Due to races clean blocks can remain cached, remove a conditional from

[dragonfly.git] / sys / kern / sys_generic.c

/*
 * Copyright (c) 1982, 1986, 1989, 1993
 *      The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * 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. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *      This product includes software developed by the University of
 *      California, Berkeley and its contributors.
 * 4. 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.
 *
 *      @(#)sys_generic.c       8.5 (Berkeley) 1/21/94
 * $FreeBSD: src/sys/kern/sys_generic.c,v 1.55.2.10 2001/03/17 10:39:32 peter Exp $
 * $DragonFly: src/sys/kern/sys_generic.c,v 1.49 2008/05/05 22:09:44 dillon Exp $
 */

#include "opt_ktrace.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/sysproto.h>
#include <sys/filedesc.h>
#include <sys/filio.h>
#include <sys/fcntl.h>
#include <sys/file.h>
#include <sys/proc.h>
#include <sys/signalvar.h>
#include <sys/socketvar.h>
#include <sys/uio.h>
#include <sys/kernel.h>
#include <sys/kern_syscall.h>
#include <sys/malloc.h>
#include <sys/mapped_ioctl.h>
#include <sys/poll.h>
#include <sys/queue.h>
#include <sys/resourcevar.h>
#include <sys/sysctl.h>
#include <sys/sysent.h>
#include <sys/buf.h>
#ifdef KTRACE
#include <sys/ktrace.h>
#endif
#include <vm/vm.h>
#include <vm/vm_page.h>
#include <sys/file2.h>

#include <machine/limits.h>

static MALLOC_DEFINE(M_IOCTLOPS, "ioctlops", "ioctl data buffer");
static MALLOC_DEFINE(M_IOCTLMAP, "ioctlmap", "mapped ioctl handler buffer");
static MALLOC_DEFINE(M_SELECT, "select", "select() buffer");
MALLOC_DEFINE(M_IOV, "iov", "large iov's");

static int      doselect(int nd, fd_set *in, fd_set *ou, fd_set *ex,
                        struct timeval *tv, int *res);
static int      pollscan (struct proc *, struct pollfd *, u_int, int *);
static int      selscan (struct proc *, fd_mask **, fd_mask **,
                        int, int *);
static int      dofileread(int, struct file *, struct uio *, int, int *);
static int      dofilewrite(int, struct file *, struct uio *, int, int *);

/*
 * Read system call.
 *
 * MPSAFE
 */
int
sys_read(struct read_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov;
        int error;

        aiov.iov_base = uap->buf;
        aiov.iov_len = uap->nbyte;
        auio.uio_iov = &aiov;
        auio.uio_iovcnt = 1;
        auio.uio_offset = -1;
        auio.uio_resid = uap->nbyte;
        auio.uio_rw = UIO_READ;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        if (auio.uio_resid < 0)
                error = EINVAL;
        else
                error = kern_preadv(uap->fd, &auio, 0, &uap->sysmsg_result);
        return(error);
}

/*
 * Positioned (Pread) read system call
 *
 * MPSAFE
 */
int
sys_extpread(struct extpread_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov;
        int error;
        int flags;

        aiov.iov_base = uap->buf;
        aiov.iov_len = uap->nbyte;
        auio.uio_iov = &aiov;
        auio.uio_iovcnt = 1;
        auio.uio_offset = uap->offset;
        auio.uio_resid = uap->nbyte;
        auio.uio_rw = UIO_READ;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        flags = uap->flags & O_FMASK;
        if (uap->offset != (off_t)-1)
                flags |= O_FOFFSET;

        if (auio.uio_resid < 0)
                error = EINVAL;
        else
                error = kern_preadv(uap->fd, &auio, flags, &uap->sysmsg_result);
        return(error);
}

/*
 * Scatter read system call.
 *
 * MPSAFE
 */
int
sys_readv(struct readv_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov[UIO_SMALLIOV], *iov = NULL;
        int error;

        error = iovec_copyin(uap->iovp, &iov, aiov, uap->iovcnt,
                             &auio.uio_resid);
        if (error)
                return (error);
        auio.uio_iov = iov;
        auio.uio_iovcnt = uap->iovcnt;
        auio.uio_offset = -1;
        auio.uio_rw = UIO_READ;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        error = kern_preadv(uap->fd, &auio, 0, &uap->sysmsg_result);

        iovec_free(&iov, aiov);
        return (error);
}


/*
 * Scatter positioned read system call.
 *
 * MPSAFE
 */
int
sys_extpreadv(struct extpreadv_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov[UIO_SMALLIOV], *iov = NULL;
        int error;
        int flags;

