sys/vfs/hammer2/hammer2_vfsops.c

   1 /*-
   2  * Copyright (c) 2011, 2012 The DragonFly Project.  All rights reserved.
   3  *
   4  * This code is derived from software contributed to The DragonFly Project
   5  * by Matthew Dillon <dillon@backplane.com>
   6  *
   7  * Redistribution and use in source and binary forms, with or without
   8  * modification, are permitted provided that the following conditions
   9  * are met:
  10  *
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in
  15  *    the documentation and/or other materials provided with the
  16  *    distribution.
  17  * 3. Neither the name of The DragonFly Project nor the names of its
  18  *    contributors may be used to endorse or promote products derived
  19  *    from this software without specific, prior written permission.
  20  *
  21  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  22  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  23  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
  24  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE
  25  * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
  26  * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
  27  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  28  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
  29  * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
  30  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
  31  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  32  * SUCH DAMAGE.
  33  */
  34 #include <sys/param.h>
  35 #include <sys/systm.h>
  36 #include <sys/kernel.h>
  37 #include <sys/nlookup.h>
  38 #include <sys/vnode.h>
  39 #include <sys/mount.h>
  40 #include <sys/fcntl.h>
  41 #include <sys/buf.h>
  42 #include <sys/uuid.h>
  43 #include <sys/vfsops.h>
  44 #include <sys/sysctl.h>
  45 #include <sys/socket.h>
  46
  47 #include "hammer2.h"
  48 #include "hammer2_disk.h"
  49 #include "hammer2_mount.h"
  50 #include "hammer2_network.h"
  51
  52 struct hammer2_sync_info {
  53         int error;
  54         int waitfor;
  55 };
  56
  57 TAILQ_HEAD(hammer2_mntlist, hammer2_mount);
  58 static struct hammer2_mntlist hammer2_mntlist;
  59 static struct lock hammer2_mntlk;
  60
  61 int hammer2_debug;
  62 int hammer2_cluster_enable = 1;
  63 int hammer2_hardlink_enable = 1;
  64 long hammer2_iod_file_read;
  65 long hammer2_iod_meta_read;
  66 long hammer2_iod_indr_read;
  67 long hammer2_iod_file_write;
  68 long hammer2_iod_meta_write;
  69 long hammer2_iod_indr_write;
  70 long hammer2_iod_volu_write;
  71 long hammer2_ioa_file_read;
  72 long hammer2_ioa_meta_read;
  73 long hammer2_ioa_indr_read;
  74 long hammer2_ioa_file_write;
  75 long hammer2_ioa_meta_write;
  76 long hammer2_ioa_indr_write;
  77 long hammer2_ioa_volu_write;
  78
  79 SYSCTL_NODE(_vfs, OID_AUTO, hammer2, CTLFLAG_RW, 0, "HAMMER2 filesystem");
  80
  81 SYSCTL_INT(_vfs_hammer2, OID_AUTO, debug, CTLFLAG_RW,
  82            &hammer2_debug, 0, "");
  83 SYSCTL_INT(_vfs_hammer2, OID_AUTO, cluster_enable, CTLFLAG_RW,
  84            &hammer2_cluster_enable, 0, "");
  85 SYSCTL_INT(_vfs_hammer2, OID_AUTO, hardlink_enable, CTLFLAG_RW,
  86            &hammer2_hardlink_enable, 0, "");
  87 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, iod_file_read, CTLFLAG_RW,
  88            &hammer2_iod_file_read, 0, "");
  89 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, iod_meta_read, CTLFLAG_RW,
  90            &hammer2_iod_meta_read, 0, "");
  91 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, iod_indr_read, CTLFLAG_RW,
  92            &hammer2_iod_indr_read, 0, "");
  93 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, iod_file_write, CTLFLAG_RW,
  94            &hammer2_iod_file_write, 0, "");
  95 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, iod_meta_write, CTLFLAG_RW,
  96            &hammer2_iod_meta_write, 0, "");
  97 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, iod_indr_write, CTLFLAG_RW,
  98            &hammer2_iod_indr_write, 0, "");
  99 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, iod_volu_write, CTLFLAG_RW,
 100            &hammer2_iod_volu_write, 0, "");
 101 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, ioa_file_read, CTLFLAG_RW,
 102            &hammer2_ioa_file_read, 0, "");
 103 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, ioa_meta_read, CTLFLAG_RW,
 104            &hammer2_ioa_meta_read, 0, "");
 105 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, ioa_indr_read, CTLFLAG_RW,
 106            &hammer2_ioa_indr_read, 0, "");
 107 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, ioa_file_write, CTLFLAG_RW,
 108            &hammer2_ioa_file_write, 0, "");
 109 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, ioa_meta_write, CTLFLAG_RW,
 110            &hammer2_ioa_meta_write, 0, "");
 111 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, ioa_indr_write, CTLFLAG_RW,
 112            &hammer2_ioa_indr_write, 0, "");
 113 SYSCTL_LONG(_vfs_hammer2, OID_AUTO, ioa_volu_write, CTLFLAG_RW,
 114            &hammer2_ioa_volu_write, 0, "");
 115
 116 static int hammer2_vfs_init(struct vfsconf *conf);
 117 static int hammer2_vfs_mount(struct mount *mp, char *path, caddr_t data,
 118                                 struct ucred *cred);
 119 static int hammer2_remount(struct mount *, char *, struct vnode *,
 120                                 struct ucred *);
 121 static int hammer2_vfs_unmount(struct mount *mp, int mntflags);
 122 static int hammer2_vfs_root(struct mount *mp, struct vnode **vpp);
 123 static int hammer2_vfs_statfs(struct mount *mp, struct statfs *sbp,
 124                                 struct ucred *cred);
 125 static int hammer2_vfs_statvfs(struct mount *mp, struct statvfs *sbp,
 126                                 struct ucred *cred);
 127 static int hammer2_vfs_sync(struct mount *mp, int waitfor);
 128 static int hammer2_vfs_vget(struct mount *mp, struct vnode *dvp,
 129                                 ino_t ino, struct vnode **vpp);
 130 static int hammer2_vfs_fhtovp(struct mount *mp, struct vnode *rootvp,
 131                                 struct fid *fhp, struct vnode **vpp);
 132 static int hammer2_vfs_vptofh(struct vnode *vp, struct fid *fhp);
 133 static int hammer2_vfs_checkexp(struct mount *mp, struct sockaddr *nam,
 134                                 int *exflagsp, struct ucred **credanonp);
 135
 136 static int hammer2_install_volume_header(hammer2_mount_t *hmp);
 137 static int hammer2_sync_scan1(struct mount *mp, struct vnode *vp, void *data);
 138 static int hammer2_sync_scan2(struct mount *mp, struct vnode *vp, void *data);
 139
 140 static void hammer2_cluster_thread_rd(void *arg);
 141 static void hammer2_cluster_thread_wr(void *arg);
 142 static int hammer2_msg_span_reply(hammer2_pfsmount_t *pmp, hammer2_msg_t *msg);
 143
 144 /*
 145  * HAMMER2 vfs operations.
