| 1 | /*	$OpenBSD: nfs.h,v 1.54 2024/05/04 10:53:37 jsg Exp $	*/ |
| 2 | /*	$NetBSD: nfs.h,v 1.10.4.1 1996/05/27 11:23:56 fvdl Exp $	*/ |
| 3 | |
| 4 | /* |
| 5 | * Copyright (c) 1989, 1993, 1995 |
| 6 | *	The Regents of the University of California. All rights reserved. |
| 7 | * |
| 8 | * This code is derived from software contributed to Berkeley by |
| 9 | * Rick Macklem at The University of Guelph. |
| 10 | * |
| 11 | * Redistribution and use in source and binary forms, with or without |
| 12 | * modification, are permitted provided that the following conditions |
| 13 | * are met: |
| 14 | * 1. Redistributions of source code must retain the above copyright |
| 15 | * notice, this list of conditions and the following disclaimer. |
| 16 | * 2. Redistributions in binary form must reproduce the above copyright |
| 17 | * notice, this list of conditions and the following disclaimer in the |
| 18 | * documentation and/or other materials provided with the distribution. |
| 19 | * 3. Neither the name of the University nor the names of its contributors |
| 20 | * may be used to endorse or promote products derived from this software |
| 21 | * without specific prior written permission. |
| 22 | * |
| 23 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
| 24 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 25 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 26 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE |
| 27 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
| 28 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
| 29 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
| 30 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
| 31 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
| 32 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
| 33 | * SUCH DAMAGE. |
| 34 | * |
| 35 | *	@(#)nfs.h	8.4 (Berkeley) 5/1/95 |
| 36 | */ |
| 37 | |
| 38 | #ifndef _NFS_NFS_H_ |
| 39 | #define _NFS_NFS_H_ |
| 40 | |
| 41 | #define NFS_TICKINTVL	5		/* Desired time for a tick (msec) */ |
| 42 | #define NFS_HZ		(hz / nfs_ticks) /* Ticks/sec */ |
| 43 | #define	NFS_TIMEO	(1 * NFS_HZ)	/* Default timeout = 1 second */ |
| 44 | #define	NFS_MINTIMEO	(1 * NFS_HZ)	/* Min timeout to use */ |
| 45 | #define	NFS_MAXTIMEO	(60 * NFS_HZ)	/* Max timeout to backoff to */ |
| 46 | #define	NFS_TIMEOUTMUL	2		/* Timeout/Delay multiplier */ |
| 47 | #define	NFS_MAXREXMIT	100		/* Stop counting after this many */ |
| 48 | #define	NFS_RETRANS	10		/* Num of retrans for soft mounts */ |
| 49 | #define	NFS_MAXGRPS	16		/* Max. size of groups list */ |
| 50 | #define	NFS_MINATTRTIMO 5		/* Attribute cache timeout in sec */ |
| 51 | #define	NFS_MAXATTRTIMO 60 |
| 52 | #define	NFS_WSIZE	8192		/* Def. write data size <= 8192 */ |
| 53 | #define	NFS_RSIZE	8192		/* Def. read data size <= 8192 */ |
| 54 | #define NFS_READDIRSIZE	8192		/* Def. readdir size */ |
| 55 | #define	NFS_DEFRAHEAD	1		/* Def. read ahead # blocks */ |
| 56 | #define	NFS_MAXRAHEAD	4		/* Max. read ahead # blocks */ |
| 57 | #define	NFS_MAXASYNCDAEMON 	20	/* Max. number async_daemons runable */ |
| 58 | |
| 59 | /* |
| 60 | * Ideally, NFS_DIRBLKSIZ should be bigger, but I've seen servers with |
| 61 | * broken NFS/ethernet drivers that won't work with anything bigger (Linux..) |
| 62 | */ |