        error = iovec_copyin(uap->iovp, &iov, aiov, uap->iovcnt,
                             &auio.uio_resid);
        if (error)
                return (error);
        auio.uio_iov = iov;
        auio.uio_iovcnt = uap->iovcnt;
        auio.uio_offset = uap->offset;
        auio.uio_rw = UIO_READ;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        flags = uap->flags & O_FMASK;
        if (uap->offset != (off_t)-1)
                flags |= O_FOFFSET;

        error = kern_preadv(uap->fd, &auio, flags, &uap->sysmsg_result);

        iovec_free(&iov, aiov);
        return(error);
}

/*
 * MPSAFE
 */
int
kern_preadv(int fd, struct uio *auio, int flags, int *res)
{
        struct thread *td = curthread;
        struct proc *p = td->td_proc;
        struct file *fp;
        int error;

        KKASSERT(p);

        fp = holdfp(p->p_fd, fd, FREAD);
        if (fp == NULL)
                return (EBADF);
        if (flags & O_FOFFSET && fp->f_type != DTYPE_VNODE) {
                error = ESPIPE;
        } else if (auio->uio_resid < 0) {
                error = EINVAL;
        } else {
                error = dofileread(fd, fp, auio, flags, res);
        }
        fdrop(fp);
        return(error);
}

/*
 * Common code for readv and preadv that reads data in
 * from a file using the passed in uio, offset, and flags.
 *
 * MPALMOSTSAFE - ktrace needs help
 */
static int
dofileread(int fd, struct file *fp, struct uio *auio, int flags, int *res)
{
        struct thread *td = curthread;
        int error;
        int len;
#ifdef KTRACE
        struct iovec *ktriov = NULL;
        struct uio ktruio;
#endif

#ifdef KTRACE
        /*
         * if tracing, save a copy of iovec
         */
        if (KTRPOINT(td, KTR_GENIO))  {
                int iovlen = auio->uio_iovcnt * sizeof(struct iovec);

                MALLOC(ktriov, struct iovec *, iovlen, M_TEMP, M_WAITOK);
                bcopy((caddr_t)auio->uio_iov, (caddr_t)ktriov, iovlen);
                ktruio = *auio;
        }
#endif
        len = auio->uio_resid;
        error = fo_read(fp, auio, fp->f_cred, flags);
        if (error) {
                if (auio->uio_resid != len && (error == ERESTART ||
                    error == EINTR || error == EWOULDBLOCK))
                        error = 0;
        }
#ifdef KTRACE
        if (ktriov != NULL) {
                if (error == 0) {
                        ktruio.uio_iov = ktriov;
                        ktruio.uio_resid = len - auio->uio_resid;
                        get_mplock();
                        ktrgenio(td->td_lwp, fd, UIO_READ, &ktruio, error);
                        rel_mplock();
                }
                FREE(ktriov, M_TEMP);
        }
#endif
        if (error == 0)
                *res = len - auio->uio_resid;

        return(error);
}

/*
 * Write system call
 *
 * MPSAFE
 */
int
sys_write(struct write_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov;
        int error;

        aiov.iov_base = (void *)(uintptr_t)uap->buf;
        aiov.iov_len = uap->nbyte;
        auio.uio_iov = &aiov;
        auio.uio_iovcnt = 1;
        auio.uio_offset = -1;
        auio.uio_resid = uap->nbyte;
        auio.uio_rw = UIO_WRITE;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        if (auio.uio_resid < 0)
                error = EINVAL;
        else
                error = kern_pwritev(uap->fd, &auio, 0, &uap->sysmsg_result);

        return(error);
}

/*
 * Pwrite system call
 *
 * MPSAFE
 */
int
sys_extpwrite(struct extpwrite_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov;
        int error;
        int flags;

        aiov.iov_base = (void *)(uintptr_t)uap->buf;
        aiov.iov_len = uap->nbyte;
        auio.uio_iov = &aiov;
        auio.uio_iovcnt = 1;
        auio.uio_offset = uap->offset;
        auio.uio_resid = uap->nbyte;
        auio.uio_rw = UIO_WRITE;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        flags = uap->flags & O_FMASK;
        if (uap->offset != (off_t)-1)
                flags |= O_FOFFSET;

        if (auio.uio_resid < 0)
                error = EINVAL;
        else
                error = kern_pwritev(uap->fd, &auio, flags, &uap->sysmsg_result);

