't actually prototype it, because some take three args!!!
*/
typedef bool_t (*xdrproc_t)(XDR *, ...);
#endif
/*
* Operations defined on a XDR handle
*
* XDR *xdrs;
* long *longp;
* char * addr;
* u_int len;
* u_int pos;
*/
#define XDR_GETLONG(xdrs, longp) \
(*(xdrs)->x_ops->x_getlong)(xdrs, longp)
#define xdr_getlong(xdrs, longp) \
(*(xdrs)->x_ops->x_getlong)(xdrs, longp)
#define XDR_PUTLONG(xdrs, longp) \
(*(xdrs)->x_ops->x_putlong)(xdrs, longp)
#define xdr_putlong(xdrs, longp) \
(*(xdrs)->x_ops->x_putlong)(xdrs, longp)
static __inline int
xdr_getint32(XDR *xdrs, int32_t *ip)
{
long l;
if (!xdr_getlong(xdrs, &l))
return (FALSE);
*ip = (int32_t)l;
return (TRUE);
}
static __inline int
xdr_putint32(XDR *xdrs, int32_t *ip)
{
long l;
l = (long)*ip;
return xdr_putlong(xdrs, &l);
}
#define XDR_GETINT32(xdrs, int32p) xdr_getint32(xdrs, int32p)
#define XDR_PUTINT32(xdrs, int32p) xdr_putint32(xdrs, int32p)
#define XDR_GETBYTES(xdrs, addr, len) \
(*(xdrs)->x_ops->x_getbytes)(xdrs, addr, len)
#define xdr_getbytes(xdrs, addr, len) \
(*(xdrs)->x_ops->x_getbytes)(xdrs, addr, len)
#define XDR_PUTBYTES(xdrs, addr, len) \
(*(xdrs)->x_ops->x_putbytes)(xdrs, addr, len)
#define xdr_putbytes(xdrs, addr, len) \
(*(xdrs)->x_ops->x_putbytes)(xdrs, addr, len)
#define XDR_GETPOS(xdrs) \
(*(xdrs)->x_ops->x_getpostn)(xdrs)
#define xdr_getpos(xdrs) \
(*(xdrs)->x_ops->x_getpostn)(xdrs)
#define XDR_SETPOS(xdrs, pos) \
(*(xdrs)->x_ops->x_setpostn)(xdrs, pos)
#define xdr_setpos(xdrs, pos) \
(*(xdrs)->x_ops->x_setpostn)(xdrs, pos)
#define XDR_INLINE(xdrs, len) \
(*(xdrs)->x_ops->x_inline)(xdrs, len)
#define xdr_inline(xdrs, len) \
(*(xdrs)->x_ops->x_inline)(xdrs, len)
#define XDR_DESTROY(xdrs) \
if ((xdrs)->x_ops->x_destroy) \
(*(xdrs)->x_ops->x_destroy)(xdrs)
#define xdr_destroy(xdrs) \
if ((xdrs)->x_ops->x_destroy) \
(*(xdrs)->x_ops->x_destroy)(xdrs)
#define XDR_CONTROL(xdrs, req, op) \
if ((xdrs)->x_ops->x_control) \
(*(xdrs)->x_ops->x_control)(xdrs, req, op)
#define xdr_control(xdrs, req, op) XDR_CONTROL(xdrs, req, op)
/*
* Solaris strips the '_t' from these types -- not sure why.
* But, let's be compatible.
*/
#define xdr_rpcvers(xdrs, versp) xdr_uint32_t(xdrs, versp)
#define xdr_rpcprog(xdrs, progp) xdr_uint32_t(xdrs, progp)
#define xdr_rpcproc(xdrs, procp) xdr_uint32_t(xdrs, procp)
#define xdr_rpcprot(xdrs, protp) xdr_uint32_t(xdrs, protp)
#define xdr_rpcport(xdrs, portp) xdr_uint32_t(xdrs, portp)
/*
* Support struct for discriminated unions.
* You create an array of xdrdiscrim structures, terminated with
* an entry with a null procedure pointer. The xdr_union routine gets
* the discriminant value and then searches the array of structures
* for a matching value. If a match is found the associated xdr routine
* is called to handle that part of the union. If there is
* no match, then a default routine may be called.
* If there is no match and no default routine it is an error.
*/
#define NULL_xdrproc_t ((xdrproc_t)0)
struct xdr_discrim {
int value;
xdrproc_t proc;
};
/*
* In-line routines for fast encode/decode of primitive data types.
* Caveat emptor: these use single memory cycles to get the
* data from the underlying buffer, and will fail to operate
* properly if the data is not aligned. The standard way to use these
* is to say:
* if ((buf = XDR_INLINE(xdrs, count)) == NULL)
* return (FALSE);
* <<< macro calls >>>
* where ``count'' is the number of bytes of data occupied
* by the primitive data types.
*
* N.B. and frozen for all time: each data type here uses 4 bytes
* of external representation.
