Current Path : /sys/fs/nwfs/ |
FreeBSD hs32.drive.ne.jp 9.1-RELEASE FreeBSD 9.1-RELEASE #1: Wed Jan 14 12:18:08 JST 2015 root@hs32.drive.ne.jp:/sys/amd64/compile/hs32 amd64 |
Current File : //sys/fs/nwfs/nwfs_vfsops.c |
/*- * Copyright (c) 1999, 2000 Boris Popov * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD: release/9.1.0/sys/fs/nwfs/nwfs_vfsops.c 230725 2012-01-29 08:03:45Z mckusick $ */ #include <sys/param.h> #include <sys/systm.h> #include <sys/proc.h> #include <sys/kernel.h> #include <sys/sysctl.h> #include <sys/vnode.h> #include <sys/mount.h> #include <sys/stat.h> #include <sys/malloc.h> #include <sys/bio.h> #include <sys/buf.h> #include <netncp/ncp.h> #include <netncp/ncp_conn.h> #include <netncp/ncp_subr.h> #include <netncp/ncp_ncp.h> #include <netncp/ncp_nls.h> #include <fs/nwfs/nwfs.h> #include <fs/nwfs/nwfs_node.h> #include <fs/nwfs/nwfs_subr.h> int nwfs_debuglevel = 0; static int nwfs_version = NWFS_VERSION; SYSCTL_DECL(_vfs_nwfs); SYSCTL_NODE(_vfs, OID_AUTO, nwfs, CTLFLAG_RW, 0, "Netware filesystem"); SYSCTL_INT(_vfs_nwfs, OID_AUTO, version, CTLFLAG_RD, &nwfs_version, 0, ""); SYSCTL_INT(_vfs_nwfs, OID_AUTO, debuglevel, CTLFLAG_RW, &nwfs_debuglevel, 0, ""); MODULE_DEPEND(nwfs, ncp, 1, 1, 1); MODULE_DEPEND(nwfs, libmchain, 1, 1, 1); static vfs_cmount_t nwfs_cmount; static vfs_mount_t nwfs_mount; static vfs_quotactl_t nwfs_quotactl; static vfs_root_t nwfs_root; static vfs_statfs_t nwfs_statfs; static vfs_unmount_t nwfs_unmount; static vfs_init_t nwfs_init; static vfs_uninit_t nwfs_uninit; static struct vfsops nwfs_vfsops = { .vfs_init = nwfs_init, .vfs_mount = nwfs_mount, .vfs_cmount = nwfs_cmount, .vfs_quotactl = nwfs_quotactl, .vfs_root = nwfs_root, .vfs_statfs = nwfs_statfs, .vfs_sync = vfs_stdsync, .vfs_uninit = nwfs_uninit, .vfs_unmount = nwfs_unmount, }; VFS_SET(nwfs_vfsops, nwfs, VFCF_NETWORK); int nwfs_pbuf_freecnt = -1; /* start out unlimited */ static int nwfsid = 1; static int nwfs_initnls(struct nwmount *nmp) { char *pc, *pe; int error = 0; #define COPY_TABLE(t,d) { \ if (t) { \ error = copyin((t), pc, 256); \ if (error) break; \ } else \ bcopy(d, pc, 256); \ (t) = pc; pc += 256; \ } nmp->m.nls.opt |= NWHP_NLS | NWHP_DOS; if ((nmp->m.flags & NWFS_MOUNT_HAVE_NLS) == 0) { nmp->m.nls.to_lower = ncp_defnls.to_lower; nmp->m.nls.to_upper = ncp_defnls.to_upper; nmp->m.nls.n2u = ncp_defnls.n2u; nmp->m.nls.u2n = ncp_defnls.u2n; return 0; } pe = malloc(256 * 4, M_NWFSDATA, M_WAITOK); pc = pe; do { COPY_TABLE(nmp->m.nls.to_lower, ncp_defnls.to_lower); COPY_TABLE(nmp->m.nls.to_upper, ncp_defnls.to_upper); COPY_TABLE(nmp->m.nls.n2u, ncp_defnls.n2u); COPY_TABLE(nmp->m.nls.u2n, ncp_defnls.u2n); } while(0); if (error) { free(pe, M_NWFSDATA); return error; } return 0; } static int nwfs_cmount(struct mntarg *ma, void *data, uint64_t flags) { struct nwfs_args args; /* will hold data from mount request */ int error; error = copyin(data, &args, sizeof(struct nwfs_args)); if (error) return (error); /* * XXX: cheap cop-out here, args contains a structure I don't * XXX: know how we should handle, and I don't see any immediate * XXX: prospect of avoiding a mount_nwfs(8) binary anyway. */ ma = mount_arg(ma, "nwfs_args", &args, sizeof args); error = kernel_mount(ma, flags); return (error); } /* * mp - path - addr in user space of mount point (ie /usr or whatever) * data - addr in user space of mount params */ static int nwfs_mount(struct mount *mp) { struct nwfs_args args; /* will hold data from mount