 146  */
 147 static struct vfsops hammer2_vfsops = {
 148         .vfs_init       = hammer2_vfs_init,
 149         .vfs_sync       = hammer2_vfs_sync,
 150         .vfs_mount      = hammer2_vfs_mount,
 151         .vfs_unmount    = hammer2_vfs_unmount,
 152         .vfs_root       = hammer2_vfs_root,
 153         .vfs_statfs     = hammer2_vfs_statfs,
 154         .vfs_statvfs    = hammer2_vfs_statvfs,
 155         .vfs_vget       = hammer2_vfs_vget,
 156         .vfs_vptofh     = hammer2_vfs_vptofh,
 157         .vfs_fhtovp     = hammer2_vfs_fhtovp,
 158         .vfs_checkexp   = hammer2_vfs_checkexp
 159 };
 160
 161 MALLOC_DEFINE(M_HAMMER2, "HAMMER2-mount", "");
 162
 163 VFS_SET(hammer2_vfsops, hammer2, 0);
 164 MODULE_VERSION(hammer2, 1);
 165
 166 static
 167 int
 168 hammer2_vfs_init(struct vfsconf *conf)
 169 {
 170         int error;
 171
 172         error = 0;
 173
 174         if (HAMMER2_BLOCKREF_BYTES != sizeof(struct hammer2_blockref))
 175                 error = EINVAL;
 176         if (HAMMER2_INODE_BYTES != sizeof(struct hammer2_inode_data))
 177                 error = EINVAL;
 178         if (HAMMER2_ALLOCREF_BYTES != sizeof(struct hammer2_allocref))
 179                 error = EINVAL;
 180         if (HAMMER2_VOLUME_BYTES != sizeof(struct hammer2_volume_data))
 181                 error = EINVAL;
 182
 183         if (error)
 184                 kprintf("HAMMER2 structure size mismatch; cannot continue.\n");
 185
 186         lockinit(&hammer2_mntlk, "mntlk", 0, 0);
 187         TAILQ_INIT(&hammer2_mntlist);
 188
 189         return (error);
 190 }
 191
 192 /*
 193  * Mount or remount HAMMER2 fileystem from physical media
 194  *
 195  *      mountroot
 196  *              mp              mount point structure
 197  *              path            NULL
 198  *              data            <unused>
 199  *              cred            <unused>
 200  *
 201  *      mount
 202  *              mp              mount point structure
 203  *              path            path to mount point
 204  *              data            pointer to argument structure in user space
 205  *                      volume  volume path (device@LABEL form)
 206  *                      hflags  user mount flags
 207  *              cred            user credentials
 208  *
 209  * RETURNS:     0       Success
 210  *              !0      error number
 211  */
 212 static
 213 int
 214 hammer2_vfs_mount(struct mount *mp, char *path, caddr_t data,
 215                   struct ucred *cred)
 216 {
 217         struct hammer2_mount_info info;
 218         hammer2_pfsmount_t *pmp;
 219         hammer2_mount_t *hmp;
 220         hammer2_key_t lhc;
 221         struct vnode *devvp;
 222         struct nlookupdata nd;
 223         hammer2_chain_t *parent;
 224         hammer2_chain_t *schain;
 225         hammer2_chain_t *rchain;
 226         char devstr[MNAMELEN];
 227         size_t size;
 228         size_t done;
 229         char *dev;
 230         char *label;
 231         int ronly = 1;
 232         int create_hmp;
 233         int error;
 234
 235         hmp = NULL;
 236         pmp = NULL;
 237         dev = NULL;
 238         label = NULL;
 239         devvp = NULL;
 240
 241         kprintf("hammer2_mount\n");
 242
 243         if (path == NULL) {
 244                 /*
 245                  * Root mount
 246                  */
 247                 bzero(&info, sizeof(info));
 248                 info.cluster_fd = -1;
 249                 return (EOPNOTSUPP);
 250         } else {
 251                 /*
 252                  * Non-root mount or updating a mount
 253                  */
 254                 error = copyin(data, &info, sizeof(info));
 255                 if (error)
 256                         return (error);
 257
 258                 error = copyinstr(info.volume, devstr, MNAMELEN - 1, &done);
 259                 if (error)
 260                         return (error);
 261
 262                 /* Extract device and label */
 263                 dev = devstr;
 264                 label = strchr(devstr, '@');
 265                 if (label == NULL ||
 266                     ((label + 1) - dev) > done) {
 267                         return (EINVAL);
 268                 }
 269                 *label = '\0';
 270                 label++;
 271                 if (*label == '\0')
 272                         return (EINVAL);
 273
 274                 if (mp->mnt_flag & MNT_UPDATE) {
 275                         /* Update mount */
 276                         /* HAMMER2 implements NFS export via mountctl */
 277                         hmp = MPTOHMP(mp);
 278                         devvp = hmp->devvp;
 279                         error = hammer2_remount(mp, path, devvp, cred);
 280                         return error;
 281                 }
 282         }
 283
 284         /*
 285          * PFS mount
 286          *
 287          * Lookup name and verify it refers to a block device.
 288          */
 289         error = nlookup_init(&nd, dev, UIO_SYSSPACE, NLC_FOLLOW);
 290         if (error == 0)
 291                 error = nlookup(&nd);
 292         if (error == 0)
 293                 error = cache_vref(&nd.nl_nch, nd.nl_cred, &devvp);
 294         nlookup_done(&nd);
 295
 296         if (error == 0) {
 297                 if (vn_isdisk(devvp, &error))
 298                         error = vfs_mountedon(devvp);
 299         }
 300
 301         /*
 302          * Determine if the device has already been mounted.  After this
 303          * check hmp will be non-NULL if we are doing the second or more
 304          * hammer2 mounts from the same device.
