/*
 * VNODEINFO.C
 *
 * cc -I/usr/src/sys vnodeinfo.c -o /usr/local/bin/vnodeinfo -lkvm
 *
 * vnodeinfo
 *
 * Dump the mountlist and related vnodes.
 *
 *
 * Copyright (c) 2004 The DragonFly Project.  All rights reserved.
 *
 * This code is derived from software contributed to The DragonFly Project
 * by Matthew Dillon <dillon@backplane.com>
 *
 * 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 DragonFly Project 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 COPYRIGHT HOLDERS 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
 * COPYRIGHT HOLDERS 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.
 *
 */

#define _KERNEL_STRUCTURES
#include <sys/param.h>
#include <sys/user.h>
#include <sys/malloc.h>
#include <sys/signalvar.h>
#include <sys/namecache.h>
#include <sys/mount.h>
#include <sys/vnode.h>
#include <sys/buf.h>

#include <vm/vm.h>
#include <vm/vm_page.h>
#include <vm/vm_kern.h>
#include <vm/vm_object.h>
#include <vm/swap_pager.h>
#include <vm/vnode_pager.h>

#include <vfs/ufs/quota.h>
#include <vfs/ufs/inode.h>

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <fcntl.h>
#include <kvm.h>
#include <nlist.h>
#include <getopt.h>

struct nlist Nl[] = {
    { "_mountlist" },
    { "_vnode_list_hash" },
    { "_ncpus" },
    { NULL }
};

static void kkread(kvm_t *kd, u_long addr, void *buf, size_t nbytes);
static struct mount *dumpmount(kvm_t *kd, struct mount *mp);
static struct vnode *dumpvp(kvm_t *kd, struct vnode *vp, int whichlist, char *vfc_name);
static void dumpbufs(kvm_t *kd, void *bufp, const char *id);
static void dumplocks(kvm_t *kd, struct lockf *lockf);
static void dumplockinfo(kvm_t *kd, struct lockf_range *item);
static int getobjpages(kvm_t *kd, struct vm_object *obj);
static int getobjvnpsize(kvm_t *kd, struct vm_object *obj);

static const struct dump_private_data {
        char vfc_name[MFSNAMELEN];
        void (*dumpfn)(kvm_t *, void *);
} dumplist[] = {
        { "", NULL }
};

int tracebufs = 0;
int tracelocks = 0;
int withnames = 0;
int fsprivate = 0;

int
main(int ac, char **av)
{
    struct mount *mp;
    struct vnode_index *vib;
    struct vnode_index vni;
    kvm_t *kd;
    int ch;
    int ncpus = 0;
    const char *corefile = NULL;
    const char *sysfile = NULL;

    while ((ch = getopt(ac, av, "alnbM:N:p")) != -1) {
        switch(ch) {
        case 'b':
            tracebufs = 1;
            break;
        case 'n':
            withnames = 1;
            break;
        case 'l':
            tracelocks = 1;
            break;
        case 'a':
            tracebufs = 1;
            tracelocks = 1;
            withnames = 1;
            fsprivate = 1;
            break;
        case 'p':
            fsprivate = 1;
            break;
        case 'M':
            corefile = optarg;
            break;
        case 'N':
            sysfile = optarg;
            break;
        default:
            fprintf(stderr, "%s [-pbnla] [-M core] [-N system]\n", av[0]);
            exit(1);
        }
    }

    if ((kd = kvm_open(sysfile, corefile, NULL, O_RDONLY, "kvm:")) == NULL) {
        perror("kvm_open");
        exit(1);
    }
    if (kvm_nlist(kd, Nl) != 0) {
        perror("kvm_nlist");
        exit(1);
    }

    /* Mount points and their private data */
    kkread(kd, Nl[0].n_value, &mp, sizeof(mp));
    while (mp)
        mp = dumpmount(kd, mp);

    /*
     * Get ncpus for the vnode lists, we could get it with a sysctl
     * but since we're reading kernel memory, take advantage of it.
     * Also read the base address of vnode_list_hash.
     */
    kkread(kd, Nl[1].n_value, &vib, sizeof(vib));
    kkread(kd, Nl[2].n_value, &ncpus, sizeof(ncpus));

    /* Per-CPU list of inactive vnodes */
    printf("INACTIVELIST {\n");

    for (int i = 0; i < ncpus; i++) {
            struct vnode *vp;

            kkread(kd, (u_long)(vib + i), &vni, sizeof(vni));
            vp = vni.inactive_list.tqh_first;
            for (; vp != NULL;)
                    vp = dumpvp(kd, vp, 0, NULL);
    }
    printf("}\n");

