i86xpv/io/privcmd.c

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */

/*
 * Copyright 2009 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 */

#include <sys/xpv_user.h>

#include <sys/types.h>
#include <sys/file.h>
#include <sys/errno.h>
#include <sys/open.h>
#include <sys/cred.h>
#include <sys/conf.h>
#include <sys/stat.h>
#include <sys/modctl.h>
#include <sys/ddi.h>
#include <sys/sunddi.h>
#include <sys/vmsystm.h>
#include <sys/sdt.h>
#include <sys/hypervisor.h>
#include <sys/xen_errno.h>
#include <sys/policy.h>

#include <vm/hat_i86.h>
#include <vm/hat_pte.h>
#include <vm/seg_mf.h>

#include <xen/sys/privcmd.h>
#include <sys/privcmd_impl.h>

static dev_info_t *privcmd_devi;

/*ARGSUSED*/
static int
privcmd_getinfo(dev_info_t *devi, ddi_info_cmd_t cmd, void *arg, void **result)
{
    switch (cmd) {
    case DDI_INFO_DEVT2DEVINFO:
    case DDI_INFO_DEVT2INSTANCE:
        break;
    default:
        return (DDI_FAILURE);
    }

    switch (getminor((dev_t)arg)) {
    case PRIVCMD_MINOR:
        break;
    default:
        return (DDI_FAILURE);
    }

    if (cmd == DDI_INFO_DEVT2INSTANCE)
        *result = 0;
    else
        *result = privcmd_devi;
    return (DDI_SUCCESS);
}

static int
privcmd_attach(dev_info_t *devi, ddi_attach_cmd_t cmd)
{
    if (cmd != DDI_ATTACH)
        return (DDI_FAILURE);

    if (ddi_create_minor_node(devi, PRIVCMD_NODE,
        S_IFCHR, PRIVCMD_MINOR, DDI_PSEUDO, 0) != DDI_SUCCESS)
        return (DDI_FAILURE);

    privcmd_devi = devi;
    ddi_report_dev(devi);
    return (DDI_SUCCESS);
}

static int
privcmd_detach(dev_info_t *devi, ddi_detach_cmd_t cmd)
{
    if (cmd != DDI_DETACH)
        return (DDI_FAILURE);
    ddi_remove_minor_node(devi, NULL);
    privcmd_devi = NULL;
    return (DDI_SUCCESS);
}

/*ARGSUSED1*/
static int
privcmd_open(dev_t *dev, int flag, int otyp, cred_t *cr)
{
    return (getminor(*dev) == PRIVCMD_MINOR ? 0 : ENXIO);
}

/*
 * Map a contiguous set of machine frames in a foreign domain.
 * Used in the following way:
 *
 *  privcmd_mmap_t p;
 *  privcmd_mmap_entry_t e;
 *
 *  addr = mmap(NULL, size, prot, MAP_SHARED, fd, 0);
 *  p.num = number of privcmd_mmap_entry_t's
 *  p.dom = domid;
 *  p.entry = &e;
 *  e.va = addr;
 *  e.mfn = mfn;
 *  e.npages = btopr(size);
 *  ioctl(fd, IOCTL_PRIVCMD_MMAP, &p);
 */
/*ARGSUSED2*/
int
do_privcmd_mmap(void *uarg, int mode, cred_t *cr)
{
    privcmd_mmap_t __mmapcmd, *mmc = &__mmapcmd;
    privcmd_mmap_entry_t *umme;
    struct as *as = curproc->p_as;
    struct seg *seg;
    int i, error = 0;

    if (ddi_copyin(uarg, mmc, sizeof (*mmc), mode))
        return (EFAULT);

    DTRACE_XPV3(mmap__start, domid_t, mmc->dom, int, mmc->num,
        privcmd_mmap_entry_t *, mmc->entry);

    if (mmc->dom == DOMID_SELF) {
        error = ENOTSUP;    /* Too paranoid? */
        goto done;
    }

    for (umme = mmc->entry, i = 0; i < mmc->num; i++, umme++) {
        privcmd_mmap_entry_t __mmapent, *mme = &__mmapent;
        caddr_t addr;

        if (ddi_copyin(umme, mme, sizeof (*mme), mode)) {
            error = EFAULT;
            break;
        }

        DTRACE_XPV3(mmap__entry, ulong_t, mme->va, ulong_t, mme->mfn,
            ulong_t, mme->npages);

        if (mme->mfn == MFN_INVALID) {
            error = EINVAL;
            break;
        }

        addr = (caddr_t)mme->va;

        /*
         * Find the segment we want to mess with, then add
         * the mfn range to the segment.
         */
        AS_LOCK_ENTER(as, RW_READER);
        if ((seg = as_findseg(as, addr, 0)) == NULL ||
            addr + mmu_ptob(mme->npages) > seg->s_base + seg->s_size)
            error = EINVAL;
        else
            error = segmf_add_mfns(seg, addr,
                mme->mfn, mme->npages, mmc->dom);
        AS_LOCK_EXIT(as);

        if (error != 0)
            break;
    }

done:
    DTRACE_XPV1(mmap__end, int, error);

    return (error);
}

/*
 * Set up the address range to map to an array of mfns in
 * a foreign domain.  Used in the following way:
 *
 *  privcmd_mmap_batch_t p;
 *
 *  addr = mmap(NULL, size, prot, MAP_SHARED, fd, 0);
 *  p.num = number of pages
 *  p.dom = domid
 *  p.addr = addr;
 *  p.arr = array of mfns, indexed 0 .. p.num - 1
 *  ioctl(fd, IOCTL_PRIVCMD_MMAPBATCH, &p);
 */
/*ARGSUSED2*/
static int
do_privcmd_mmapbatch(void *uarg, int mode, cred_t *cr)
{
    privcmd_mmapbatch_t __mmapbatch, *mmb = &__mmapbatch;
    struct as *as = curproc->p_as;
    struct seg *seg;
    int i, error = 0;
    caddr_t addr;
    ulong_t *ulp;

    if (ddi_copyin(uarg, mmb, sizeof (*mmb), mode))
        return (EFAULT);