 305          */
 306         lockmgr(&hammer2_mntlk, LK_EXCLUSIVE);
 307         TAILQ_FOREACH(hmp, &hammer2_mntlist, mntentry) {
 308                 if (hmp->devvp == devvp)
 309                         break;
 310         }
 311
 312         /*
 313          * Open the device if this isn't a secondary mount
 314          */
 315         if (hmp) {
 316                 create_hmp = 0;
 317         } else {
 318                 create_hmp = 1;
 319                 if (error == 0 && vcount(devvp) > 0)
 320                         error = EBUSY;
 321
 322                 /*
 323                  * Now open the device
 324                  */
 325                 if (error == 0) {
 326                         ronly = ((mp->mnt_flag & MNT_RDONLY) != 0);
 327                         vn_lock(devvp, LK_EXCLUSIVE | LK_RETRY);
 328                         error = vinvalbuf(devvp, V_SAVE, 0, 0);
 329                         if (error == 0) {
 330                                 error = VOP_OPEN(devvp,
 331                                                  ronly ? FREAD : FREAD | FWRITE,
 332                                                  FSCRED, NULL);
 333                         }
 334                         vn_unlock(devvp);
 335                 }
 336                 if (error && devvp) {
 337                         vrele(devvp);
 338                         devvp = NULL;
 339                 }
 340                 if (error) {
 341                         lockmgr(&hammer2_mntlk, LK_RELEASE);
 342                         return error;
 343                 }
 344         }
 345
 346         /*
 347          * Block device opened successfully, finish initializing the
 348          * mount structure.
 349          *
 350          * From this point on we have to call hammer2_unmount() on failure.
 351          */
 352         pmp = kmalloc(sizeof(*pmp), M_HAMMER2, M_WAITOK | M_ZERO);
 353         mp->mnt_data = (qaddr_t)pmp;
 354         pmp->mp = mp;
 355         kmalloc_create(&pmp->mmsg, "HAMMER2-pfsmsg");
 356         lockinit(&pmp->msglk, "h2msg", 0, 0);
 357         TAILQ_INIT(&pmp->msgq);
 358         RB_INIT(&pmp->staterd_tree);
 359         RB_INIT(&pmp->statewr_tree);
 360
 361         if (create_hmp) {
 362                 hmp = kmalloc(sizeof(*hmp), M_HAMMER2, M_WAITOK | M_ZERO);
 363                 hmp->ronly = ronly;
 364                 hmp->devvp = devvp;
 365                 kmalloc_create(&hmp->minode, "HAMMER2-inodes");
 366                 kmalloc_create(&hmp->mchain, "HAMMER2-chains");
 367                 TAILQ_INSERT_TAIL(&hammer2_mntlist, hmp, mntentry);
 368         }
 369         ccms_domain_init(&pmp->ccms_dom);
 370         pmp->hmp = hmp;
 371         ++hmp->pmp_count;
 372         lockmgr(&hammer2_mntlk, LK_RELEASE);
 373         kprintf("hammer2_mount hmp=%p pmpcnt=%d\n", hmp, hmp->pmp_count);
 374
 375         mp->mnt_flag = MNT_LOCAL;
 376         mp->mnt_kern_flag |= MNTK_ALL_MPSAFE;   /* all entry pts are SMP */
 377
 378         if (create_hmp) {
 379                 /*
 380                  * vchain setup. vchain.data is special cased to NULL.
 381                  * vchain.refs is initialized and will never drop to 0.
 382                  */
 383                 hmp->vchain.refs = 1;
 384                 hmp->vchain.data = (void *)&hmp->voldata;
 385                 hmp->vchain.bref.type = HAMMER2_BREF_TYPE_VOLUME;
 386                 hmp->vchain.bref.data_off = 0 | HAMMER2_PBUFRADIX;
 387                 hmp->vchain.bref_flush = hmp->vchain.bref;
 388                 ccms_cst_init(&hmp->vchain.cst, NULL);
 389                 /* hmp->vchain.u.xxx is left NULL */
 390                 lockinit(&hmp->alloclk, "h2alloc", 0, 0);
 391                 lockinit(&hmp->voldatalk, "voldata", 0, LK_CANRECURSE);
 392
 393                 /*
 394                  * Install the volume header
 395                  */
 396                 error = hammer2_install_volume_header(hmp);
 397                 if (error) {
 398                         hammer2_vfs_unmount(mp, MNT_FORCE);
 399                         return error;
 400                 }
 401         }
 402
 403         /*
 404          * required mount structure initializations
 405          */
 406         mp->mnt_stat.f_iosize = HAMMER2_PBUFSIZE;
 407         mp->mnt_stat.f_bsize = HAMMER2_PBUFSIZE;
 408
 409         mp->mnt_vstat.f_frsize = HAMMER2_PBUFSIZE;
 410         mp->mnt_vstat.f_bsize = HAMMER2_PBUFSIZE;
 411
 412         /*
 413          * Optional fields
 414          */
 415         mp->mnt_iosize_max = MAXPHYS;
 416
 417         /*
 418          * First locate the super-root inode, which is key 0 relative to the
 419          * volume header's blockset.
 420          *
 421          * Then locate the root inode by scanning the directory keyspace
 422          * represented by the label.
 423          */
 424         if (create_hmp) {
 425                 parent = &hmp->vchain;
 426                 hammer2_chain_lock(hmp, parent, HAMMER2_RESOLVE_ALWAYS);
 427                 schain = hammer2_chain_lookup(hmp, &parent,
 428                                       HAMMER2_SROOT_KEY, HAMMER2_SROOT_KEY, 0);
 429                 hammer2_chain_unlock(hmp, parent);
 430                 if (schain == NULL) {
 431                         kprintf("hammer2_mount: invalid super-root\n");
 432                         hammer2_vfs_unmount(mp, MNT_FORCE);
 433                         return EINVAL;
 434                 }
 435                 hammer2_chain_ref(hmp, schain); /* for hmp->schain */
 436                 hmp->schain = schain;           /* left locked */
 437         } else {
 438                 schain = hmp->schain;
 439                 hammer2_chain_lock(hmp, schain, HAMMER2_RESOLVE_ALWAYS);
 440         }
 441
 442         parent = schain;
 443         lhc = hammer2_dirhash(label, strlen(label));
 444         rchain = hammer2_chain_lookup(hmp, &parent,
 445                                       lhc, lhc + HAMMER2_DIRHASH_LOMASK,
 446                                       0);
 447         while (rchain) {
 448                 if (rchain->bref.type == HAMMER2_BREF_TYPE_INODE &&
 449                     rchain->u.ip &&
 450                     strcmp(label, rchain->data->ipdata.filename) == 0) {
 451                         break;
 452                 }
 453                 rchain = hammer2_chain_next(hmp, &parent, rchain,
 454                                             lhc, lhc + HAMMER2_DIRHASH_LOMASK,
 455                                             0);
 456         }
 457         hammer2_chain_unlock(hmp, parent);
 458         if (rchain == NULL) {
 459                 kprintf("hammer2_mount: PFS label not found\n");
 460                 hammer2_vfs_unmount(mp, MNT_FORCE);
 461                 return EINVAL;
 462         }
 463         if (rchain->flags & HAMMER2_CHAIN_MOUNTED) {
 464                 hammer2_chain_unlock(hmp, rchain);
 465                 kprintf("hammer2_mount: PFS label already mounted!\n");
 466                 hammer2_vfs_unmount(mp, MNT_FORCE);
 467                 return EBUSY;
 468         }
 469         atomic_set_int(&rchain->flags, HAMMER2_CHAIN_MOUNTED);
 470
 471         hammer2_chain_ref(hmp, rchain); /* for pmp->rchain */
 472         hammer2_chain_unlock(hmp, rchain);
 473         pmp->rchain = rchain;           /* left held & unlocked */
 474         pmp->iroot = rchain->u.ip;      /* implied hold from rchain */
 475         pmp->iroot->pmp = pmp;
 476
 477         kprintf("iroot %p\n", pmp->iroot);
 478
 479         /*
 480          * Ref the cluster management messaging descriptor.  The mount
 481          * program deals with the other end of the communications pipe.
