/* $OpenBSD: kvm_file2.c,v 1.18 2011/12/14 17:33:46 guenther Exp $ */ /* * Copyright (c) 2009 Todd C. Miller * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ /*- * Copyright (c) 1989, 1992, 1993 * The Regents of the University of California. 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. * 3. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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. */ /* * Extended file list interface for kvm. pstat, fstat and netstat are * users of this code, so we've factored it out into a separate module. * Thus, we keep this grunge out of the other kvm applications (i.e., * most other applications are interested only in open/close/read/nlist). */ #define __need_process #include #include #include #include #define _KERNEL #include #include #include #undef _KERNEL #include #include #include #include #include #include #include #include #include #include #include #include #include #define _KERNEL #include #include #undef _KERNEL #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef INET6 #include #include #endif #include #include #include #include #include #include #include "kvm_private.h" static struct kinfo_file2 *kvm_deadfile2_byfile(kvm_t *, int, int, size_t, int *); static struct kinfo_file2 *kvm_deadfile2_byid(kvm_t *, int, int, size_t, int *); static int fill_file2(kvm_t *, struct kinfo_file2 *, struct file *, struct vnode *, struct proc *, int); static int filestat(kvm_t *, struct kinfo_file2 *, struct vnode *); LIST_HEAD(proclist, proc); struct kinfo_file2 * kvm_getfile2(kvm_t *kd, int op, int arg, size_t esize, int *cnt) { int mib[6], rv; size_t size; if (kd->filebase != NULL) { free(kd->filebase); /* * Clear this pointer in case this call fails. Otherwise, * kvm_close() will free it again. */ kd->filebase = 0; } if (ISALIVE(kd)) { mib[0] = CTL_KERN; mib[1] = KERN_FILE2; mib[2] = op; mib[3] = arg; mib[4] = esize; mib[5] = 0; /* find size and alloc buffer */ rv = sysctl(mib, 6, NULL, &size, NULL, 0); if (rv == -1) { if (kd->vmfd != -1) goto deadway; _kvm_syserr(kd, kd->program, "kvm_getfile2"); return (NULL); } kd->filebase = _kvm_malloc(kd, size); if (kd->filebase == NULL) return (NULL); /* get actual data */ mib[5] = size / esize; rv = sysctl(mib, 6, kd->filebase, &size, NULL, 0); if (rv == -1) { _kvm_syserr(kd, kd->program, "kvm_getfile2"); return (NULL); } *cnt = size / esize; return ((struct kinfo_file2 *)kd->filebase); } else { if (esize > sizeof(struct kinfo_file2)) { _kvm_syserr(kd, kd->program, "kvm_getfile2: unknown fields requested: libkvm out of date?"); return (NULL); } deadway: switch (op) { case KERN_FILE_BYFILE: if (arg != 0) { _kvm_err(kd, kd->program, "%s: invalid argument"); return (NULL); } return (kvm_deadfile2_byfile(kd, op, arg, esize, cnt)); break; case KERN_FILE_BYPID: case KERN_FILE_BYUID: return (kvm_deadfile2_byid(kd, op, arg, esize, cnt)); break; default: return (NULL); } } } static struct