        return(error);
}

/*
 * MPSAFE
 */
int
sys_writev(struct writev_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov[UIO_SMALLIOV], *iov = NULL;
        int error;

        error = iovec_copyin(uap->iovp, &iov, aiov, uap->iovcnt,
                             &auio.uio_resid);
        if (error)
                return (error);
        auio.uio_iov = iov;
        auio.uio_iovcnt = uap->iovcnt;
        auio.uio_offset = -1;
        auio.uio_rw = UIO_WRITE;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        error = kern_pwritev(uap->fd, &auio, 0, &uap->sysmsg_result);

        iovec_free(&iov, aiov);
        return (error);
}


/*
 * Gather positioned write system call
 *
 * MPSAFE
 */
int
sys_extpwritev(struct extpwritev_args *uap)
{
        struct thread *td = curthread;
        struct uio auio;
        struct iovec aiov[UIO_SMALLIOV], *iov = NULL;
        int error;
        int flags;

        error = iovec_copyin(uap->iovp, &iov, aiov, uap->iovcnt,
                             &auio.uio_resid);
        if (error)
                return (error);
        auio.uio_iov = iov;
        auio.uio_iovcnt = uap->iovcnt;
        auio.uio_offset = uap->offset;
        auio.uio_rw = UIO_WRITE;
        auio.uio_segflg = UIO_USERSPACE;
        auio.uio_td = td;

        flags = uap->flags & O_FMASK;
        if (uap->offset != (off_t)-1)
                flags |= O_FOFFSET;

        error = kern_pwritev(uap->fd, &auio, flags, &uap->sysmsg_result);

        iovec_free(&iov, aiov);
        return(error);
}

/*
 * MPSAFE
 */
int
kern_pwritev(int fd, struct uio *auio, int flags, int *res)
{
        struct thread *td = curthread;
        struct proc *p = td->td_proc;
        struct file *fp;
        int error;

        KKASSERT(p);

        fp = holdfp(p->p_fd, fd, FWRITE);
        if (fp == NULL)
                return (EBADF);
        else if ((flags & O_FOFFSET) && fp->f_type != DTYPE_VNODE) {
                error = ESPIPE;
        } else {
                error = dofilewrite(fd, fp, auio, flags, res);
        }
        
        fdrop(fp);
        return (error);
}

/*
 * Common code for writev and pwritev that writes data to
 * a file using the passed in uio, offset, and flags.
 *
 * MPALMOSTSAFE - ktrace needs help
 */
static int
dofilewrite(int fd, struct file *fp, struct uio *auio, int flags, int *res)
{       
        struct thread *td = curthread;
        struct lwp *lp = td->td_lwp;
        int error;
        int len;
#ifdef KTRACE
        struct iovec *ktriov = NULL;
        struct uio ktruio;
#endif

#ifdef KTRACE
        /*
         * if tracing, save a copy of iovec and uio
         */
        if (KTRPOINT(td, KTR_GENIO))  {
                int iovlen = auio->uio_iovcnt * sizeof(struct iovec);

                MALLOC(ktriov, struct iovec *, iovlen, M_TEMP, M_WAITOK);
                bcopy((caddr_t)auio->uio_iov, (caddr_t)ktriov, iovlen);
                ktruio = *auio;
        }
#endif
        len = auio->uio_resid;
        error = fo_write(fp, auio, fp->f_cred, flags);
        if (error) {
                if (auio->uio_resid != len && (error == ERESTART ||
                    error == EINTR || error == EWOULDBLOCK))
                        error = 0;
                /* Socket layer is responsible for issuing SIGPIPE. */
                if (error == EPIPE) {
                        get_mplock();
                        lwpsignal(lp->lwp_proc, lp, SIGPIPE);
                        rel_mplock();
                }
        }
#ifdef KTRACE
        if (ktriov != NULL) {
                if (error == 0) {
                        ktruio.uio_iov = ktriov;
                        ktruio.uio_resid = len - auio->uio_resid;
                        get_mplock();
                        ktrgenio(lp, fd, UIO_WRITE, &ktruio, error);
                        rel_mplock();
                }
                FREE(ktriov, M_TEMP);
        }
#endif
        if (error == 0)
                *res = len - auio->uio_resid;

        return(error);
}

/*
 * Ioctl system call
 */
/* ARGSUSED */
int
sys_ioctl(struct ioctl_args *uap)
{
        return(mapped_ioctl(uap->fd, uap->com, uap->data, NULL));
}

struct ioctl_map_entry {
        const char *subsys;
        struct ioctl_map_range *cmd_ranges;
        LIST_ENTRY(ioctl_map_entry) entries;
};