 482          */
 483         pmp->msg_fp = holdfp(curproc->p_fd, info.cluster_fd, -1);
 484         if (pmp->msg_fp == NULL) {
 485                 kprintf("hammer2_mount: bad cluster_fd!\n");
 486                 hammer2_vfs_unmount(mp, MNT_FORCE);
 487                 return EBADF;
 488         }
 489         lwkt_create(hammer2_cluster_thread_rd, pmp, &pmp->msgrd_td,
 490                     NULL, 0, -1, "hammer2-msgrd");
 491         lwkt_create(hammer2_cluster_thread_wr, pmp, &pmp->msgwr_td,
 492                     NULL, 0, -1, "hammer2-msgwr");
 493
 494         /*
 495          * Finish setup
 496          */
 497         vfs_getnewfsid(mp);
 498         vfs_add_vnodeops(mp, &hammer2_vnode_vops, &mp->mnt_vn_norm_ops);
 499         vfs_add_vnodeops(mp, &hammer2_spec_vops, &mp->mnt_vn_spec_ops);
 500         vfs_add_vnodeops(mp, &hammer2_fifo_vops, &mp->mnt_vn_fifo_ops);
 501
 502         copyinstr(info.volume, mp->mnt_stat.f_mntfromname, MNAMELEN - 1, &size);
 503         bzero(mp->mnt_stat.f_mntfromname + size, MNAMELEN - size);
 504         bzero(mp->mnt_stat.f_mntonname, sizeof(mp->mnt_stat.f_mntonname));
 505         copyinstr(path, mp->mnt_stat.f_mntonname,
 506                   sizeof(mp->mnt_stat.f_mntonname) - 1,
 507                   &size);
 508
 509         /*
 510          * Initial statfs to prime mnt_stat.
 511          */
 512         hammer2_vfs_statfs(mp, &mp->mnt_stat, cred);
 513
 514         return 0;
 515 }
 516
 517 static
 518 int
 519 hammer2_remount(struct mount *mp, char *path, struct vnode *devvp,
 520                 struct ucred *cred)
 521 {
 522         return (0);
 523 }
 524
 525 static
 526 int
 527 hammer2_vfs_unmount(struct mount *mp, int mntflags)
 528 {
 529         hammer2_pfsmount_t *pmp;
 530         hammer2_mount_t *hmp;
 531         int flags;
 532         int error = 0;
 533         int ronly = ((mp->mnt_flag & MNT_RDONLY) != 0);
 534         struct vnode *devvp;
 535
 536         pmp = MPTOPMP(mp);
 537         hmp = pmp->hmp;
 538         flags = 0;
 539
 540         if (mntflags & MNT_FORCE)
 541                 flags |= FORCECLOSE;
 542
 543         hammer2_mount_exlock(hmp);
 544
 545         /*
 546          * If mount initialization proceeded far enough we must flush
 547          * its vnodes.
 548          */
 549         if (pmp->iroot)
 550                 error = vflush(mp, 0, flags);
 551
 552         if (error)
 553                 return error;
 554
 555         lockmgr(&hammer2_mntlk, LK_EXCLUSIVE);
 556         --hmp->pmp_count;
 557         kprintf("hammer2_unmount hmp=%p pmpcnt=%d\n", hmp, hmp->pmp_count);
 558
 559         /*
 560          * Flush any left over chains.  The voldata lock is only used
 561          * to synchronize against HAMMER2_CHAIN_MODIFIED_AUX.
 562          */
 563         hammer2_voldata_lock(hmp);
 564         if (hmp->vchain.flags & (HAMMER2_CHAIN_MODIFIED |
 565                                  HAMMER2_CHAIN_MODIFIED_AUX |
 566                                  HAMMER2_CHAIN_SUBMODIFIED)) {
 567                 hammer2_voldata_unlock(hmp);
 568                 hammer2_vfs_sync(mp, MNT_WAIT);
 569         } else {
 570                 hammer2_voldata_unlock(hmp);
 571         }
 572         if (hmp->pmp_count == 0) {
 573                 if (hmp->vchain.flags & (HAMMER2_CHAIN_MODIFIED |
 574                                          HAMMER2_CHAIN_MODIFIED_AUX |
 575                                          HAMMER2_CHAIN_SUBMODIFIED)) {
 576                         kprintf("hammer2_unmount: chains left over after "
 577                                 "final sync\n");
 578                         if (hammer2_debug & 0x0010)
 579                                 Debugger("entered debugger");
 580                 }
 581         }
 582
 583         /*
 584          * Cleanup the root and super-root chain elements (which should be
 585          * clean).
 586          */
 587         pmp->iroot = NULL;
 588         if (pmp->rchain) {
 589                 atomic_clear_int(&pmp->rchain->flags, HAMMER2_CHAIN_MOUNTED);
 590                 KKASSERT(pmp->rchain->refs == 1);
 591                 hammer2_chain_drop(hmp, pmp->rchain);
 592                 pmp->rchain = NULL;
 593         }
 594         ccms_domain_uninit(&pmp->ccms_dom);
 595
 596         /*
 597          * Ask the cluster controller to go away
 598          */
 599         atomic_set_int(&pmp->msg_ctl, HAMMER2_CLUSTERCTL_KILL);
 600         while (pmp->msgrd_td || pmp->msgwr_td) {
 601                 wakeup(&pmp->msg_ctl);
 602                 tsleep(pmp, 0, "clstrkl", hz);
 603         }
 604
 605         /*
 606          * Drop communications descriptor
 607          */
 608         if (pmp->msg_fp) {
 609                 fdrop(pmp->msg_fp);
 610                 pmp->msg_fp = NULL;
 611         }
 612
 613         /*
 614          * If no PFS's left drop the master hammer2_mount for the device.