| 63 | #define	NFS_DIRBLKSIZ	1024		/* Must be a multiple of DIRBLKSIZ */ |
| 64 | #define NFS_READDIRBLKSIZ	512	/* Size of read dir blocks. XXX */ |
| 65 | |
| 66 | /* |
| 67 | * Oddballs |
| 68 | */ |
| 69 | #define NFS_CMPFH(n, f, s) \ |
| 70 | 	((n)->n_fhsize == (s) && !bcmp((caddr_t)(n)->n_fhp, (caddr_t)(f), (s))) |
| 71 | #define NFS_ISV3(v)	(VFSTONFS((v)->v_mount)->nm_flag & NFSMNT_NFSV3) |
| 72 | #define NFS_SRVMAXDATA(n) \ |
| 73 | 		(((n)->nd_flag & ND_NFSV3) ? (((n)->nd_nam2) ? \ |
| 74 | 		 NFS_MAXDGRAMDATA : NFS_MAXDATA) : NFS_V2MAXDATA) |
| 75 | |
| 76 | /* |
| 77 | * The B_INVAFTERWRITE flag should be set to whatever is required by the |
| 78 | * buffer cache code to say "Invalidate the block after it is written back". |
| 79 | */ |
| 80 | #define	B_INVAFTERWRITE	B_INVAL |
| 81 | |
| 82 | /* |
| 83 | * Structures for the nfssvc(2) syscall. |
| 84 | * Not that anyone besides nfsd(8) should ever use it. |
| 85 | */ |
| 86 | struct nfsd_args { |
| 87 | 	int	sock;		/* Socket to serve */ |
| 88 | 	caddr_t	name;		/* Client addr for connection based sockets */ |
| 89 | 	int	namelen;	/* Length of name */ |
| 90 | }; |
| 91 | |
| 92 | struct nfsd_srvargs { |
| 93 | 	struct nfsd	*nsd_nfsd;	/* Pointer to in kernel nfsd struct */ |
| 94 | 	uid_t		nsd_uid;	/* Effective uid mapped to cred */ |
| 95 | 	u_int32_t	nsd_haddr;	/* IP address of client */ |
| 96 | 	struct xucred	nsd_cr;		/* Cred. uid maps to */ |
| 97 | 	int		nsd_authlen;	/* Length of auth string (ret) */ |
| 98 | 	u_char		*nsd_authstr;	/* Auth string (ret) */ |
| 99 | 	int		nsd_verflen;	/* and the verifier */ |
| 100 | 	u_char		*nsd_verfstr; |
| 101 | 	struct timeval	nsd_timestamp;	/* timestamp from verifier */ |
| 102 | 	u_int32_t	nsd_ttl;	/* credential ttl (sec) */ |
| 103 | }; |
| 104 | |
| 105 | /* |
| 106 | * Stats structure |
| 107 | */ |
| 108 | struct nfsstats { |
| 109 | 	uint64_t	attrcache_hits; |
| 110 | 	uint64_t	attrcache_misses; |
| 111 | 	uint64_t	lookupcache_hits; |
| 112 | 	uint64_t	lookupcache_misses; |
| 113 | 	uint64_t	direofcache_hits; |
| 114 | 	uint64_t	direofcache_misses; |
| 115 | 	uint64_t	biocache_reads; |
| 116 | 	uint64_t	read_bios; |
| 117 | 	uint64_t	read_physios; |
| 118 | 	uint64_t	biocache_writes; |
| 119 | 	uint64_t	write_bios; |
| 120 | 	uint64_t	write_physios; |
| 121 | 	uint64_t	biocache_readlinks; |
| 122 | 	uint64_t	readlink_bios; |
| 123 | 	uint64_t	biocache_readdirs; |
| 124 | 	uint64_t	readdir_bios; |
| 125 | 	uint64_t	rpccnt[NFS_NPROCS]; |
| 126 | 	uint64_t	rpcretries; |
| 127 | 	uint64_t	srvrpccnt[NFS_NPROCS]; |
| 128 | 	uint64_t	srvrpc_errs; |
| 129 | 	uint64_t	srv_errs; |
| 130 | 	uint64_t	rpcrequests; |
| 131 | 	uint64_t	rpctimeouts; |
| 132 | 	uint64_t	rpcunexpected; |
| 133 | 	uint64_t	rpcinvalid; |
| 134 | 	uint64_t	srvcache_inproghits; |
| 135 | 	uint64_t	srvcache_idemdonehits; |
| 136 | 	uint64_t	srvcache_nonidemdonehits; |
| 137 | 	uint64_t	srvcache_misses; |
| 138 | 	uint64_t	forcedsync; |
| 139 | 	uint64_t	srvnqnfs_leases; |
| 140 | 	uint64_t	srvnqnfs_maxleases; |
| 141 | 	uint64_t	srvnqnfs_getleases; |
| 142 | 	uint64_t	srvvop_writes; |
| 143 | }; |
| 144 | |
| 145 | /* |
| 146 | * Flags for nfssvc() system call. |
| 147 | */ |
| 148 | #define	NFSSVC_NFSD	0x004 |
| 149 | #define	NFSSVC_ADDSOCK	0x008 |
| 150 | |
| 151 | /* |
| 152 | * fs.nfs sysctl(3) identifiers |