/*
 * The true heart of all ioctl syscall handlers (native, emulation).
 * If map != NULL, it will be searched for a matching entry for com,
 * and appropriate conversions/conversion functions will be utilized.
 */
int
mapped_ioctl(int fd, u_long com, caddr_t uspc_data, struct ioctl_map *map)
{
        struct thread *td = curthread;
        struct proc *p = td->td_proc;
        struct ucred *cred;
        struct file *fp;
        struct ioctl_map_range *iomc = NULL;
        int error;
        u_int size;
        u_long ocom = com;
        caddr_t data, memp;
        int tmp;
#define STK_PARAMS      128
        union {
            char stkbuf[STK_PARAMS];
            long align;
        } ubuf;

        KKASSERT(p);
        cred = p->p_ucred;

        fp = holdfp(p->p_fd, fd, FREAD|FWRITE);
        if (fp == NULL)
                return(EBADF);

        if (map != NULL) {      /* obey translation map */
                u_long maskcmd;
                struct ioctl_map_entry *e;

                maskcmd = com & map->mask;

                LIST_FOREACH(e, &map->mapping, entries) {
                        for (iomc = e->cmd_ranges; iomc->start != 0 ||
                             iomc->maptocmd != 0 || iomc->wrapfunc != NULL ||
                             iomc->mapfunc != NULL;
                             iomc++) {
                                if (maskcmd >= iomc->start &&
                                    maskcmd <= iomc->end)
                                        break;
                        }

                        /* Did we find a match? */
                        if (iomc->start != 0 || iomc->maptocmd != 0 ||
                            iomc->wrapfunc != NULL || iomc->mapfunc != NULL)
                                break;
                }

                if (iomc == NULL ||
                    (iomc->start == 0 && iomc->maptocmd == 0
                     && iomc->wrapfunc == NULL && iomc->mapfunc == NULL)) {
                        kprintf("%s: 'ioctl' fd=%d, cmd=0x%lx ('%c',%d) not implemented\n",
                               map->sys, fd, maskcmd,
                               (int)((maskcmd >> 8) & 0xff),
                               (int)(maskcmd & 0xff));
                        error = EINVAL;
                        goto done;
                }

                /*
                 * If it's a non-range one to one mapping, maptocmd should be
                 * correct. If it's a ranged one to one mapping, we pass the
                 * original value of com, and for a range mapped to a different
                 * range, we always need a mapping function to translate the
                 * ioctl to our native ioctl. Ex. 6500-65ff <-> 9500-95ff
                 */
                if (iomc->start == iomc->end && iomc->maptocmd == iomc->maptoend) {
                        com = iomc->maptocmd;
                } else if (iomc->start == iomc->maptocmd && iomc->end == iomc->maptoend) {
                        if (iomc->mapfunc != NULL)
                                com = iomc->mapfunc(iomc->start, iomc->end,
                                                    iomc->start, iomc->end,
                                                    com, com);
                } else {
                        if (iomc->mapfunc != NULL) {
                                com = iomc->mapfunc(iomc->start, iomc->end,
                                                    iomc->maptocmd, iomc->maptoend,
                                                    com, ocom);
                        } else {
                                kprintf("%s: Invalid mapping for fd=%d, cmd=%#lx ('%c',%d)\n",
                                       map->sys, fd, maskcmd,
                                       (int)((maskcmd >> 8) & 0xff),
                                       (int)(maskcmd & 0xff));
                                error = EINVAL;
                                goto done;
                        }
                }
        }

        switch (com) {
        case FIONCLEX:
                error = fclrfdflags(p->p_fd, fd, UF_EXCLOSE);
                goto done;
        case FIOCLEX:
                error = fsetfdflags(p->p_fd, fd, UF_EXCLOSE);
                goto done;
        }