    /* Per-CPU list of active vnodes */
    printf("ACTIVELIST {\n");
    for (int i = 0; i < ncpus; i++) {
            struct vnode *vp;

            kkread(kd, (u_long)(vib + i), &vni, sizeof(vni));
            vp = vni.active_list.tqh_first;
            for (; vp;)
                    vp = dumpvp(kd, vp, 0,
                        NULL);
    }
    printf("}\n");
    return(0);
}

static struct mount *
dumpmount(kvm_t *kd, struct mount *mp)
{
    struct mount mnt;
    struct vnode *vp;
    struct vfsconf vfc;

    kkread(kd, (u_long)mp, &mnt, sizeof(mnt));
    printf("MOUNTPOINT %s on %s {\n",
        mnt.mnt_stat.f_mntfromname, mnt.mnt_stat.f_mntonname);
    printf("    lk_flags %08x count %016jx holder = %p\n",
        mnt.mnt_lock.lk_flags, mnt.mnt_lock.lk_count,
        mnt.mnt_lock.lk_lockholder);
    printf("    mnt_flag %08x mnt_kern_flag %08x\n",
        mnt.mnt_flag, mnt.mnt_kern_flag);
    printf("    mnt_nvnodelistsize %d\n", mnt.mnt_nvnodelistsize);
    printf("    mnt_stat.f_fsid %08x %08x\n", mnt.mnt_stat.f_fsid.val[0],
        mnt.mnt_stat.f_fsid.val[1]);

    /* Dump fs private node data */
    kkread(kd, (u_long)mnt.mnt_vfc, &vfc, sizeof(vfc));
    vp = mnt.mnt_nvnodelist.tqh_first;
    while (vp)
            vp = dumpvp(kd, vp, 1, vfc.vfc_name);

    printf("}\n");

    return(mnt.mnt_list.tqe_next);
}

static const char *
vtype(enum vtype type)
{
    static char buf[32];

    switch(type) {
    case VNON:
        return("VNON");
    case VREG:
        return("VREG");
    case VDIR:
        return("VDIR");
    case VBLK:
        return("VBLK");
    case VCHR:
        return("VCHR");
    case VLNK:
        return("VLNK");
    case VSOCK:
        return("VSOCK");
    case VFIFO:
        return("VFIFO");
    case VBAD:
        return("VBAD");
    default:
        break;
    }
    snprintf(buf, sizeof(buf), "%d", (int)type);
    return(buf);
}

static struct vnode *
dumpvp(kvm_t *kd, struct vnode *vp, int whichlist, char *vfc_name)
{
    struct vnode vn;

    kkread(kd, (u_long)vp, &vn, sizeof(vn));

    printf("    vnode %p.%d refcnt %08x auxcnt %d type=%s flags %08x",
        vp, vn.v_state, vn.v_refcnt, vn.v_auxrefs, vtype(vn.v_type), vn.v_flag);

    if ((vn.v_flag & VOBJBUF) && vn.v_object) {
        int npages = getobjpages(kd, vn.v_object);
        int vnpsize = getobjvnpsize(kd, vn.v_object);
        if (npages || vnpsize)
            printf(" vmobjpgs=%d vnpsize=%d", npages, vnpsize);
    }

    if (vn.v_flag & VROOT)
        printf(" VROOT");
    if (vn.v_flag & VTEXT)
        printf(" VTEXT");
    if (vn.v_flag & VSYSTEM)
        printf(" VSYSTEM");
    if (vn.v_flag & VISTTY)
        printf(" VISTTY");
#ifdef VXLOCK
    if (vn.v_flag & VXLOCK)
        printf(" VXLOCK");
    if (vn.v_flag & VXWANT)
        printf(" VXWANT");
#endif
#ifdef VRECLAIMED
    if (vn.v_flag & VRECLAIMED)
        printf(" VRECLAIMED");
    if (vn.v_flag & VINACTIVE)
        printf(" VINACTIVE");
#endif
    if (vn.v_flag & VOBJBUF)
        printf(" VOBJBUF");
#ifdef VSWAPCACHE
    if (vn.v_flag & VSWAPCACHE)
        printf(" VSWAPCACHE");
#endif
    switch(vn.v_flag & (VAGE0 | VAGE1)) {
    case 0:
        printf(" VAGE0");
        break;
    case VAGE0:
        printf(" VAGE1");
        break;
    case VAGE1:
        printf(" VAGE2");
        break;
    case VAGE0 | VAGE1:
        printf(" VAGE3");
        break;
    }
#ifdef VDOOMED
    if (vn.v_flag & VDOOMED)
        printf(" VDOOMED");
#endif
#ifdef VINFREE
    if (vn.v_flag & VINFREE)
        printf(" VINFREE");
#endif
    if (vn.v_flag & VONWORKLST)
        printf(" VONWORKLST");
    if (vn.v_flag & VOBJDIRTY)
        printf(" VOBJDIRTY");
    if (vn.v_flag & VMAYHAVELOCKS)
        printf(" VMAYHAVELOCKS");

    printf("\n");

    if (vn.v_lock.lk_count || vn.v_lock.lk_lockholder != NULL) {
        printf("\tlk_flags %08x count %016jx holder = %p\n",
            vn.v_lock.lk_flags, vn.v_lock.lk_count,
            vn.v_lock.lk_lockholder);
    }