request */ int error; struct nwmount *nmp = NULL; struct ncp_conn *conn = NULL; struct ncp_handle *handle = NULL; struct vnode *vp; struct thread *td; char *pc,*pe; td = curthread; if (mp->mnt_flag & MNT_ROOTFS) return (EOPNOTSUPP); if (mp->mnt_flag & MNT_UPDATE) { nwfs_printf("MNT_UPDATE not implemented"); return (EOPNOTSUPP); } error = vfs_copyopt(mp->mnt_optnew, "nwfs_args", &args, sizeof args); if (error) return (error); if (args.version != NWFS_VERSION) { nwfs_printf("mount version mismatch: kernel=%d, mount=%d\n",NWFS_VERSION,args.version); return (1); } error = ncp_conn_getbyref(args.connRef, td , td->td_ucred,NCPM_EXECUTE,&conn); if (error) { nwfs_printf("invalid connection refernce %d\n",args.connRef); return (error); } error = ncp_conn_gethandle(conn, NULL, &handle); if (error) { nwfs_printf("can't get connection handle\n"); return (error); } ncp_conn_unlock(conn, td); /* we keep the ref */ mp->mnt_stat.f_iosize = conn->buffer_size; /* We must malloc our own mount info */ nmp = malloc(sizeof(struct nwmount), M_NWFSDATA, M_WAITOK | M_ZERO); if (nmp == NULL) { nwfs_printf("could not alloc nwmount\n"); error = ENOMEM; goto bad; } mp->mnt_data = nmp; nmp->connh = handle; nmp->n_root = NULL; nmp->n_id = nwfsid++; nmp->m = args; nmp->m.file_mode = (nmp->m.file_mode & (S_IRWXU|S_IRWXG|S_IRWXO)) | S_IFREG; nmp->m.dir_mode = (nmp->m.dir_mode & (S_IRWXU|S_IRWXG|S_IRWXO)) | S_IFDIR; if ((error = nwfs_initnls(nmp)) != 0) goto bad; pc = mp->mnt_stat.f_mntfromname; pe = pc+sizeof(mp->mnt_stat.f_mntfromname); bzero(pc, MNAMELEN); *(pc++) = '/'; pc = index(strncpy(pc, conn->li.server, pe-pc-2),0); if (pc < pe-1) { *(pc++) = ':'; pc=index(strncpy(pc, conn->li.user, pe-pc-2),0); if (pc < pe-1) { *(pc++) = '/'; strncpy(pc, nmp->m.mounted_vol, pe-pc-2); } } /* protect against invalid mount points */ nmp->m.mount_point[sizeof(nmp->m.mount_point)-1] = '\0'; vfs_getnewfsid(mp); error = nwfs_root(mp, LK_EXCLUSIVE, &vp); if (error) goto bad; /* * Lose the lock but keep the ref. */ VOP_UNLOCK(vp, 0); NCPVODEBUG("rootvp.vrefcnt=%d\n",vrefcnt(vp)); return error; bad: if (nmp) free(nmp, M_NWFSDATA); if (handle) ncp_conn_puthandle(handle, NULL, 0); return error; } /* Unmount the filesystem described by mp. */ static int nwfs_unmount(struct mount *mp, int mntflags) { struct thread *td; struct nwmount *nmp = VFSTONWFS(mp); struct ncp_conn *conn; int error, flags; NCPVODEBUG("nwfs_unmount: flags=%04x\n",mntflags); td = curthread; flags = 0; if (mntflags & MNT_FORCE) flags |= FORCECLOSE; /* There is 1 extra root vnode reference from nwfs_mount(). */ error = vflush(mp, 1, flags, td); if (error) return (error); conn = NWFSTOCONN(nmp); ncp_conn_puthandle(nmp->connh,NULL,0); if (ncp_conn_lock(conn, td, td->td_ucred,NCPM_WRITE | NCPM_EXECUTE) == 0) { if(ncp_conn_free(conn)) ncp_conn_unlock(conn, td); } mp->mnt_data = NULL; if (nmp->m.flags & NWFS_MOUNT_HAVE_NLS) free(nmp->m.nls.to_lower, M_NWFSDATA); free(nmp, M_NWFSDATA); MNT_ILOCK(mp); mp->mnt_flag &= ~MNT_LOCAL; MNT_IUNLOCK(mp); return (error); } /* Return locked vnode to root of a filesystem */ static int nwfs_root(struct mount *mp, int flags, struct vnode **vpp) { struct vnode *vp; struct nwmount *nmp; struct nwnode *np; struct ncp_conn *conn; struct nw_entry_info fattr; struct thread *td; struct ucred *cred; int error, nsf, opt; u_char vol; td = curthread; cred = td->td_ucred; nmp = VFSTONWFS(mp); conn = NWFSTOCONN(nmp); if (nmp->n_root) { *vpp = NWTOV(nmp->n_root); while (vget(*vpp, LK_EXCLUSIVE, curthread) != 0) ; return 0; } error = ncp_lookup_volume(conn, nmp->m.mounted_vol, &vol, &nmp->n_rootent.f_id, td, cred); if (error) return ENOENT; nmp->n_volume = vol; error = ncp_get_namespaces(conn, vol, &nsf, td, cred); if (error) return ENOENT; if (nsf & NW_NSB_OS2) { NCPVODEBUG("volume %s has os2 namespace\n",nmp->m.mounted_vol); if ((nmp->m.flags & NWFS_MOUNT_NO_OS2) == 0) { nmp->name_space = NW_NS_OS2; nmp->m.nls.opt &= ~NWHP_DOS; } } opt = nmp->m.nls.opt; nsf = opt & (NWHP_UPPER | NWHP_LOWER); if (opt & NWHP_DOS) { if (nsf == (NWHP_UPPER | NWHP_LOWER)) { nmp->m.nls.opt &= ~(NWHP_LOWER | NWHP_UPPER); } else if (nsf == 0) { nmp->m.nls.opt |= NWHP_LOWER; } } else { if (nsf == (NWHP_UPPER | NWHP_LOWER)) { nmp->m.nls.opt &= ~(NWHP_LOWER | NWHP_UPPER); } } if (nmp->m.root_path[0]) { nmp->m.root_path[0]--; error = ncp_obtain_info(nmp, nmp->n_rootent.f_id, -nmp->m.root_path[0], nmp->m.root_path, &fattr, td, cred); if (error) { NCPFATAL("Invalid root path specified\n"); return ENOENT; } nmp->n_rootent.f_parent = fattr.dirEntNum; nmp->m.root_path[0]++; error = ncp_obtain_info(nmp, nmp->n_rootent.f_id, -nmp->m.root_path[0], nmp->m.root_path, &fattr, td, cred); if (error) { NCPFATAL("Invalid root path specified\n"); return ENOENT; } nmp->n_rootent.f_id = fattr.dirEntNum; } else { error = ncp_obtain_info(nmp, nmp->n_rootent.f_id, 0, NULL, &fattr, td, cred); if (error) { NCPFATAL("Can't obtain volume info\n"); return ENOENT; } fattr.nameLen = strlen(strcpy(fattr.entryName, "#.ROOT")); nmp->n_rootent.f_parent = nmp->n_rootent.f_id; } error = nwfs_nget(mp, nmp->n_rootent, &fattr, NULL, &vp); if (error) return (error); vp->v_vflag |= VV_ROOT; np = VTONW(vp); if (nmp->m.root_path[0] == 0) np->n_flag |= NVOLUME; nmp->n_root = np; /* error = VOP_GETATTR(vp, &vattr, cred, td); if (error) { vput(vp); NCPFATAL("Can't get root directory entry\n"); return error; }*/ *vpp = vp; return (0); } /* * Do operations associated with quotas, not supported */ /* ARGSUSED */ static int nwfs_quotactl(mp, cmd, uid, arg) struct mount *mp; int cmd; uid_t uid; void *arg; { NCPVODEBUG("return EOPNOTSUPP\n"); return (EOPNOTSUPP); } /*ARGSUSED*/ int nwfs_init(struct vfsconf *vfsp) { nwfs_hash_init(); nwfs_pbuf_freecnt = nswbuf / 2 + 1; NCPVODEBUG("always happy to load!\n"); return (0); } /*ARGSUSED*/ int nwfs_uninit(struct vfsconf *vfsp) { nwfs_hash_free(); NCPVODEBUG("unloaded\n"); return (0); } /* * nwfs_statfs call */ int nwfs_statfs(mp, sbp) struct mount *mp; struct statfs *sbp; { struct nwmount *nmp = VFSTONWFS(mp); struct thread *td = curthread; int error = 0, secsize; struct nwnode *np = nmp->n_root; struct ncp_volume_info vi; if (np == NULL) return EINVAL; error = ncp_get_volume_info_with_number(NWFSTOCONN(nmp), nmp->n_volume, &vi, td, td->td_ucred); if (error) return error; secsize = 512; /* XXX how to get real value ??? */ /* fundamental filesystem block size */ sbp->f_bsize = vi.sectors_per_block*secsize; /* optimal transfer block size */ sbp->f_iosize = NWFSTOCONN(nmp)->buffer_size; /* total data blocks in filesystem */ sbp->f_blocks= vi.total_blocks; /* free blocks in fs */ sbp->f_bfree = vi.free_blocks + vi.purgeable_blocks; /* free blocks avail to non-superuser */ sbp->f_bavail= vi.free_blocks+vi.purgeable_blocks; /* total file nodes in filesystem */ sbp->f_files = vi.total_dir_entries; /* free file nodes in fs */ sbp->f_ffree = vi.available_dir_entries; sbp->f_flags = 0; /* copy of mount exported flags */ return 0; }