        /*
         * Interpret high order word to find amount of data to be
         * copied to/from the user's address space.
         */
        size = IOCPARM_LEN(com);
        if (size > IOCPARM_MAX) {
                error = ENOTTY;
                goto done;
        }

        memp = NULL;
        if (size > sizeof (ubuf.stkbuf)) {
                memp = kmalloc(size, M_IOCTLOPS, M_WAITOK);
                data = memp;
        } else {
                data = ubuf.stkbuf;
        }
        if ((com & IOC_IN) != 0) {
                if (size != 0) {
                        error = copyin(uspc_data, data, (u_int)size);
                        if (error) {
                                if (memp != NULL)
                                        kfree(memp, M_IOCTLOPS);
                                goto done;
                        }
                } else {
                        *(caddr_t *)data = uspc_data;
                }
        } else if ((com & IOC_OUT) != 0 && size) {
                /*
                 * Zero the buffer so the user always
                 * gets back something deterministic.
                 */
                bzero(data, size);
        } else if ((com & IOC_VOID) != 0) {
                *(caddr_t *)data = uspc_data;
        }

        switch (com) {
        case FIONBIO:
                if ((tmp = *(int *)data))
                        fp->f_flag |= FNONBLOCK;
                else
                        fp->f_flag &= ~FNONBLOCK;
                error = 0;
                break;

        case FIOASYNC:
                if ((tmp = *(int *)data))
                        fp->f_flag |= FASYNC;
                else
                        fp->f_flag &= ~FASYNC;
                error = fo_ioctl(fp, FIOASYNC, (caddr_t)&tmp, cred);
                break;

        default:
                /*
                 *  If there is a override function,
                 *  call it instead of directly routing the call
                 */
                if (map != NULL && iomc->wrapfunc != NULL)
                        error = iomc->wrapfunc(fp, com, ocom, data, cred);
                else
                        error = fo_ioctl(fp, com, data, cred);
                /*
                 * Copy any data to user, size was
                 * already set and checked above.
                 */
                if (error == 0 && (com & IOC_OUT) != 0 && size != 0)
                        error = copyout(data, uspc_data, (u_int)size);
                break;
        }
        if (memp != NULL)
                kfree(memp, M_IOCTLOPS);
done:
        fdrop(fp);
        return(error);
}

int
mapped_ioctl_register_handler(struct ioctl_map_handler *he)
{
        struct ioctl_map_entry *ne;

        KKASSERT(he != NULL && he->map != NULL && he->cmd_ranges != NULL &&
                 he->subsys != NULL && *he->subsys != '\0');

        ne = kmalloc(sizeof(struct ioctl_map_entry), M_IOCTLMAP, M_WAITOK);

        ne->subsys = he->subsys;
        ne->cmd_ranges = he->cmd_ranges;

        LIST_INSERT_HEAD(&he->map->mapping, ne, entries);

        return(0);
}

int
mapped_ioctl_unregister_handler(struct ioctl_map_handler *he)
{
        struct ioctl_map_entry *ne;

        KKASSERT(he != NULL && he->map != NULL && he->cmd_ranges != NULL);

        LIST_FOREACH(ne, &he->map->mapping, entries) {
                if (ne->cmd_ranges != he->cmd_ranges)
                        continue;
                LIST_REMOVE(ne, entries);
                kfree(ne, M_IOCTLMAP);
                return(0);
        }
        return(EINVAL);
}

static int      nselcoll;       /* Select collisions since boot */
int     selwait;
SYSCTL_INT(_kern, OID_AUTO, nselcoll, CTLFLAG_RD, &nselcoll, 0, "");

/*
 * Select system call.
 */
int
sys_select(struct select_args *uap)
{
        struct timeval ktv;
        struct timeval *ktvp;
        int error;

        /*
         * Get timeout if any.
         */
        if (uap->tv != NULL) {
                error = copyin(uap->tv, &ktv, sizeof (ktv));
                if (error)
                        return (error);
                error = itimerfix(&ktv);
                if (error)
                        return (error);
                ktvp = &ktv;
        } else {
                ktvp = NULL;
        }

        /*
         * Do real work.
         */
        error = doselect(uap->nd, uap->in, uap->ou, uap->ex, ktvp,
                        &uap->sysmsg_result);

        return (error);
}


/*
 * Pselect system call.
 */
int
sys_pselect(struct pselect_args *uap)
{
        struct thread *td = curthread;
        struct lwp *lp = td->td_lwp;
        struct timespec kts;
        struct timeval ktv;
        struct timeval *ktvp;
        sigset_t sigmask;
        int error;

        /*
         * Get timeout if any and convert it.
         * Round up during conversion to avoid timeout going off early.
         */
        if (uap->ts != NULL) {
                error = copyin(uap->ts, &kts, sizeof (kts));
                if (error)
                        return (error);
                ktv.tv_sec = kts.tv_sec;
                ktv.tv_usec = (kts.tv_nsec + 999) / 1000;
                error = itimerfix(&ktv);
                if (error)
                        return (error);
                ktvp = &ktv;
        } else {
                ktvp = NULL;
        }