kinfo_file2 * kvm_deadfile2_byfile(kvm_t *kd, int op, int arg, size_t esize, int *cnt) { size_t size; struct nlist nl[3], *p; int buflen = kd->arglen, n = 0; char *where = kd->argspc; struct kinfo_file2 *kf = NULL; struct file *fp, file; struct filelist filehead; int nfiles; nl[0].n_name = "_filehead"; nl[1].n_name = "_nfiles"; nl[2].n_name = 0; if (kvm_nlist(kd, nl) != 0) { for (p = nl; p->n_type != 0; ++p) ; _kvm_err(kd, kd->program, "%s: no such symbol", p->n_name); return (NULL); } if (KREAD(kd, nl[0].n_value, &filehead)) { _kvm_err(kd, kd->program, "can't read filehead"); return (NULL); } if (KREAD(kd, nl[1].n_value, &nfiles)) { _kvm_err(kd, kd->program, "can't read nfiles"); return (NULL); } size = (nfiles + 10) * sizeof(struct kinfo_file2); kd->filebase = _kvm_malloc(kd, size); if (kd->filebase == NULL) return (NULL); LIST_FOREACH(fp, &filehead, f_list) { if (buflen < sizeof(struct kinfo_file2)) break; if (KREAD(kd, (long)fp, &file)) { _kvm_err(kd, kd->program, "can't read kfp"); return (NULL); } kf = (struct kinfo_file2 *)where; where += sizeof(struct kinfo_file2); buflen -= sizeof(struct kinfo_file2); n++; if (fill_file2(kd, kf, fp, NULL, NULL, 0) == -1) return (NULL); } if (n != nfiles) { _kvm_err(kd, kd->program, "inconsistent nfiles"); return (NULL); } *cnt = n; return (kf); } static struct kinfo_file2 * kvm_deadfile2_byid(kvm_t *kd, int op, int arg, size_t esize, int *cnt) { size_t size; struct nlist nl[5], *np; int buflen = kd->arglen, n = 0; char *where = kd->argspc; struct kinfo_file2 *kf = NULL; struct file *fp, file; struct filelist filehead; struct filedesc0 filed0; #define filed filed0.fd_fd struct proclist allproc; struct proc *p, proc; struct process process; struct pcred pcred; struct ucred ucred; int i, nfiles, nprocs; nl[0].n_name = "_filehead"; nl[1].n_name = "_nfiles"; nl[2].n_name = "_nprocs"; nl[3].n_name = "_allproc"; nl[4].n_name = 0; if (kvm_nlist(kd, nl) != 0) { for (np = nl; np->n_type != 0; ++np) ; _kvm_err(kd, kd->program, "%s: no such symbol", np->n_name); return (NULL); } if (KREAD(kd, nl[0].n_value, &filehead)) { _kvm_err(kd, kd->program, "can't read filehead"); return (NULL); } if (KREAD(kd, nl[1].n_value, &nfiles)) { _kvm_err(kd, kd->program, "can't read nfiles"); return (NULL); } if (KREAD(kd, nl[2].n_value, &nprocs)) { _kvm_err(kd, kd->program, "can't read nprocs"); return (NULL); } if (KREAD(kd, nl[3].n_value, &allproc)) { _kvm_err(kd, kd->program, "can't read allproc"); return (NULL); } /* this may be more room than we need but counting is expensive */ size = (nfiles + 10) * sizeof(struct kinfo_file2); kd->filebase = _kvm_malloc(kd, size); if (kd->filebase == NULL) return (NULL); LIST_FOREACH(p, &allproc, p_list) { if (buflen < sizeof(struct kinfo_file2)) break; if (KREAD(kd, (u_long)p, &proc)) { _kvm_err(kd, kd->program, "can't read proc at %x", p); return (NULL); } /* skip system, embryonic and undead processes */ if ((proc.p_flag & P_SYSTEM) || proc.p_stat == SIDL || proc.p_stat == SZOMB) continue; if (op == KERN_FILE_BYPID) { if (arg > 0 && proc.p_pid != (pid_t)arg) { /* not the pid we are looking for */ continue; } } else /* if (op == KERN_FILE_BYUID) */ { if (arg > 0 && proc.p_ucred->cr_uid != (uid_t)arg) { /* not the uid we are looking for */ continue; } } if (proc.p_fd == NULL || proc.p_p == NULL) continue; if (KREAD(kd, (u_long)proc.p_p, &process)) { _kvm_err(kd, kd->program, "can't read process at %x", proc.p_p); return (NULL); } proc.p_p = &process; if (KREAD(kd, (u_long)process.ps_cred, &pcred) == 0) KREAD(kd, (u_long)pcred.pc_ucred, &ucred); process.ps_cred = &pcred; pcred.pc_ucred = &ucred; if (KREAD(kd, (u_long)proc.p_fd, &filed0)) { _kvm_err(kd, kd->program, "can't read filedesc at %x", proc.p_fd); return (NULL); } proc.p_fd = &filed; if (proc.p_textvp) { if (buflen < sizeof(struct kinfo_file2)) goto done; kf = (struct kinfo_file2 *)where; where += sizeof(struct kinfo_file2); buflen -= sizeof(struct kinfo_file2); n++; if (fill_file2(kd, kf, NULL, proc.p_textvp, &proc, KERN_FILE_TEXT) == -1) return (NULL); } if (filed.fd_cdir) { if (buflen < sizeof(struct kinfo_file2)) goto done; kf = (struct kinfo_file2 *)where; where += sizeof(struct kinfo_file2); buflen -= sizeof(struct kinfo_file2); n++; if (fill_file2(kd, kf, NULL, filed.fd_cdir, &proc, KERN_FILE_CDIR) == -1) return (NULL); } if (filed.fd_rdir) { if (buflen < sizeof(struct kinfo_file2)) goto done; kf = (struct kinfo_file2 *)where; where += sizeof(struct kinfo_file2); buflen -= sizeof(struct kinfo_file2); n++; if (fill_file2(kd, kf, NULL, filed.fd_rdir, &proc, KERN_FILE_RDIR) == -1) return (NULL); } if (process.ps_tracevp) { if (buflen < sizeof(struct kinfo_file2)) goto done; kf = (struct kinfo_file2 *)where; where += sizeof(struct kinfo_file2); buflen -= sizeof(struct kinfo_file2); n++; if (fill_file2(kd, kf, NULL, process.ps_tracevp, &proc, KERN_FILE_TRACE) == -1) return (NULL); } if (filed.fd_nfiles < 0 || filed.fd_lastfile >= filed.fd_nfiles || filed.fd_freefile > filed.fd_lastfile + 1) { _kvm_err(kd, kd->program, "filedesc corrupted at %x for pid %d", proc.p_fd, proc.p_pid); return (NULL); } for (i = 0; i < filed.fd_nfiles; i++) { if (buflen < sizeof(struct kinfo_file2)) goto done; if ((fp = filed.fd_ofiles[i]) == NULL) continue; if (KREAD(kd, (u_long)fp, &file)) { _kvm_err(kd, kd->program, "can't read file"); return (NULL); } kf = (struct kinfo_file2 *)where; where += sizeof(struct kinfo_file2); buflen -= sizeof(struct kinfo_file2); n++; if (fill_file2(kd, kf, &file, NULL, &proc, i) == -1) return (NULL); } } done: *cnt = n; return (kf); } static int fill_file2(kvm_t *kd, struct kinfo_file2 *kf, struct file *fp, struct vnode *vp, struct proc *p, int fd) { struct ucred f_cred; memset(kf, 0, sizeof(*kf)); kf->fd_fd = fd; /* might not really be an fd */ if (fp != NULL) { /* Fill in f_cred */ if (KREAD(kd, (u_long)fp->f_cred, &f_cred)) { _kvm_err(kd, kd->program, "can't read f_cred"); return (-1); } fp->f_cred = &f_cred; kf->f_fileaddr = PTRTOINT64(fp); kf->f_flag = fp->f_flag; kf->f_iflags = fp->f_iflags; kf->f_type = fp->f_type; kf->f_count = fp->f_count; kf->f_msgcount = fp->f_msgcount; kf->f_ucred = PTRTOINT64(fp->f_cred); kf->f_uid = fp->f_cred->cr_uid; kf->f_gid = fp->f_cred->cr_gid; kf->f_ops = PTRTOINT64(fp->f_ops); kf->f_offset = fp->f_offset; kf->f_data = PTRTOINT64(fp->f_data); kf->f_usecount = fp->f_usecount; if (getuid() == 0 || p->p_ucred->cr_uid == fp->f_cred->cr_uid) { kf->f_rxfer = fp->f_rxfer; kf->f_rwfer = fp->f_wxfer; kf->f_seek = fp->f_seek; kf->f_rbytes = fp->f_rbytes; kf->f_wbytes = fp->f_rbytes; } } else if (vp != NULL) { /* fake it */ kf->f_type = DTYPE_VNODE; kf->f_flag = FREAD; if (fd == KERN_FILE_TRACE) kf->f_flag |= FWRITE; } /* information about the object associated with this file */ switch (kf->f_type) { case DTYPE_VNODE: { struct vnode vbuf; struct mount mount; if (KREAD(kd, (u_long)(fp ? fp->f_data : vp), &vbuf)) { _kvm_err(kd, kd->program, "can't read vnode"); return (-1); } vp = &vbuf; if (KREAD(kd, (u_long)vp->v_mount, &mount)) { _kvm_err(kd, kd->program, "can't read v_mount"); return (-1); } vp->v_mount = &mount; kf->v_un = PTRTOINT64(vp->v_un.vu_socket); kf->v_type = vp->v_type; kf->v_tag = vp->v_tag; kf->v_flag = vp->v_flag; kf->v_data = PTRTOINT64(vp->v_data); kf->v_mount = PTRTOINT64(vp->v_mount); strlcpy(kf->f_mntonname, vp->v_mount->mnt_stat.f_mntonname, sizeof(kf->f_mntonname)); /* Fill in va_fsid, va_fileid, va_mode, va_size, va_rdev */ filestat(kd, kf, vp); break; } case DTYPE_SOCKET: { struct socket sock; struct protosw protosw; struct domain domain; if (KREAD(kd, (u_long)fp->f_data, &sock)) { _kvm_err(kd, kd->program, "can't read socket"); return (-1); } kf->so_type = sock.so_type; kf->so_state = sock.so_state; kf->so_pcb = PTRTOINT64(sock.so_pcb); if (KREAD(kd, (u_long)sock.so_proto, &protosw)) { _kvm_err(kd, kd->program, "can't read protosw"); return (-1); } kf->so_protocol = protosw.pr_protocol; if (KREAD(kd, (u_long)protosw.pr_domain, &domain)) { _kvm_err(kd, kd->program, "can't read domain"); return (-1); } kf->so_family = domain.dom_family; if (!sock.so_pcb) break; switch (kf->so_family) { case AF_INET: { struct inpcb inpcb; if (KREAD(kd, (u_long)sock.so_pcb, &inpcb)) { _kvm_err(kd, kd->program, "can't read inpcb"); return (-1); } kf->inp_ppcb = PTRTOINT64(inpcb.inp_ppcb); kf->inp_lport = inpcb.inp_lport; kf->inp_laddru[0] = inpcb.inp_laddr.s_addr; kf->inp_fport = inpcb.inp_fport; kf->inp_faddru[0] = inpcb.inp_faddr.s_addr; break; } case AF_INET6: { struct inpcb inpcb; #define s6_addr32 __u6_addr.