    if (withnames && TAILQ_FIRST(&vn.v_namecache)) {
        struct namecache ncp;
        int nlen;
        char buf[1024];

        kkread(kd, (u_long)TAILQ_FIRST(&vn.v_namecache), &ncp, sizeof(ncp));
        if ((nlen = ncp.nc_nlen) >= sizeof(buf))
                nlen = sizeof(buf) - 1;
        if (nlen < 0)
                nlen = 0;
        if (nlen) {
                kkread(kd, (u_long)ncp.nc_name, buf, nlen);
                buf[nlen] = 0;
                printf("\tfilename %s\n", buf);
        }
    }

    if (tracebufs) {
        if (vn.v_rbclean_tree.rbh_root) {
            printf("\tCLEAN BUFFERS\n");
            dumpbufs(kd, vn.v_rbclean_tree.rbh_root, "ROOT");
        }
        if (vn.v_rbdirty_tree.rbh_root) {
            printf("\tDIRTY BUFFERS\n");
            dumpbufs(kd, vn.v_rbdirty_tree.rbh_root, "ROOT");
        }
    }

    if (tracelocks) {
        if (vn.v_tag == VT_UFS && vn.v_data) {
            struct inode *ip = vn.v_data;
            struct lockf lockf;

            kkread(kd, (u_long)&ip->i_lockf, &lockf, sizeof(lockf));
            dumplocks(kd, &lockf);
        }
    }

    if (fsprivate && vfc_name) {
            /*
             * Actually find whether the filesystem can dump
             * detailed inode information out of the vnode
             */
            const struct dump_private_data *dpd;

            for (dpd = dumplist; dpd->dumpfn != NULL; dpd++) {
                    if ((strcmp(dpd->vfc_name, vfc_name) == 0) &&
                        vn.v_data != NULL)
                            dpd->dumpfn(kd, vn.v_data);
            }
    }

    if (whichlist)
        return(vn.v_nmntvnodes.tqe_next);
    else
        return(vn.v_list.tqe_next);
}

static void
dumpbufs(kvm_t *kd, void *bufp, const char *id)
{
        struct buf buf;

        kkread(kd, (u_long)bufp, &buf, sizeof(buf));
        printf("\t    %-8s %p loffset %012lx/%05x foffset %08lx",
                id, bufp,
                buf.b_bio1.bio_offset,
                buf.b_bufsize,
                buf.b_bio2.bio_offset);
        printf(" q=%d count=%016jx flags=%08x refs=%08x dep=%p",
                buf.b_qindex, buf.b_lock.lk_count,
                buf.b_flags, buf.b_refs, buf.b_dep.lh_first);
        printf("\n");

        if (buf.b_rbnode.rbe_left)
            dumpbufs(kd, buf.b_rbnode.rbe_left, "LEFT");
        if (buf.b_rbnode.rbe_right)
            dumpbufs(kd, buf.b_rbnode.rbe_right, "RIGHT");
}

static void
dumplocks(kvm_t *kd, struct lockf *lockf)
{
        struct lockf_range item;
        struct lockf_range *scan;

        if ((scan = TAILQ_FIRST(&lockf->lf_range)) != NULL) {
                printf("\tLOCKS\n");
                do {
                        kkread(kd, (u_long)scan, &item, sizeof(item));
                        dumplockinfo(kd, &item);
                } while ((scan = TAILQ_NEXT(&item, lf_link)) != NULL);
                printf("\n");
        }
        if ((scan = TAILQ_FIRST(&lockf->lf_blocked)) != NULL) {
                printf("\tBLKED\n");
                do {
                        kkread(kd, (u_long)scan, &item, sizeof(item));
                        dumplockinfo(kd, &item);
                } while ((scan = TAILQ_NEXT(&item, lf_link)) != NULL);
                printf("\n");
        }

}

static void
dumplockinfo(kvm_t *kd, struct lockf_range *item)
{
        int ownerpid;

        if (item->lf_owner && (item->lf_flags & F_POSIX)) {
                kkread(kd, (u_long)&item->lf_owner->p_pid,
                        &ownerpid, sizeof(ownerpid));
        } else {
                ownerpid = -1;
        }

        printf("\t    ty=%d flgs=%04x %ld-%ld owner=%d\n",
                item->lf_type, item->lf_flags,
                item->lf_start, item->lf_end,
                ownerpid
        );
}

static
int
getobjpages(kvm_t *kd, struct vm_object *obj)
{
        struct vm_object vmobj;

        kkread(kd, (u_long)obj, &vmobj, sizeof(vmobj));
        return(vmobj.resident_page_count);
}

static
int
getobjvnpsize(kvm_t *kd, struct vm_object *obj)
{
        struct vm_object vmobj;

        kkread(kd, (u_long)obj, &vmobj, sizeof(vmobj));
        return ((int)vmobj.size);
}

static void
kkread(kvm_t *kd, u_long addr, void *buf, size_t nbytes)
{
    if (kvm_read(kd, addr, buf, nbytes) != nbytes) {
        perror("kvm_read");
        exit(1);
    }
}