 615          */
 616         if (hmp->pmp_count == 0) {
 617                 if (hmp->schain) {
 618                         KKASSERT(hmp->schain->refs == 1);
 619                         hammer2_chain_drop(hmp, hmp->schain);
 620                         hmp->schain = NULL;
 621                 }
 622
 623                 /*
 624                  * Finish up with the device vnode
 625                  */
 626                 if ((devvp = hmp->devvp) != NULL) {
 627                         vinvalbuf(devvp, (ronly ? 0 : V_SAVE), 0, 0);
 628                         hmp->devvp = NULL;
 629                         VOP_CLOSE(devvp, (ronly ? FREAD : FREAD|FWRITE));
 630                         vrele(devvp);
 631                         devvp = NULL;
 632                 }
 633         }
 634         hammer2_mount_unlock(hmp);
 635
 636         pmp->mp = NULL;
 637         pmp->hmp = NULL;
 638         mp->mnt_data = NULL;
 639
 640         kmalloc_destroy(&pmp->mmsg);
 641
 642         kfree(pmp, M_HAMMER2);
 643         if (hmp->pmp_count == 0) {
 644                 TAILQ_REMOVE(&hammer2_mntlist, hmp, mntentry);
 645                 kmalloc_destroy(&hmp->minode);
 646                 kmalloc_destroy(&hmp->mchain);
 647                 kfree(hmp, M_HAMMER2);
 648         }
 649         lockmgr(&hammer2_mntlk, LK_RELEASE);
 650         return (error);
 651 }
 652
 653 static
 654 int
 655 hammer2_vfs_vget(struct mount *mp, struct vnode *dvp,
 656              ino_t ino, struct vnode **vpp)
 657 {
 658         kprintf("hammer2_vget\n");
 659         return (EOPNOTSUPP);
 660 }
 661
 662 static
 663 int
 664 hammer2_vfs_root(struct mount *mp, struct vnode **vpp)
 665 {
 666         hammer2_pfsmount_t *pmp;
 667         hammer2_mount_t *hmp;
 668         int error;
 669         struct vnode *vp;
 670
 671         pmp = MPTOPMP(mp);
 672         hmp = pmp->hmp;
 673         hammer2_mount_exlock(hmp);
 674         if (pmp->iroot == NULL) {
 675                 *vpp = NULL;
 676                 error = EINVAL;
 677         } else {
 678                 hammer2_chain_lock(hmp, &pmp->iroot->chain,
 679                                    HAMMER2_RESOLVE_ALWAYS |
 680                                    HAMMER2_RESOLVE_SHARED);
 681                 vp = hammer2_igetv(pmp->iroot, &error);
 682                 hammer2_chain_unlock(hmp, &pmp->iroot->chain);
 683                 *vpp = vp;
 684                 if (vp == NULL)
 685                         kprintf("vnodefail\n");
 686         }
 687         hammer2_mount_unlock(hmp);
 688
 689         return (error);
 690 }
 691
 692 /*
 693  * Filesystem status
 694  *
 695  * XXX incorporate pmp->iroot->ip_data.inode_quota and data_quota
 696  */
 697 static
 698 int
 699 hammer2_vfs_statfs(struct mount *mp, struct statfs *sbp, struct ucred *cred)
 700 {
 701         hammer2_pfsmount_t *pmp;
 702         hammer2_mount_t *hmp;
 703
 704         pmp = MPTOPMP(mp);
 705         hmp = MPTOHMP(mp);
 706
 707         mp->mnt_stat.f_files = pmp->iroot->ip_data.inode_count +
 708                                pmp->iroot->delta_icount;
 709         mp->mnt_stat.f_ffree = 0;
 710         mp->mnt_stat.f_blocks = hmp->voldata.allocator_size / HAMMER2_PBUFSIZE;
 711         mp->mnt_stat.f_bfree = (hmp->voldata.allocator_size -
 712                                 hmp->voldata.allocator_beg) / HAMMER2_PBUFSIZE;
 713         mp->mnt_stat.f_bavail = mp->mnt_stat.f_bfree;
 714
 715         *sbp = mp->mnt_stat;
 716         return (0);
 717 }
 718
 719 static
 720 int
 721 hammer2_vfs_statvfs(struct mount *mp, struct statvfs *sbp, struct ucred *cred)
 722 {
 723         hammer2_pfsmount_t *pmp;
 724         hammer2_mount_t *hmp;
 725
 726         pmp = MPTOPMP(mp);
 727         hmp = MPTOHMP(mp);
 728
 729         mp->mnt_vstat.f_bsize = HAMMER2_PBUFSIZE;
 730         mp->mnt_vstat.f_files = pmp->iroot->ip_data.inode_count +
 731                                 pmp->iroot->delta_icount;
 732         mp->mnt_vstat.f_ffree = 0;
 733         mp->mnt_vstat.f_blocks = hmp->voldata.allocator_size / HAMMER2_PBUFSIZE;
 734         mp->mnt_vstat.f_bfree = (hmp->voldata.allocator_size -
 735                                  hmp->voldata.allocator_beg) / HAMMER2_PBUFSIZE;
 736         mp->mnt_vstat.f_bavail = mp->mnt_vstat.f_bfree;
 737
 738         *sbp = mp->mnt_vstat;
 739         return (0);
 740 }
 741
 742 /*
 743  * Sync the entire filesystem; this is called from the filesystem syncer
 744  * process periodically and whenever a user calls sync(1) on the hammer
 745  * mountpoint.
 746  *
 747  * Currently is actually called from the syncer! \o/
 748  *
 749  * This task will have to snapshot the state of the dirty inode chain.
 750  * From that, it will have to make sure all of the inodes on the dirty
 751  * chain have IO initiated. We make sure that io is initiated for the root
 752  * block.
 753  *
 754  * If waitfor is set, we wait for media to acknowledge the new rootblock.
 755  *
 756  * THINKS: side A vs side B, to have sync not stall all I/O?