__u6_addr32 if (KREAD(kd, (u_long)sock.so_pcb, &inpcb)) { _kvm_err(kd, kd->program, "can't read inpcb"); return (-1); } kf->inp_ppcb = PTRTOINT64(inpcb.inp_ppcb); kf->inp_lport = inpcb.inp_lport; kf->inp_laddru[0] = inpcb.inp_laddr6.s6_addr32[0]; kf->inp_laddru[1] = inpcb.inp_laddr6.s6_addr32[1]; kf->inp_laddru[2] = inpcb.inp_laddr6.s6_addr32[2]; kf->inp_laddru[3] = inpcb.inp_laddr6.s6_addr32[3]; kf->inp_fport = inpcb.inp_fport; kf->inp_faddru[0] = inpcb.inp_laddr6.s6_addr32[0]; kf->inp_faddru[1] = inpcb.inp_faddr6.s6_addr32[1]; kf->inp_faddru[2] = inpcb.inp_faddr6.s6_addr32[2]; kf->inp_faddru[3] = inpcb.inp_faddr6.s6_addr32[3]; break; } case AF_UNIX: { struct unpcb unpcb; if (KREAD(kd, (u_long)sock.so_pcb, &unpcb)) { _kvm_err(kd, kd->program, "can't read unpcb"); return (-1); } kf->unp_conn = PTRTOINT64(unpcb.unp_conn); break; } } break; } case DTYPE_PIPE: { struct pipe pipe; if (KREAD(kd, (u_long)fp->f_data, &pipe)) { _kvm_err(kd, kd->program, "can't read pipe"); return (-1); } kf->pipe_peer = PTRTOINT64(pipe.pipe_peer); kf->pipe_state = pipe.pipe_state; break; } case DTYPE_KQUEUE: { struct kqueue kqi; if (KREAD(kd, (u_long)fp->f_data, &kqi)) { _kvm_err(kd, kd->program, "can't read kqi"); return (-1); } kf->kq_count = kqi.kq_count; kf->kq_state = kqi.kq_state; break; } case DTYPE_SYSTRACE: { struct fsystrace f; if (KREAD(kd, (u_long)fp->f_data, &f)) { _kvm_err(kd, kd->program, "can't read fsystrace"); return (-1); } kf->str_npolicies = f.npolicies; break; } } /* per-process information for KERN_FILE_BY[PU]ID */ if (p != NULL) { kf->p_pid = p->p_pid; kf->p_uid = p->p_ucred->cr_uid; kf->p_gid = p->p_ucred->cr_gid; strlcpy(kf->p_comm, p->p_comm, sizeof(kf->p_comm)); if (p->p_fd != NULL) kf->fd_ofileflags = p->p_fd->fd_ofileflags[fd]; } return (0); } mode_t _kvm_getftype(enum vtype v_type) { mode_t ftype = 0; switch (v_type) { case VREG: ftype = S_IFREG; break; case VDIR: ftype = S_IFDIR; break; case VBLK: ftype = S_IFBLK; break; case VCHR: ftype = S_IFCHR; break; case VLNK: ftype = S_IFLNK; break; case VSOCK: ftype = S_IFSOCK; break; case VFIFO: ftype = S_IFIFO; break; case VNON: case VBAD: break; } return (ftype); } static int ufs_filestat(kvm_t *kd, struct kinfo_file2 *kf, struct vnode *vp) { struct inode inode; struct ufs1_dinode di1; if (KREAD(kd, (u_long)VTOI(vp), &inode)) { _kvm_err(kd, kd->program, "can't read inode at %p", VTOI(vp)); return (-1); } if (KREAD(kd, (u_long)inode.i_din1, &di1)) { _kvm_err(kd, kd->program, "can't read dinode at %p", inode.i_din1); return (-1); } inode.i_din1 = &di1; kf->va_fsid = inode.i_dev & 0xffff; kf->va_fileid = (long)inode.i_number; kf->va_mode = inode.i_ffs1_mode; kf->va_size = inode.i_ffs1_size; kf->va_rdev = inode.i_ffs1_rdev; return (0); } static int ext2fs_filestat(kvm_t *kd, struct kinfo_file2 *kf, struct vnode *vp) { struct inode inode; struct ext2fs_dinode e2di; if (KREAD(kd, (u_long)VTOI(vp), &inode)) { _kvm_err(kd, kd->program, "can't read inode at %p", VTOI(vp)); return (-1); } if (KREAD(kd, (u_long)inode.i_e2din, &e2di)) { _kvm_err(kd, kd->program, "can't read dinode at %p", inode.i_e2din); return (-1); } inode.i_e2din = &e2di; kf->va_fsid = inode.i_dev & 