        /*
         * Install temporary signal mask if any provided.
         */
        if (uap->sigmask != NULL) {
                error = copyin(uap->sigmask, &sigmask, sizeof(sigmask));
                if (error)
                        return (error);
                lp->lwp_oldsigmask = lp->lwp_sigmask;
                SIG_CANTMASK(sigmask);
                lp->lwp_sigmask = sigmask;
        }

        /*
         * Do real job.
         */
        error = doselect(uap->nd, uap->in, uap->ou, uap->ex, ktvp,
                        &uap->sysmsg_result);

        if (uap->sigmask != NULL) {
                /* doselect() responsible for turning ERESTART into EINTR */
                KKASSERT(error != ERESTART);
                if (error == EINTR) {
                        /*
                         * We can't restore the previous signal mask now
                         * because it could block the signal that interrupted
                         * us.  So make a note to restore it after executing
                         * the handler.
                         */
                        lp->lwp_flag |= LWP_OLDMASK;
                } else {
                        /*
                         * No handler to run. Restore previous mask immediately.
                         */
                        lp->lwp_sigmask = lp->lwp_oldsigmask;
                }
        }

        return (error);
}

/*
 * Common code for sys_select() and sys_pselect().
 *
 * in, out and ex are userland pointers.  tv must point to validated
 * kernel-side timeout value or NULL for infinite timeout.  res must
 * point to syscall return value.
 */
static int
doselect(int nd, fd_set *in, fd_set *ou, fd_set *ex, struct timeval *tv,
                int *res)
{
        struct lwp *lp = curthread->td_lwp;
        struct proc *p = curproc;

        /*
         * The magic 2048 here is chosen to be just enough for FD_SETSIZE
         * infds with the new FD_SETSIZE of 1024, and more than enough for
         * FD_SETSIZE infds, outfds and exceptfds with the old FD_SETSIZE
         * of 256.
         */
        fd_mask s_selbits[howmany(2048, NFDBITS)];
        fd_mask *ibits[3], *obits[3], *selbits, *sbp;
        struct timeval atv, rtv, ttv;
        int ncoll, error, timo;
        u_int nbufbytes, ncpbytes, nfdbits;

        if (nd < 0)
                return (EINVAL);
        if (nd > p->p_fd->fd_nfiles)
                nd = p->p_fd->fd_nfiles;   /* forgiving; slightly wrong */

        /*
         * Allocate just enough bits for the non-null fd_sets.  Use the
         * preallocated auto buffer if possible.
         */
        nfdbits = roundup(nd, NFDBITS);
        ncpbytes = nfdbits / NBBY;
        nbufbytes = 0;
        if (in != NULL)
                nbufbytes += 2 * ncpbytes;
        if (ou != NULL)
                nbufbytes += 2 * ncpbytes;
        if (ex != NULL)
                nbufbytes += 2 * ncpbytes;
        if (nbufbytes <= sizeof s_selbits)
                selbits = &s_selbits[0];
        else
                selbits = kmalloc(nbufbytes, M_SELECT, M_WAITOK);

        /*
         * Assign pointers into the bit buffers and fetch the input bits.
         * Put the output buffers together so that they can be bzeroed
         * together.
         */
        sbp = selbits;
#define getbits(name, x) \
        do {                                                            \
                if (name == NULL)                                       \
                        ibits[x] = NULL;                                \
                else {                                                  \
                        ibits[x] = sbp + nbufbytes / 2 / sizeof *sbp;   \
                        obits[x] = sbp;                                 \
                        sbp += ncpbytes / sizeof *sbp;                  \
                        error = copyin(name, ibits[x], ncpbytes);       \
                        if (error != 0)                                 \
                                goto done;                              \
                }                                                       \
        } while (0)
        getbits(in, 0);
        getbits(ou, 1);
        getbits(ex, 2);
#undef  getbits
        if (nbufbytes != 0)
                bzero(selbits, nbufbytes / 2);

        if (tv != NULL) {
                atv = *tv;
                getmicrouptime(&rtv);
                timevaladd(&atv, &rtv);
        } else {
                atv.tv_sec = 0;
                atv.tv_usec = 0;
        }
        timo = 0;
retry:
        ncoll = nselcoll;
        lp->lwp_flag |= LWP_SELECT;
        error = selscan(p, ibits, obits, nd, res);
        if (error || *res)
                goto done;
        if (atv.tv_sec || atv.tv_usec) {
                getmicrouptime(&rtv);
                if (timevalcmp(&rtv, &atv, >=))
                        goto done;
                ttv = atv;
                timevalsub(&ttv, &rtv);
                timo = ttv.tv_sec > 24 * 60 * 60 ?
                    24 * 60 * 60 * hz : tvtohz_high(&ttv);
        }
        crit_enter();
        if ((lp->lwp_flag & LWP_SELECT) == 0 || nselcoll != ncoll) {
                crit_exit();
                goto retry;
        }
        lp->lwp_flag &= ~LWP_SELECT;