 757  */
 758 static
 759 int
 760 hammer2_vfs_sync(struct mount *mp, int waitfor)
 761 {
 762         struct hammer2_sync_info info;
 763         hammer2_mount_t *hmp;
 764         int flags;
 765         int error;
 766         int haswork;
 767
 768         hmp = MPTOHMP(mp);
 769
 770         flags = VMSC_GETVP;
 771         if (waitfor & MNT_LAZY)
 772                 flags |= VMSC_ONEPASS;
 773
 774         info.error = 0;
 775         info.waitfor = MNT_NOWAIT;
 776         vmntvnodescan(mp, flags | VMSC_NOWAIT,
 777                       hammer2_sync_scan1,
 778                       hammer2_sync_scan2, &info);
 779         if (info.error == 0 && (waitfor & MNT_WAIT)) {
 780                 info.waitfor = waitfor;
 781                     vmntvnodescan(mp, flags,
 782                                   hammer2_sync_scan1,
 783                                   hammer2_sync_scan2, &info);
 784
 785         }
 786 #if 0
 787         if (waitfor == MNT_WAIT) {
 788                 /* XXX */
 789         } else {
 790                 /* XXX */
 791         }
 792 #endif
 793         hammer2_chain_lock(hmp, &hmp->vchain, HAMMER2_RESOLVE_ALWAYS);
 794         if (hmp->vchain.flags & (HAMMER2_CHAIN_MODIFIED |
 795                                  HAMMER2_CHAIN_MODIFIED_AUX |
 796                                  HAMMER2_CHAIN_SUBMODIFIED)) {
 797                 hammer2_chain_flush(hmp, &hmp->vchain, 0);
 798                 haswork = 1;
 799         } else {
 800                 haswork = 0;
 801         }
 802         hammer2_chain_unlock(hmp, &hmp->vchain);
 803
 804         error = 0;
 805
 806         if ((waitfor & MNT_LAZY) == 0) {
 807                 waitfor = MNT_NOWAIT;
 808                 vn_lock(hmp->devvp, LK_EXCLUSIVE | LK_RETRY);
 809                 error = VOP_FSYNC(hmp->devvp, waitfor, 0);
 810                 vn_unlock(hmp->devvp);
 811         }
 812
 813         if (error == 0 && haswork) {
 814                 struct buf *bp;
 815
 816                 /*
 817                  * Synchronize the disk before flushing the volume
 818                  * header.
 819                  */
 820                 bp = getpbuf(NULL);
 821                 bp->b_bio1.bio_offset = 0;
 822                 bp->b_bufsize = 0;
 823                 bp->b_bcount = 0;
 824                 bp->b_cmd = BUF_CMD_FLUSH;
 825                 bp->b_bio1.bio_done = biodone_sync;
 826                 bp->b_bio1.bio_flags |= BIO_SYNC;
 827                 vn_strategy(hmp->devvp, &bp->b_bio1);
 828                 biowait(&bp->b_bio1, "h2vol");
 829                 relpbuf(bp, NULL);
 830
 831                 /*
 832                  * Then we can safely flush the volume header.  Volume
 833                  * data is locked separately to prevent ioctl functions
 834                  * from deadlocking due to a configuration issue.
 835                  */
 836                 bp = getblk(hmp->devvp, 0, HAMMER2_PBUFSIZE, 0, 0);
 837                 hammer2_voldata_lock(hmp);
 838                 bcopy(&hmp->voldata, bp->b_data, HAMMER2_PBUFSIZE);
 839                 hammer2_voldata_unlock(hmp);
 840                 bawrite(bp);
 841         }
 842         return (error);
 843 }
 844
 845 /*
 846  * Sync passes.
 847  *
 848  * NOTE: We don't test SUBMODIFIED or MOVED here because the fsync code
 849  *       won't flush on those flags.  The syncer code above will do a
 850  *       general meta-data flush globally that will catch these flags.
 851  */
 852 static int
 853 hammer2_sync_scan1(struct mount *mp, struct vnode *vp, void *data)
 854 {
 855         hammer2_inode_t *ip;
 856
 857         ip = VTOI(vp);
 858         if (vp->v_type == VNON || ip == NULL ||
 859             ((ip->chain.flags & (HAMMER2_CHAIN_MODIFIED |
 860                                  HAMMER2_CHAIN_DIRTYEMBED)) == 0 &&
 861              RB_EMPTY(&vp->v_rbdirty_tree))) {
 862                 return(-1);
 863         }
 864         return(0);
 865 }
 866
 867 static int
 868 hammer2_sync_scan2(struct mount *mp, struct vnode *vp, void *data)
 869 {
 870         struct hammer2_sync_info *info = data;
 871         hammer2_inode_t *ip;
 872         int error;
 873
 874         ip = VTOI(vp);
 875         if (vp->v_type == VNON || vp->v_type == VBAD ||
 876             ((ip->chain.flags & (HAMMER2_CHAIN_MODIFIED |
 877                                  HAMMER2_CHAIN_DIRTYEMBED)) == 0 &&
 878             RB_EMPTY(&vp->v_rbdirty_tree))) {
 879                 return(0);
 880         }
 881         error = VOP_FSYNC(vp, MNT_NOWAIT, 0);
 882         if (error)
 883                 info->error = error;
 884         return(0);
 885 }
 886
 887 static
 888 int
 889 hammer2_vfs_vptofh(struct vnode *vp, struct fid *fhp)
 890 {
 891         return (0);
 892 }
 893
 894 static
 895 int
 896 hammer2_vfs_fhtovp(struct mount *mp, struct vnode *rootvp,
 897                struct fid *fhp, struct vnode **vpp)
 898 {
 899         return (0);
 900 }
 901
 902 static
 903 int
 904 hammer2_vfs_checkexp(struct mount *mp, struct sockaddr *nam,
 905                  int *exflagsp, struct ucred **credanonp)
 906 {
 907         return (0);
 908 }
 909
 910 /*
 911  * Support code for hammer2_mount().  Read, verify, and install the volume
 912  * header into the HMP
 913  *
 914  * XXX read four volhdrs and use the one with the highest TID whos CRC
 915  *     matches.
 916  *
 917  * XXX check iCRCs.
 918  *
 919  * XXX For filesystems w/ less than 4 volhdrs, make sure to not write to
 920  *     nonexistant locations.