*/
#define IXDR_GET_INT32(buf) ((int32_t)__ntohl((uint32_t)*(buf)++))
#define IXDR_PUT_INT32(buf, v) (*(buf)++ =(int32_t)__htonl((uint32_t)v))
#define IXDR_GET_U_INT32(buf) ((uint32_t)IXDR_GET_INT32(buf))
#define IXDR_PUT_U_INT32(buf, v) IXDR_PUT_INT32((buf), ((int32_t)(v)))
#define IXDR_GET_UINT32(buf) ((uint32_t)IXDR_GET_INT32(buf))
#define IXDR_PUT_UINT32(buf, v) IXDR_PUT_INT32((buf), ((int32_t)(v)))
#define IXDR_GET_LONG(buf) ((long)__ntohl((uint32_t)*(buf)++))
#define IXDR_PUT_LONG(buf, v) (*(buf)++ =(int32_t)__htonl((uint32_t)v))
#define IXDR_GET_BOOL(buf) ((bool_t)IXDR_GET_LONG(buf))
#define IXDR_GET_ENUM(buf, t) ((t)IXDR_GET_LONG(buf))
#define IXDR_GET_U_LONG(buf) ((u_long)IXDR_GET_LONG(buf))
#define IXDR_GET_SHORT(buf) ((short)IXDR_GET_LONG(buf))
#define IXDR_GET_U_SHORT(buf) ((u_short)IXDR_GET_LONG(buf))
#define IXDR_PUT_BOOL(buf, v) IXDR_PUT_LONG((buf), (v))
#define IXDR_PUT_ENUM(buf, v) IXDR_PUT_LONG((buf), (v))
#define IXDR_PUT_U_LONG(buf, v) IXDR_PUT_LONG((buf), (v))
#define IXDR_PUT_SHORT(buf, v) IXDR_PUT_LONG((buf), (v))
#define IXDR_PUT_U_SHORT(buf, v) IXDR_PUT_LONG((buf), (v))
/*
* These are the "generic" xdr routines.
*/
__BEGIN_DECLS
extern bool_t xdr_void(void);
extern bool_t xdr_int(XDR *, int *);
extern bool_t xdr_u_int(XDR *, u_int *);
extern bool_t xdr_long(XDR *, long *);
extern bool_t xdr_u_long(XDR *, u_long *);
extern bool_t xdr_short(XDR *, short *);
extern bool_t xdr_u_short(XDR *, u_short *);
extern bool_t xdr_int16_t(XDR *, int16_t *);
extern bool_t xdr_uint16_t(XDR *, uint16_t *);
extern bool_t xdr_int32_t(XDR *, int32_t *);
extern bool_t xdr_uint32_t(XDR *, uint32_t *);
extern bool_t xdr_int64_t(XDR *, int64_t *);
extern bool_t xdr_uint64_t(XDR *, uint64_t *);
extern bool_t xdr_bool(XDR *, bool_t *);
extern bool_t xdr_enum(XDR *, enum_t *);
extern bool_t xdr_array(XDR *, char **, u_int *, u_int, u_int, xdrproc_t);
extern bool_t xdr_bytes(XDR *, char **, u_int *, u_int);
extern bool_t xdr_opaque(XDR *, char *, u_int);
extern bool_t xdr_string(XDR *, char **, u_int);
extern bool_t xdr_union(XDR *, enum_t *, char *, const struct xdr_discrim *, xdrproc_t);
extern bool_t xdr_char(XDR *, char *);
extern bool_t xdr_u_char(XDR *, u_char *);
extern bool_t xdr_vector(XDR *, char *, u_int, u_int, xdrproc_t);
extern bool_t xdr_float(XDR *, float *);
extern bool_t xdr_double(XDR *, double *);
extern bool_t xdr_quadruple(XDR *, long double *);
extern bool_t xdr_reference(XDR *, char **, u_int, xdrproc_t);
extern bool_t xdr_pointer(XDR *, char **, u_int, xdrproc_t);
extern bool_t xdr_wrapstring(XDR *, char **);
extern void xdr_free(xdrproc_t, void *);
extern bool_t xdr_hyper(XDR *, quad_t *);
extern bool_t xdr_u_hyper(XDR *, u_quad_t *);
extern bool_t xdr_longlong_t(XDR *, quad_t *);
extern bool_t xdr_u_longlong_t(XDR *, u_quad_t *);
extern unsigned long xdr_sizeof(xdrproc_t func, void *data);
__END_DECLS
/*
* Common opaque bytes objects used by many rpc protocols;
* declared here due to commonality.
*/
#define MAX_NETOBJ_SZ 1024
struct netobj {
u_int n_len;
char *n_bytes;
};
typedef struct netobj netobj;
extern bool_t xdr_netobj(XDR *, struct netobj *);
/*
* These are the public routines for the various implementations of
* xdr streams.