        error = tsleep((caddr_t)&selwait, PCATCH, "select", timo);
        
        crit_exit();
        if (error == 0)
                goto retry;
done:
        lp->lwp_flag &= ~LWP_SELECT;
        /* select is not restarted after signals... */
        if (error == ERESTART)
                error = EINTR;
        if (error == EWOULDBLOCK)
                error = 0;
#define putbits(name, x) \
        if (name && (error2 = copyout(obits[x], name, ncpbytes))) \
                error = error2;
        if (error == 0) {
                int error2;

                putbits(in, 0);
                putbits(ou, 1);
                putbits(ex, 2);
#undef putbits
        }
        if (selbits != &s_selbits[0])
                kfree(selbits, M_SELECT);
        return (error);
}

static int
selscan(struct proc *p, fd_mask **ibits, fd_mask **obits, int nfd, int *res)
{
        int msk, i, fd;
        fd_mask bits;
        struct file *fp;
        int n = 0;
        /* Note: backend also returns POLLHUP/POLLERR if appropriate. */
        static int flag[3] = { POLLRDNORM, POLLWRNORM, POLLRDBAND };

        for (msk = 0; msk < 3; msk++) {
                if (ibits[msk] == NULL)
                        continue;
                for (i = 0; i < nfd; i += NFDBITS) {
                        bits = ibits[msk][i/NFDBITS];
                        /* ffs(int mask) not portable, fd_mask is long */
                        for (fd = i; bits && fd < nfd; fd++, bits >>= 1) {
                                if (!(bits & 1))
                                        continue;
                                fp = holdfp(p->p_fd, fd, -1);
                                if (fp == NULL)
                                        return (EBADF);
                                if (fo_poll(fp, flag[msk], fp->f_cred)) {
                                        obits[msk][(fd)/NFDBITS] |=
                                            ((fd_mask)1 << ((fd) % NFDBITS));
                                        n++;
                                }
                                fdrop(fp);
                        }
                }
        }
        *res = n;
        return (0);
}

/*
 * Poll system call.
 */
int
sys_poll(struct poll_args *uap)
{
        struct pollfd *bits;
        struct pollfd smallbits[32];
        struct timeval atv, rtv, ttv;
        int ncoll, error = 0, timo;
        u_int nfds;
        size_t ni;
        struct lwp *lp = curthread->td_lwp;
        struct proc *p = curproc;

        nfds = uap->nfds;
        /*
         * This is kinda bogus.  We have fd limits, but that is not
         * really related to the size of the pollfd array.  Make sure
         * we let the process use at least FD_SETSIZE entries and at
         * least enough for the current limits.  We want to be reasonably
         * safe, but not overly restrictive.
         */
        if (nfds > p->p_rlimit[RLIMIT_NOFILE].rlim_cur && nfds > FD_SETSIZE)
                return (EINVAL);
        ni = nfds * sizeof(struct pollfd);
        if (ni > sizeof(smallbits))
                bits = kmalloc(ni, M_TEMP, M_WAITOK);
        else
                bits = smallbits;
        error = copyin(uap->fds, bits, ni);
        if (error)
                goto done;
        if (uap->timeout != INFTIM) {
                atv.tv_sec = uap->timeout / 1000;
                atv.tv_usec = (uap->timeout % 1000) * 1000;
                if (itimerfix(&atv)) {
                        error = EINVAL;
                        goto done;
                }
                getmicrouptime(&rtv);
                timevaladd(&atv, &rtv);
        } else {
                atv.tv_sec = 0;
                atv.tv_usec = 0;
        }
        timo = 0;
retry:
        ncoll = nselcoll;
        lp->lwp_flag |= LWP_SELECT;
        error = pollscan(p, bits, nfds, &uap->sysmsg_result);
        if (error || uap->sysmsg_result)
                goto done;
        if (atv.tv_sec || atv.tv_usec) {
                getmicrouptime(&rtv);
                if (timevalcmp(&rtv, &atv, >=))
                        goto done;
                ttv = atv;
                timevalsub(&ttv, &rtv);
                timo = ttv.tv_sec > 24 * 60 * 60 ?
                    24 * 60 * 60 * hz : tvtohz_high(&ttv);
        } 
        crit_enter();
        if ((lp->lwp_flag & LWP_SELECT) == 0 || nselcoll != ncoll) {
                crit_exit();
                goto retry;
        }
        lp->lwp_flag &= ~LWP_SELECT;
        error = tsleep((caddr_t)&selwait, PCATCH, "poll", timo);
        crit_exit();
        if (error == 0)
                goto retry;
done:
        lp->lwp_flag &= ~LWP_SELECT;
        /* poll is not restarted after signals... */
        if (error == ERESTART)
                error = EINTR;
        if (error == EWOULDBLOCK)
                error = 0;
        if (error == 0) {
                error = copyout(bits, uap->fds, ni);
                if (error)
                        goto out;
        }
out:
        if (ni > sizeof(smallbits))
                kfree(bits, M_TEMP);
        return (error);
}