 921  *
 922  * XXX Record selected volhdr and ring updates to each of 4 volhdrs
 923  */
 924 static
 925 int
 926 hammer2_install_volume_header(hammer2_mount_t *hmp)
 927 {
 928         hammer2_volume_data_t *vd;
 929         struct buf *bp;
 930         hammer2_crc32_t crc0, crc, bcrc0, bcrc;
 931         int error_reported;
 932         int error;
 933         int valid;
 934         int i;
 935
 936         error_reported = 0;
 937         error = 0;
 938         valid = 0;
 939         bp = NULL;
 940
 941         /*
 942          * There are up to 4 copies of the volume header (syncs iterate
 943          * between them so there is no single master).  We don't trust the
 944          * volu_size field so we don't know precisely how large the filesystem
 945          * is, so depend on the OS to return an error if we go beyond the
 946          * block device's EOF.
 947          */
 948         for (i = 0; i < HAMMER2_NUM_VOLHDRS; i++) {
 949                 error = bread(hmp->devvp, i * HAMMER2_ZONE_BYTES64,
 950                               HAMMER2_VOLUME_BYTES, &bp);
 951                 if (error) {
 952                         brelse(bp);
 953                         bp = NULL;
 954                         continue;
 955                 }
 956
 957                 vd = (struct hammer2_volume_data *) bp->b_data;
 958                 if ((vd->magic != HAMMER2_VOLUME_ID_HBO) &&
 959                     (vd->magic != HAMMER2_VOLUME_ID_ABO)) {
 960                         brelse(bp);
 961                         bp = NULL;
 962                         continue;
 963                 }
 964
 965                 if (vd->magic == HAMMER2_VOLUME_ID_ABO) {
 966                         /* XXX: Reversed-endianness filesystem */
 967                         kprintf("hammer2: reverse-endian filesystem detected");
 968                         brelse(bp);
 969                         bp = NULL;
 970                         continue;
 971                 }
 972
 973                 crc = vd->icrc_sects[HAMMER2_VOL_ICRC_SECT0];
 974                 crc0 = hammer2_icrc32(bp->b_data + HAMMER2_VOLUME_ICRC0_OFF,
 975                                       HAMMER2_VOLUME_ICRC0_SIZE);
 976                 bcrc = vd->icrc_sects[HAMMER2_VOL_ICRC_SECT1];
 977                 bcrc0 = hammer2_icrc32(bp->b_data + HAMMER2_VOLUME_ICRC1_OFF,
 978                                        HAMMER2_VOLUME_ICRC1_SIZE);
 979                 if ((crc0 != crc) || (bcrc0 != bcrc)) {
 980                         kprintf("hammer2 volume header crc "
 981                                 "mismatch copy #%d\t%08x %08x",
 982                                 i, crc0, crc);
 983                         error_reported = 1;
 984                         brelse(bp);
 985                         bp = NULL;
 986                         continue;
 987                 }
 988                 if (valid == 0 || hmp->voldata.mirror_tid < vd->mirror_tid) {
 989                         valid = 1;
 990                         hmp->voldata = *vd;
 991                 }
 992                 brelse(bp);
 993                 bp = NULL;
 994         }
 995         if (valid) {
 996                 error = 0;
 997                 if (error_reported)
 998                         kprintf("hammer2: a valid volume header was found\n");
 999         } else {
1000                 error = EINVAL;
1001                 kprintf("hammer2: no valid volume headers found!\n");
1002         }
1003         return (error);
1004 }
1005
1006 /*
1007  * Cluster controller thread.  Perform messaging functions.  We have one
1008  * thread for the reader and one for the writer.  The writer handles
1009  * shutdown requests (which should break the reader thread).
1010  */
1011 static
1012 void
1013 hammer2_cluster_thread_rd(void *arg)
1014 {
1015         hammer2_pfsmount_t *pmp = arg;
1016         hammer2_msg_hdr_t hdr;
1017         hammer2_msg_t *msg;
1018         hammer2_state_t *state;
1019         size_t hbytes;
1020         int error = 0;
1021
1022         while ((pmp->msg_ctl & HAMMER2_CLUSTERCTL_KILL) == 0) {
1023                 /*
1024                  * Retrieve the message from the pipe or socket.
1025                  */
1026                 error = fp_read(pmp->msg_fp, &hdr, sizeof(hdr),
1027                                 NULL, 1, UIO_SYSSPACE);
1028                 if (error)
1029                         break;
1030                 if (hdr.magic != HAMMER2_MSGHDR_MAGIC) {
1031                         kprintf("hammer2: msgrd: bad magic: %04x\n",
1032                                 hdr.magic);
1033                         error = EINVAL;
1034                         break;
1035                 }
1036                 hbytes = (hdr.cmd & HAMMER2_MSGF_SIZE) * HAMMER2_MSG_ALIGN;
1037                 if (hbytes < sizeof(hdr) || hbytes > HAMMER2_MSGAUX_MAX) {
1038                         kprintf("hammer2: msgrd: bad header size %zd\n",
1039                                 hbytes);
1040                         error = EINVAL;
1041                         break;
1042                 }
1043                 msg = kmalloc(offsetof(struct hammer2_msg, any) + hbytes,
1044                               pmp->mmsg, M_WAITOK | M_ZERO);
1045                 msg->any.head = hdr;
1046                 msg->hdr_size = hbytes;
1047                 if (hbytes > sizeof(hdr)) {
1048                         error = fp_read(pmp->msg_fp, &msg->any.head + 1,
1049                                         hbytes - sizeof(hdr),
1050                                         NULL, 1, UIO_SYSSPACE);
1051                         if (error) {
1052                                 kprintf("hammer2: short msg received\n");
1053                                 error = EINVAL;
1054                                 break;
1055                         }
1056                 }
1057                 msg->aux_size = hdr.aux_bytes * HAMMER2_MSG_ALIGN;
1058                 if (msg->aux_size > HAMMER2_MSGAUX_MAX) {
1059                         kprintf("hammer2: illegal msg payload size %zd\n",
1060                                 msg->aux_size);
1061                         error = EINVAL;
1062                         break;
1063                 }
1064                 if (msg->aux_size) {
1065                         msg->aux_data = kmalloc(msg->aux_size, pmp->mmsg,
1066                                                 M_WAITOK | M_ZERO);
1067                         error = fp_read(pmp->msg_fp, msg->aux_data,
1068                                         msg->aux_size,
1069                                         NULL, 1, UIO_SYSSPACE);
1070                         if (error) {
1071                                 kprintf("hammer2: short msg "
1072                                         "payload received\n");
1073                                 break;
1074                         }
1075                 }
1076
1077                 /*
1078                  * State machine tracking, state assignment for msg,
1079                  * returns error and discard status.  Errors are fatal
1080                  * to the connection except for EALREADY which forces
1081                  * a discard without execution.