0xffff; kf->va_fileid = (long)inode.i_number; kf->va_mode = inode.i_e2fs_mode; kf->va_size = inode.i_e2fs_size; kf->va_rdev = 0; /* XXX */ return (0); } static int msdos_filestat(kvm_t *kd, struct kinfo_file2 *kf, struct vnode *vp) { struct denode de; struct msdosfsmount mp; if (KREAD(kd, (u_long)VTODE(vp), &de)) { _kvm_err(kd, kd->program, "can't read denode at %p", VTODE(vp)); return (-1); } if (KREAD(kd, (u_long)de.de_pmp, &mp)) { _kvm_err(kd, kd->program, "can't read mount struct at %p", de.de_pmp); return (-1); } kf->va_fsid = de.de_dev & 0xffff; kf->va_fileid = 0; /* XXX see msdosfs_vptofh() for more info */ kf->va_mode = (mp.pm_mask & 0777) | _kvm_getftype(vp->v_type); kf->va_size = de.de_FileSize; kf->va_rdev = 0; /* msdosfs doesn't support device files */ return (0); } static int nfs_filestat(kvm_t *kd, struct kinfo_file2 *kf, struct vnode *vp) { struct nfsnode nfsnode; if (KREAD(kd, (u_long)VTONFS(vp), &nfsnode)) { _kvm_err(kd, kd->program, "can't read nfsnode at %p", VTONFS(vp)); return (-1); } kf->va_fsid = nfsnode.n_vattr.va_fsid; kf->va_fileid = nfsnode.n_vattr.va_fileid; kf->va_size = nfsnode.n_size; kf->va_rdev = nfsnode.n_vattr.va_rdev; kf->va_mode = (mode_t)nfsnode.n_vattr.va_mode | _kvm_getftype(vp->v_type); return (0); } static int nnpfs_filestat(kvm_t *kd, struct kinfo_file2 *kf, struct vnode *vp) { struct nnpfs_node nnpfs_node; if (KREAD(kd, (u_long)VNODE_TO_XNODE(vp), &nnpfs_node)) { _kvm_err(kd, kd->program, "can't read nnpfs_node at %p", VTOI(vp)); return (-1); } kf->va_fsid = nnpfs_node.attr.va_fsid; kf->va_fileid = (long)nnpfs_node.attr.va_fileid; kf->va_mode = nnpfs_node.attr.va_mode; kf->va_size = nnpfs_node.attr.va_size; kf->va_rdev = nnpfs_node.attr.va_rdev; return (0); } static int spec_filestat(kvm_t *kd, struct kinfo_file2 *kf, struct vnode *vp) { struct specinfo specinfo; struct vnode parent; if (KREAD(kd, (u_long)vp->v_specinfo, &specinfo)) { _kvm_err(kd, kd->program, "can't read specinfo at %p", vp->v_specinfo); return (-1); } vp->v_specinfo = &specinfo; if (KREAD(kd, (u_long)vp->v_specparent, &parent)) { _kvm_err(kd, kd->program, "can't read parent vnode at %p", vp->v_specparent); return (-1); } if (ufs_filestat(kd, kf, vp)) return (-1); return (0); } static int filestat(kvm_t *kd, struct kinfo_file2 *kf, struct vnode *vp) { int ret = 0; if (vp->v_type != VNON && vp->v_type != VBAD) { switch (vp->v_tag) { case VT_UFS: case VT_MFS: ret = ufs_filestat(kd, kf, vp); break; case VT_NFS: ret = nfs_filestat(kd, kf, vp); break; case VT_EXT2FS: ret = ext2fs_filestat(kd, kf, vp); break; case VT_ISOFS: ret = _kvm_stat_cd9660(kd, kf, vp); break; case VT_MSDOSFS: ret = msdos_filestat(kd, kf, vp); break; case VT_NNPFS: ret = nnpfs_filestat(kd, kf, vp); break; case VT_UDF: ret = _kvm_stat_udf(kd, kf, vp); break; case VT_NTFS: ret = _kvm_stat_ntfs(kd, kf, vp); break; case VT_NON: if (vp->v_flag & VCLONE) ret = spec_filestat(kd, kf, vp); break; } } return (ret); }