static int
pollscan(struct proc *p, struct pollfd *fds, u_int nfd, int *res)
{
        int i;
        struct file *fp;
        int n = 0;

        for (i = 0; i < nfd; i++, fds++) {
                if (fds->fd >= p->p_fd->fd_nfiles) {
                        fds->revents = POLLNVAL;
                        n++;
                } else if (fds->fd < 0) {
                        fds->revents = 0;
                } else {
                        fp = holdfp(p->p_fd, fds->fd, -1);
                        if (fp == NULL) {
                                fds->revents = POLLNVAL;
                                n++;
                        } else {
                                /*
                                 * Note: backend also returns POLLHUP and
                                 * POLLERR if appropriate.
                                 */
                                fds->revents = fo_poll(fp, fds->events,
                                                        fp->f_cred);
                                if (fds->revents != 0)
                                        n++;
                                fdrop(fp);
                        }
                }
        }
        *res = n;
        return (0);
}

/*
 * OpenBSD poll system call.
 * XXX this isn't quite a true representation..  OpenBSD uses select ops.
 */
int
sys_openbsd_poll(struct openbsd_poll_args *uap)
{
        return (sys_poll((struct poll_args *)uap));
}

/*ARGSUSED*/
int
seltrue(cdev_t dev, int events)
{
        return (events & (POLLIN | POLLOUT | POLLRDNORM | POLLWRNORM));
}

/*
 * Record a select request.  A global wait must be used since a process/thread
 * might go away after recording its request.
 */
void
selrecord(struct thread *selector, struct selinfo *sip)
{
        struct proc *p;
        struct lwp *lp = NULL;

        if (selector->td_lwp == NULL)
                panic("selrecord: thread needs a process");

        if (sip->si_pid == selector->td_proc->p_pid &&
            sip->si_tid == selector->td_lwp->lwp_tid)
                return;
        if (sip->si_pid && (p = pfind(sip->si_pid)))
                lp = lwp_rb_tree_RB_LOOKUP(&p->p_lwp_tree, sip->si_tid);
        if (lp != NULL && lp->lwp_wchan == (caddr_t)&selwait) {
                sip->si_flags |= SI_COLL;
        } else {
                sip->si_pid = selector->td_proc->p_pid;
                sip->si_tid = selector->td_lwp->lwp_tid;
        }
}

/*
 * Do a wakeup when a selectable event occurs.
 */
void
selwakeup(struct selinfo *sip)
{
        struct proc *p;
        struct lwp *lp = NULL;

        if (sip->si_pid == 0)
                return;
        if (sip->si_flags & SI_COLL) {
                nselcoll++;
                sip->si_flags &= ~SI_COLL;
                wakeup((caddr_t)&selwait);      /* YYY fixable */
        }
        p = pfind(sip->si_pid);
        sip->si_pid = 0;
        if (p == NULL)
                return;
        lp = lwp_rb_tree_RB_LOOKUP(&p->p_lwp_tree, sip->si_tid);
        if (lp == NULL)
                return;

        crit_enter();
        if (lp->lwp_wchan == (caddr_t)&selwait) {
                /*
                 * Flag the process to break the tsleep when
                 * setrunnable is called, but only call setrunnable
                 * here if the process is not in a stopped state.
                 */
                lp->lwp_flag |= LWP_BREAKTSLEEP;
                if (p->p_stat != SSTOP)
                        setrunnable(lp);
        } else if (lp->lwp_flag & LWP_SELECT) {
                lp->lwp_flag &= ~LWP_SELECT;
        }
        crit_exit();
}