1082                  */
1083                 error = hammer2_state_msgrx(pmp, msg);
1084                 if (error) {
1085                         hammer2_msg_free(pmp, msg);
1086                         if (error == EALREADY)
1087                                 error = 0;
1088                 } else if (msg->state) {
1089                         error = msg->state->func(pmp, msg);
1090                         hammer2_state_cleanuprx(pmp, msg);
1091                 } else {
1092                         error = hammer2_msg_adhoc_input(pmp, msg);
1093                         hammer2_state_cleanuprx(pmp, msg);
1094                 }
1095                 msg = NULL;
1096         }
1097
1098         if (error)
1099                 kprintf("hammer2: msg read failed error %d\n", error);
1100
1101         lockmgr(&pmp->msglk, LK_EXCLUSIVE);
1102         if (msg) {
1103                 if (msg->state && msg->state->msg == msg)
1104                         msg->state->msg = NULL;
1105                 hammer2_msg_free(pmp, msg);
1106         }
1107
1108         if ((state = pmp->freerd_state) != NULL) {
1109                 pmp->freerd_state = NULL;
1110                 hammer2_state_free(state);
1111         }
1112
1113         while ((state = RB_ROOT(&pmp->staterd_tree)) != NULL) {
1114                 RB_REMOVE(hammer2_state_tree, &pmp->staterd_tree, state);
1115                 hammer2_state_free(state);
1116         }
1117         lockmgr(&pmp->msglk, LK_RELEASE);
1118
1119         fp_shutdown(pmp->msg_fp, SHUT_RDWR);
1120         pmp->msgrd_td = NULL;
1121         /* pmp can be ripped out from under us at this point */
1122         wakeup(pmp);
1123         lwkt_exit();
1124 }
1125
1126 static
1127 void
1128 hammer2_cluster_thread_wr(void *arg)
1129 {
1130         hammer2_pfsmount_t *pmp = arg;
1131         hammer2_msg_t *msg = NULL;
1132         hammer2_state_t *state;
1133         ssize_t res;
1134         int error = 0;
1135
1136         /*
1137          * Initiate a SPAN transaction registering our PFS with the other
1138          * end using {source}=1.  The transaction is left open.
1139          */
1140         msg = hammer2_msg_alloc(pmp, 1, 0,
1141                                 HAMMER2_LNK_SPAN | HAMMER2_MSGF_CREATE);
1142         hammer2_msg_write(pmp, msg, hammer2_msg_span_reply);
1143
1144         /*
1145          * Transmit loop
1146          */
1147         msg = NULL;
1148         lockmgr(&pmp->msglk, LK_EXCLUSIVE);
1149
1150         while ((pmp->msg_ctl & HAMMER2_CLUSTERCTL_KILL) == 0 && error == 0) {
1151                 lksleep(&pmp->msg_ctl, &pmp->msglk, 0, "msgwr", hz);
1152                 while ((msg = TAILQ_FIRST(&pmp->msgq)) != NULL) {
1153                         /*
1154                          * Remove msg from the transmit queue and do
1155                          * persist and half-closed state handling.
1156                          */
1157                         TAILQ_REMOVE(&pmp->msgq, msg, qentry);
1158                         lockmgr(&pmp->msglk, LK_RELEASE);
1159
1160                         error = hammer2_state_msgtx(pmp, msg);
1161                         if (error == EALREADY) {
1162                                 error = 0;
1163                                 hammer2_msg_free(pmp, msg);
1164                                 lockmgr(&pmp->msglk, LK_EXCLUSIVE);
1165                                 continue;
1166                         }
1167                         if (error)
1168                                 break;
1169
1170                         /*
1171                          * Dump the message to the pipe or socket.
1172                          */
1173                         error = fp_write(pmp->msg_fp, &msg->any, msg->hdr_size,
1174                                          &res, UIO_SYSSPACE);
1175                         if (error || res != msg->hdr_size) {
1176                                 if (error == 0)
1177                                         error = EINVAL;
1178                                 lockmgr(&pmp->msglk, LK_EXCLUSIVE);
1179                                 break;
1180                         }
1181                         if (msg->aux_size) {
1182                                 error = fp_write(pmp->msg_fp,
1183                                                  msg->aux_data, msg->aux_size,
1184                                                  &res, UIO_SYSSPACE);
1185                                 if (error || res != msg->aux_size) {
1186                                         if (error == 0)
1187                                                 error = EINVAL;
1188                                         lockmgr(&pmp->msglk, LK_EXCLUSIVE);
1189                                         break;
1190                                 }
1191                         }
1192                         hammer2_state_cleanuptx(pmp, msg);
1193                         lockmgr(&pmp->msglk, LK_EXCLUSIVE);
1194                 }
1195         }
1196
1197         /*
1198          * Cleanup messages pending transmission and release msgq lock.
1199          */
1200         if (error)
1201                 kprintf("hammer2: msg write failed error %d\n", error);
1202
1203         if (msg) {
1204                 if (msg->state && msg->state->msg == msg)
1205                         msg->state->msg = NULL;
1206                 hammer2_msg_free(pmp, msg);
1207         }
1208
1209         while ((msg = TAILQ_FIRST(&pmp->msgq)) != NULL) {
1210                 TAILQ_REMOVE(&pmp->msgq, msg, qentry);
1211                 if (msg->state && msg->state->msg == msg)
1212                         msg->state->msg = NULL;
1213                 hammer2_msg_free(pmp, msg);
1214         }
1215
1216         if ((state = pmp->freewr_state) != NULL) {
1217                 pmp->freewr_state = NULL;
1218                 hammer2_state_free(state);
1219         }
1220
1221         while ((state = RB_ROOT(&pmp->statewr_tree)) != NULL) {
1222                 RB_REMOVE(hammer2_state_tree, &pmp->statewr_tree, state);
1223                 hammer2_state_free(state);
1224         }
1225         lockmgr(&pmp->msglk, LK_RELEASE);
1226
1227         /*
1228          * Cleanup descriptor, be sure the read size is shutdown so the
1229          * (probably blocked) read operations returns an error.
1230          *
1231          * pmp can be ripped out from under us once msgwr_td is set to NULL.
1232          */
1233         fp_shutdown(pmp->msg_fp, SHUT_RDWR);
1234         pmp->msgwr_td = NULL;
1235         wakeup(pmp);
1236         lwkt_exit();
1237 }
1238
1239 static int
1240 hammer2_msg_span_reply(hammer2_pfsmount_t *pmp, hammer2_msg_t *msg)
1241 {
1242         kprintf("SPAN REPLY\n");
1243         return(0);
1244 }