| 153 | */ |
| 154 | #define	NFS_NFSSTATS	1	/* struct: struct nfsstats */ |
| 155 | #define	NFS_NIOTHREADS	2	/* number of i/o threads */ |
| 156 | #define	NFS_MAXID	3 |
| 157 | |
| 158 | #define FS_NFS_NAMES { \ |
| 159 | 			{ 0, 0 }, \ |
| 160 | 			{ "nfsstats", CTLTYPE_STRUCT }, \ |
| 161 | 			{ "iothreads", CTLTYPE_INT } \ |
| 162 | } |
| 163 | |
| 164 | /* |
| 165 | * The set of signals the interrupt an I/O in progress for NFSMNT_INT mounts. |
| 166 | * What should be in this set is open to debate, but I believe that since |
| 167 | * I/O system calls on ufs are never interrupted by signals the set should |
| 168 | * be minimal. My reasoning is that many current programs that use signals |
| 169 | * such as SIGALRM will not expect file I/O system calls to be interrupted |
| 170 | * by them and break. |
| 171 | */ |
| 172 | #ifdef _KERNEL |
| 173 | extern int nfs_niothreads; |
| 174 | |
| 175 | struct uio; struct buf; struct vattr; struct nameidata;	/* XXX */ |
| 176 | |
| 177 | #define	NFSINT_SIGMASK	(sigmask(SIGINT)|sigmask(SIGTERM)|sigmask(SIGKILL)| \ |
| 178 | 			 sigmask(SIGHUP)|sigmask(SIGQUIT)) |
| 179 | |
| 180 | /* |
| 181 | * Socket errors ignored for connectionless sockets?? |
| 182 | * For now, ignore them all |
| 183 | */ |
| 184 | #define	NFSIGNORE_SOERROR(s, e) \ |
| 185 | 		((e) != EINTR && (e) != ERESTART && (e) != EWOULDBLOCK && \ |
| 186 | 		((s) & PR_CONNREQUIRED) == 0) |
| 187 | |
| 188 | /* |
| 189 | * Nfs outstanding request list element |
| 190 | */ |
| 191 | struct nfsreq { |
| 192 | 	TAILQ_ENTRY(nfsreq) r_chain; |
| 193 | 	struct mbuf	*r_mreq; |
| 194 | 	struct mbuf	*r_mrep; |
| 195 | 	struct mbuf	*r_md; |
| 196 | 	caddr_t		r_dpos; |
| 197 | 	struct nfsmount *r_nmp; |
| 198 | 	struct vnode	*r_vp; |
| 199 | 	u_int32_t	r_xid; |
| 200 | 	int		r_flags;	/* flags on request, see below */ |
| 201 | 	int		r_rexmit;	/* current retrans count */ |
| 202 | 	int		r_timer;	/* tick counter on reply */ |
| 203 | 	int		r_procnum;	/* NFS procedure number */ |
| 204 | 	int		r_rtt;		/* RTT for rpc */ |
| 205 | 	struct proc	*r_procp;	/* Proc that did I/O system call */ |
| 206 | }; |
| 207 | |
| 208 | /* Flag values for r_flags */ |
| 209 | #define R_TIMING	0x01		/* timing request (in mntp) */ |
| 210 | #define R_SENT		0x02		/* request has been sent */ |
| 211 | #define	R_SOFTTERM	0x04		/* soft mnt, too many retries */ |
| 212 | #define	R_INTR		0x08		/* intr mnt, signal pending */ |
| 213 | #define	R_SOCKERR	0x10		/* Fatal error on socket */ |
| 214 | #define	R_TPRINTFMSG	0x20		/* Did a tprintf msg. */ |
| 215 | #define	R_MUSTRESEND	0x40		/* Must resend request */ |
| 216 | |
| 217 | /* |
| 218 | * On fast networks, the estimator will try to reduce the |
| 219 | * timeout lower than the latency of the server's disks, |
| 220 | * which results in too many timeouts, so cap the lower |
| 221 | * bound. |
| 222 | */ |
| 223 | #define NFS_MINRTO	(NFS_HZ >> 2) |
| 224 | |
| 225 | /* |
| 226 | * Keep the RTO from increasing to unreasonably large values |
| 227 | * when a server is not responding. |
| 228 | */ |
| 229 | #define NFS_MAXRTO	(20 * NFS_HZ) |
| 230 | |
| 231 | enum nfs_rto_timers { |
| 232 | 	NFS_DEFAULT_TIMER, |
| 233 | 	NFS_GETATTR_TIMER, |
| 234 | 	NFS_LOOKUP_TIMER, |
| 235 | 	NFS_READ_TIMER, |
| 236 | 	NFS_WRITE_TIMER, |
| 237 | }; |