*/
__BEGIN_DECLS
/* XDR using memory buffers */
extern void xdrmem_create(XDR *, char *, u_int, enum xdr_op);
/* XDR using mbufs */
struct mbuf;
extern void xdrmbuf_create(XDR *, struct mbuf *, enum xdr_op);
Implement support for RPCSEC_GSS authentication to both the NFS client and server. This replaces the RPC implementation of the NFS client and server with the newer RPC implementation originally developed (actually ported from the userland sunrpc code) to support the NFS Lock Manager. I have tested this code extensively and I believe it is stable and that performance is at least equal to the legacy RPC implementation. The NFS code currently contains support for both the new RPC implementation and the older legacy implementation inherited from the original NFS codebase. The default is to use the new implementation - add the NFS_LEGACYRPC option to fall back to the old code. When I merge this support back to RELENG_7, I will probably change this so that users have to 'opt in' to get the new code. To use RPCSEC_GSS on either client or server, you must build a kernel which includes the KGSSAPI option and the crypto device. On the userland side, you must build at least a new libc, mountd, mount_nfs and gssd. You must install new versions of /etc/rc.d/gssd and /etc/rc.d/nfsd and add 'gssd_enable=YES' to /etc/rc.conf. As long as gssd is running, you should be able to mount an NFS filesystem from a server that requires RPCSEC_GSS authentication. The mount itself can happen without any kerberos credentials but all access to the filesystem will be denied unless the accessing user has a valid ticket file in the standard place (/tmp/krb5cc_<uid>). There is currently no support for situations where the ticket file is in a different place, such as when the user logged in via SSH and has delegated credentials from that login. This restriction is also present in Solaris and Linux. In theory, we could improve this in future, possibly using Brooks Davis' implementation of variant symlinks. Supporting RPCSEC_GSS on a server is nearly as simple. You must create service creds for the server in the form 'nfs/<fqdn>@<REALM>' and install them in /etc/krb5.keytab. The standard heimdal utility ktutil makes this fairly easy. After the service creds have been created, you can add a '-sec=krb5' option to /etc/exports and restart both mountd and nfsd. The only other difference an administrator should notice is that nfsd doesn't fork to create service threads any more. In normal operation, there will be two nfsd processes, one in userland waiting for TCP connections and one in the kernel handling requests. The latter process will create as many kthreads as required - these should be visible via 'top -H'. The code has some support for varying the number of service threads according to load but initially at least, nfsd uses a fixed number of threads according to the value supplied to its '-n' option. Sponsored by: Isilon Systems MFC after: 1 month
2008-11-03 10:38:00 +00:00
extern void xdrmbuf_append(XDR *, struct mbuf *);
extern struct mbuf * xdrmbuf_getall(XDR *);
Add the new kernel-mode NFS Lock Manager. To use it instead of the user-mode lock manager, build a kernel with the NFSLOCKD option and add '-k' to 'rpc_lockd_flags' in rc.conf. Highlights include: * Thread-safe kernel RPC client - many threads can use the same RPC client handle safely with replies being de-multiplexed at the socket upcall (typically driven directly by the NIC interrupt) and handed off to whichever thread matches the reply. For UDP sockets, many RPC clients can share the same socket. This allows the use of a single privileged UDP port number to talk to an arbitrary number of remote hosts. * Single-threaded kernel RPC server. Adding support for multi-threaded server would be relatively straightforward and would follow approximately the Solaris KPI. A single thread should be sufficient for the NLM since it should rarely block in normal operation. * Kernel mode NLM server supporting cancel requests and granted callbacks. I've tested the NLM server reasonably extensively - it passes both my own tests and the NFS Connectathon locking tests running on Solaris, Mac OS X and Ubuntu Linux. * Userland NLM client supported. While the NLM server doesn't have support for the local NFS client's locking needs, it does have to field async replies and granted callbacks from remote NLMs that the local client has contacted. We relay these replies to the userland rpc.lockd over a local domain RPC socket. * Robust deadlock detection for the local lock manager. In particular it will detect deadlocks caused by a lock request that covers more than one blocking request. As required by the NLM protocol, all deadlock detection happens synchronously - a user is guaranteed that if a lock request isn't rejected immediately, the lock will eventually be granted. The old system allowed for a 'deferred deadlock' condition where a blocked lock request could wake up and find that some other deadlock-causing lock owner had beaten them to the lock. * Since both local and remote locks are managed by the same kernel locking code, local and remote processes can safely use file locks for mutual exclusion. Local processes have no fairness advantage compared to remote processes when contending to lock a region that has just been unlocked - the local lock manager enforces a strict first-come first-served model for both local and remote lockers. Sponsored by: Isilon Systems PR: 95247 107555 115524 116679 MFC after: 2 weeks
2008-03-26 15:23:12 +00:00
/* XDR pseudo records for tcp */
extern void xdrrec_create(XDR *, u_int, u_int, void *,
int (*)(void *, void *, int),
int (*)(void *, void *, int));
/* make end of xdr record */
extern bool_t xdrrec_endofrecord(XDR *, int);
/* move to beginning of next record */
extern bool_t xdrrec_skiprecord(XDR *);
/* true if no more input */
extern bool_t xdrrec_eof(XDR *);
extern u_int xdrrec_readbytes(XDR *, caddr_t, u_int);
__END_DECLS
#endif /* !_KRPC_XDR_H */