    DTRACE_XPV3(mmapbatch__start, domid_t, mmb->dom, int, mmb->num,
        caddr_t, mmb->addr);

    addr = (caddr_t)mmb->addr;
    AS_LOCK_ENTER(as, RW_READER);
    if ((seg = as_findseg(as, addr, 0)) == NULL ||
        addr + ptob(mmb->num) > seg->s_base + seg->s_size) {
        error = EINVAL;
        goto done;
    }

    for (i = 0, ulp = mmb->arr;
        i < mmb->num; i++, addr += PAGESIZE, ulp++) {
        mfn_t mfn;

        if (fulword(ulp, &mfn) != 0) {
            error = EFAULT;
            break;
        }

        if (mfn == MFN_INVALID) {
            /*
             * This mfn is invalid and should not be added to
             * segmf, as we'd only cause an immediate EFAULT when
             * we tried to fault it in.
             */
            mfn |= XEN_DOMCTL_PFINFO_XTAB;
            continue;
        }

        if (segmf_add_mfns(seg, addr, mfn, 1, mmb->dom) == 0)
            continue;

        /*
         * Tell the process that this MFN could not be mapped, so it
         * won't later try to access it.
         */
        mfn |= XEN_DOMCTL_PFINFO_XTAB;
        if (sulword(ulp, mfn) != 0) {
            error = EFAULT;
            break;
        }
    }

done:
    AS_LOCK_EXIT(as);

    DTRACE_XPV3(mmapbatch__end, int, error, struct seg *, seg, caddr_t,
        mmb->addr);

    return (error);
}

/*ARGSUSED*/
static int
privcmd_ioctl(dev_t dev, int cmd, intptr_t arg, int mode, cred_t *cr, int *rval)
{
    if (secpolicy_xvm_control(cr))
        return (EPERM);

    /*
     * Everything is a -native- data type.
     */
    if ((mode & FMODELS) != FNATIVE)
        return (EOVERFLOW);

    switch (cmd) {
    case IOCTL_PRIVCMD_HYPERCALL:
        return (do_privcmd_hypercall((void *)arg, mode, cr, rval));
    case IOCTL_PRIVCMD_MMAP:
        if (DOMAIN_IS_PRIVILEGED(xen_info))
            return (do_privcmd_mmap((void *)arg, mode, cr));
        break;
    case IOCTL_PRIVCMD_MMAPBATCH:
        if (DOMAIN_IS_PRIVILEGED(xen_info))
            return (do_privcmd_mmapbatch((void *)arg, mode, cr));
        break;
    default:
        break;
    }
    return (EINVAL);
}

/*
 * The real magic happens in the segmf segment driver.
 */
/*ARGSUSED8*/
static int
privcmd_segmap(dev_t dev, off_t off, struct as *as, caddr_t *addrp,
    off_t len, uint_t prot, uint_t maxprot, uint_t flags, cred_t *cr)
{
    struct segmf_crargs a;
    int error;

    if (secpolicy_xvm_control(cr))
        return (EPERM);

    as_rangelock(as);
    if ((flags & MAP_FIXED) == 0) {
        map_addr(addrp, len, (offset_t)off, 0, flags);
        if (*addrp == NULL) {
            error = ENOMEM;
            goto rangeunlock;
        }
    } else {
        /*
         * User specified address
         */
        (void) as_unmap(as, *addrp, len);
    }

    /*
     * The mapping *must* be MAP_SHARED at offset 0.
     *
     * (Foreign pages are treated like device memory; the
     * ioctl interface allows the backing objects to be
     * arbitrarily redefined to point at any machine frame.)
     */
    if ((flags & MAP_TYPE) != MAP_SHARED || off != 0) {
        error = EINVAL;
        goto rangeunlock;
    }

    a.dev = dev;
    a.prot = (uchar_t)prot;
    a.maxprot = (uchar_t)maxprot;
    error = as_map(as, *addrp, len, segmf_create, &a);

rangeunlock:
    as_rangeunlock(as);
    return (error);
}

static struct cb_ops privcmd_cb_ops = {
    privcmd_open,
    nulldev,    /* close */
    nodev,      /* strategy */
    nodev,      /* print */
    nodev,      /* dump */
    nodev,      /* read */
    nodev,      /* write */
    privcmd_ioctl,
    nodev,      /* devmap */
    nodev,      /* mmap */
    privcmd_segmap,
    nochpoll,   /* poll */
    ddi_prop_op,
    NULL,
    D_64BIT | D_NEW | D_MP
};

static struct dev_ops privcmd_dv_ops = {
    DEVO_REV,
    0,
    privcmd_getinfo,
    nulldev,        /* identify */
    nulldev,        /* probe */
    privcmd_attach,
    privcmd_detach,
    nodev,          /* reset */
    &privcmd_cb_ops,
    0,          /* struct bus_ops */
    NULL,           /* power */
    ddi_quiesce_not_needed,     /* quiesce */
};

static struct modldrv modldrv = {
    &mod_driverops,
    "privcmd driver",
    &privcmd_dv_ops
};

static struct modlinkage modl = {
    MODREV_1,
    &modldrv
};

int
_init(void)
{
    return (mod_install(&modl));
}

int
_fini(void)
{
    return (mod_remove(&modl));
}

int
_info(struct modinfo *modinfo)
{
    return (mod_info(&modl, modinfo));
}