| 238 | #define NFS_MAX_TIMER	(NFS_WRITE_TIMER) |
| 239 | |
| 240 | #define NFS_INITRTT	(NFS_HZ << 3) |
| 241 | |
| 242 | /* |
| 243 | * Network address hash list element |
| 244 | */ |
| 245 | union nethostaddr { |
| 246 | 	u_int32_t had_inetaddr; |
| 247 | 	struct mbuf *had_nam; |
| 248 | }; |
| 249 | |
| 250 | struct nfssvc_sock { |
| 251 | 	TAILQ_ENTRY(nfssvc_sock) ns_chain; /* List of all nfssvc_sock's */ |
| 252 | 	struct file	*ns_fp;		/* fp from the... */ |
| 253 | 	struct socket	*ns_so;		/* ...socket this struct wraps */ |
| 254 | 	struct mbuf	*ns_nam;	/* MT_SONAME of client */ |
| 255 | 	struct mbuf	*ns_raw;	/* head of unpeeked mbufs */ |
| 256 | 	struct mbuf	*ns_rawend;	/* tail of unpeeked mbufs */ |
| 257 | 	struct mbuf	*ns_rec;	/* queued RPC records */ |
| 258 | 	struct mbuf	*ns_recend;	/* last queued RPC record */ |
| 259 | 	struct mbuf	*ns_frag;	/* end of record fragment */ |
| 260 | 	int		ns_flag;	/* socket status flags */ |
| 261 | 	int		ns_solock;	/* lock for connected socket */ |
| 262 | 	int		ns_cc;		/* actual chars queued */ |
| 263 | 	int		ns_reclen;	/* length of first queued record */ |
| 264 | 	u_int32_t	ns_sref;	/* # of refs to this struct */ |
| 265 | }; |
| 266 | |
| 267 | /* Bits for "ns_flag" */ |
| 268 | #define	SLP_VALID	0x01	/* connection is usable */ |
| 269 | #define	SLP_DOREC	0x02	/* receive operation required */ |
| 270 | #define	SLP_NEEDQ	0x04	/* connection has data to queue from socket */ |
| 271 | #define	SLP_DISCONN	0x08	/* connection is closed */ |
| 272 | #define	SLP_GETSTREAM	0x10	/* extracting RPC from TCP connection */ |
| 273 | #define	SLP_LASTFRAG	0x20	/* last fragment received on TCP connection */ |
| 274 | #define	SLP_ALLFLAGS	0xff	/* convenience */ |
| 275 | |
| 276 | /* |
| 277 | * One of these structures is allocated for each nfsd. |
| 278 | */ |
| 279 | struct nfsd { |
| 280 | 	TAILQ_ENTRY(nfsd) nfsd_chain;	/* List of all nfsd's */ |
| 281 | 	int		nfsd_flag;	/* NFSD_ flags */ |
| 282 | 	struct nfssvc_sock *nfsd_slp;	/* Current socket */ |
| 283 | 	struct proc	*nfsd_procp;	/* Proc ptr */ |
| 284 | 	struct nfsrv_descript *nfsd_nd;	/* Associated nfsrv_descript */ |
| 285 | }; |
| 286 | |
| 287 | /* Bits for "nfsd_flag" */ |
| 288 | #define	NFSD_WAITING	0x01 |
| 289 | #define	NFSD_REQINPROG	0x02 |
| 290 | |
| 291 | /* |
| 292 | * This structure is used by the server for describing each request. |
| 293 | */ |
| 294 | struct nfsrv_descript { |
| 295 | 	struct mbuf		*nd_mrep;	/* Request mbuf list */ |
| 296 | 	struct mbuf		*nd_md;		/* Current dissect mbuf */ |
| 297 | 	struct mbuf		*nd_nam;	/* and socket addr */ |
| 298 | 	struct mbuf		*nd_nam2;	/* return socket addr */ |
| 299 | 	caddr_t			nd_dpos;	/* Current dissect pos */ |
| 300 | 	unsigned int		nd_procnum;	/* RPC # */ |
| 301 | 	int			nd_flag;	/* nd_flag */ |
| 302 | 	int			nd_repstat;	/* Reply status */ |
| 303 | 	u_int32_t		nd_retxid;	/* Reply xid */ |
| 304 | 	struct ucred		nd_cr;		/* Credentials */ |
| 305 | }; |
| 306 | |
| 307 | /* Bits for "nd_flag" */ |
| 308 | #define ND_NFSV3	0x08 |
| 309 | |
| 310 | extern struct pool nfsreqpl; |
| 311 | extern struct pool nfs_node_pool; |
| 312 | extern TAILQ_HEAD(nfsdhead, nfsd) nfsd_head; |
| 313 | extern int nfsd_head_flag; |
| 314 | #define	NFSD_CHECKSLP	0x01 |
| 315 | |
| 316 | #endif	/* _KERNEL */ |
| 317 | #endif /* _NFS_NFS_H */ |